Handbook

of
NOISE
MEASUREMENT
by Arnold P. G. Pete'son
and Ervin E. Gross, Jr.

Pr u 50

SEVENTH EDITION
Copyright 1963, 1967, 1972, 1974 by General Radio Company,
Concord, Massachusetts U.S.A.
Form No. 5301-8111-L

aGeneral Radio

PRINTED IN U.S.A.
 Preface

A new chapter on hearing conservation by our col-

leagues at the Grason-Stadler Company is a particularly
noteworthy addition to this handbook. It reflects the in-
creased concern for hearing damage from noise exposure,
which is now subject to extensive monitoring as a result
of the Occupational Safety and Health Act.
The increased availability and use of more sophisticated
noise measurement systems has led us to expand the section
on analysis into aseparate chapter, which includes informa-
tion on digital as well as analog systems.
Many new instruments and an electret transducer that is
new in the measurement field are reviewed here.
Many of the other sections have been revised to bring
them up to date and to present a picture of the present
state of the art from our point of view. Those areas that are
still controversial should be evident from our comments
and the references.
In this edition the references are more specific and more
extensive in order to make it easier for the reader to pursue
a particular subject area more thoroughly than is possible
here, but we have not attempted to produce acomprehen-
sive listing.
We appreciate the comments and corrections received
from users of the previous editions and from our colleagues.
In particular, we gratefully acknowledge the extensive help
received from R. A. Boole, W. R. Kundert, and W. E.
Collins in the preparation of this edition.

Arnold P. G. Peterson
Ervin E. Gross
 Table of Contents
Chapter 1. Introduction 1

Chapter 2. What Are Noise And Vibration? 3

2.1 Introduction 3
2.2 The Decibel - What Is It? 4
2.3 Power Level 5
2.4 Sound-Pressure Level 7
2.5 Sound Level 7
2.6 Combining Decibels 9
2.7 Vibration 9
2.8 Summary 13

Chapter 3. What Noise And Vibration Do And How Much

Is Acceptable? 15

3.1 Why We Measure Noise 15

3.2 Psychoacoustical Experiments 15
3.3 Thresholds Of Hearing And Tolerance 17
3.4 "What Noise Annoys An Oyster?" . 20
3.5 Rating The Loudness Of A Sound 21
3.6 Perceived-Noise Level 27
3.7 Perceived Level - Stevens's Mark VII 28
3.8 Effect Of Tonal Components 33
3.9 Effect Of Duration 33
3.10 NC Curves 34
3.11 Noise And Number Index 35
3.12 Noise-Pollution Level 35
3.13 Masking - " ICan't Hear You When The Water's
Running" 36
3.14 Critical Ratio And Critical Bandwidth 40
3.15 Additional Hearing Characteristics 41
3.16 Weighted Sound Levels 42
3.17 Residential Noise Levels 45
3.18 Effect Of Noise On Work Output 46
3.19 Non-Auditory Effects Of Noise On Man 47
3.20 What Vibration Does 47
Chapter 4. Hearing-Conservation Programs In Industry 57

4.1 Introduction 57
4.2 The Human Ear 57
4.3 Noise-Induced Hearing Loss 58
4.4 Hearing-Conservation Programs 60
4.5 Hearing Monitoring 63
4.6 Personnel 67
4.7 Audiometry Program 68
4.8 Appropriate Legislation 69

Chapter 5. Analysis 73

5.1 Introduction 73
5.2 Analysis In Frequency Bands 74
5.3 Time Series And Sampling 78
5.4 Analysis In The Time Domain 81
5.5 Analysis In Amplitude Domain 85
5.6 Techniques And Characteristics Of Analysis 89
5.7 Two-Signal Functions 104
5.8 Analog Vs Digital 105

Chapter 6. Instrumentation For Noise And Vibration

Measurement 109

6.1 General 109

6.2 Transducers - Transducer Systems 110
6.3 Sound-Level Meters 116
6.4 Noise- Exposure Meter 118
6.5 Vibration Meter 118
6.6 Analyzers 120
6.7 Graphic Recorder 128
6.8 Magnetic-Tape Recorder 129
6.9 Impulse-Noise Measuring Instruments 131
6.10 Random-Noise Generators 131
6.11 Continuously Adjustable Filter 132
6.12 Tone- Burst Generator 133
6.13 Automatic- Level Regulator 134
6.14 Oscilloscope 134
6.15 Vibration Shakers 135
6.16 Calibrators 136
6.17 Stroboscopes 137
6.18 Systems, Computers, And Peripheral Equipment 140

Chapter 7. What Noise And Vibration Measurements

Should Be Made 151

7.1 Introduction 151

7.2 Devices That Are Noisy Or Vibrate Excessively . 152
7.3 Environmental Noise 160
7.4 Architectural Acoustics 162
 Chapter 8. Techniques, Precautions, And Calibrations . . 169

8.1 Introduction
169
8.2 Sound Measurements 170
8.3 Sound Fields
188
8.4 Vibration Measurement Techniques 206

Chapter 9. Noise And Vibration Control 217

9.1 Introduction
217
9.2 Noise Control At The Source 217
9.3 Control Of The Path Of Sound 218
9.4 Summary Of Noise Reduction Procedures 222
9.5 Vibration Reduction 222

Chapter 10. Some Case Histories 233

10.1 Noise From An Air Compressor 233

10.2 Braiding-Machine Noise 235
10.3 Reduction Of Flutter In A Tape Recorder 236
10.4 An Oil-Pump Problem
236
10.5 Summaries Of Some Examples From The literature 240

Appendixes

I Decibel Conversion Tables 244

Chart for Combining Levels of Uncorrelated Noise
Signals
252
Ill Table for Converting Loudness to Loudness Level . 253
IV Vibration Conversion Charts 254
V Definitions
256
VI Words Commonly Used to Describe Sounds 269
VII Standards and Journals 270
VIII Hearing-Loss Statutes
275
IX Catalog Section
278
Index
318
 In any of these, sound-measuring instruments and systems can help to
assess the nature of the problem, and they can help in determining what to do
to subdue the troublesome noise.
The study of mechanical vibration is closely related to that of sound,
because sound is produced by the transfer of mechanical vibration to air.
Hence, the process of quieting a machine or device often includes astudy of
the vibrations involved.
Conversely, high-energy acoustical noise, such as generated by powerful jet
or rocket engines, can produce vibrations that can weaken structural members
of avehicle or cause electronic components to fail.
Other important effects of vibration include: human discomfort and
fatigue from excessive vibration of avehicle, fatigue and rupture of structural
members, and increased maintenance of machines, appliances, vehicles, and
other devices.
Vibration, then, is asource not only of noise, annoyance, and discomfort,
but often of danger as well. The present refinement of high-speed planes,
ships, and automobiles could never have been achieved without thorough
measurement and study of mechanical vibration.
The instruments used in sound and vibration measurement are mainly
electronic. Furthermore, some of the concepts and techniques developed by
electronics engineers and physicists for dealing with random or interfering
signals (for which they have borrowed the term " noise") are now used in
sound and vibration studies.
The purpose of this book is to help those who are faced, possibly for the
first time, with the necessity of making noise measurements. It attempts to
clarify the terminology and definitions used in these measurements, to
describe the measuring instruments and their use, to aid the prospective user
in selecting the proper equipment for the measurements he must make, and
to show how these measurements can be interpreted to solve typical prob-
lems.
Although some may wish to read the chapters of this book in sequence,
many will find it more convenient to consult the table of contents or the
index to find the sections of immediate interest. They then can refer to the
other sections of the book as they need further information. For example, if
hearing conservation is of primary concern, Chapter 4 could be read first.
Chapter 7 ("What Noise and Vibration Measurements Should be Made")
could be consulted if aspecific noise problem is at hand. The reader can then
find further details on the instruments recommended (Chapter 6) and on the
techniques of use (Chapter 8).
Some sections of this book are marked by adiamond to indicate that they
might well be omitted during an initial reading, since they are highly special-
ized or very technical.

2
 In any of these, sound-measuring instruments and systems can help to
assess the nature of the problem, and they can help in determining what to do
to subdue the troublesome noise.
The study of mechanical vibration is closely related to that of sound,
because sound is produced by the transfer of mechanical vibration to air.
Hence, the process of quieting a machine or device often includes astudy of
the vibrations involved.
Conversely, high-energy acoustical noise, such as generated by powerful jet
or rocket engines, can produce vibrations that can weaken structural members
of avehicle or cause electronic components to fail.
Other important effects of vibration include: human discomfort and
fatigue from excessive vibration of avehicle, fatigue and rupture of structural
members, and increased maintenance of machines, appliances, vehicles, and
other devices.
Vibration, then, is asource not only of noise, annoyance, and discomfort,
but often of danger as well. The present refinement of high-speed planes,
ships, and automobiles could never have been achieved without thorough
measurement and study of mechanical vibration.
The instruments used in sound and vibration measurement are mainly
electronic. Furthermore, some of the concepts and techniques developed by
electronics engineers and physicists for dealing with random or interfering
signals (for which they have borrowed the term " noise") are now used in
sound and vibration studies.
The purpose of this book is to help those who are faced, possibly for the
first time, with the necessity of making noise measurements. It attempts to
clarify the terminology and definitions used in these measurements, to
describe the measuring instruments and their use, to aid the prospective user
in selecting the proper equipment for the measurements he must make, and
to show how these measurements can be interpreted to solve typical prob-
lems.
Although some may wish to read the chapters of this book in sequence,
many will find it more convenient to consult the table of contents or the
index to find the sections of immediate interest. They then can refer to the
other sections of the book as they need further information. For example, if
hearing conservation is of primary concern, Chapter 4 could be read first.
Chapter 7 ("What Noise and Vibration Measurements Should be Made")
could be consulted if aspecific noise problem is at hand. The reader can then
find further details on the instruments recommended (Chapter 6) and on the
techniques of use (Chapter 8).
Some sections of this book are marked by adiamond to indicate that they
might well be omitted during an initial reading, since they are highly special-
ized or very technical.

2
Chapter 2

What are Noise

and Vibration
2.1 INTRODUCTION.
When an object moves back and forth, it is said to vibrate. This vibration
disturbs the air particles near the object and sets them vibrating, producing a
variation in normal atmospheric pressure. The disturbance spreads and, when
the pressure variations reach our ear drums, they too are set to vibrating. This
vibration of our ear drums is translated by our complicated hearing mecha-
nisms into the sensation we call "sound."
To put it in more general terms, sound in the physical sense is avibration
of particles in a gas, a liquid, or a solid. The measurement and control of
airborne sound is the basic subject of this book. Because the chief sources of
sounds in air are vibrations of solid objects, the measurement and control of
vibration will also be discussed. Vibrations of and in solids often have
important effects other than those classified as sound, and some of these will
also be included.
We have mentioned that a sound disturbance spreads. The speed with
which it spreads depends on the mass and on the elastic properties of the
material. In air the speed is about 1100 feet/second (about 750 miles/hour)
or about 340 meters/second; in sea water it is about 1490 meters/second. The
speed of sound has been popularized in aerodynamic concepts of the sound
barrier and the supersonic transport, and its effects are commonly observed in
echoes and in the apparent delay between a flash of lightning and the
accompanying thunder.
The variation in normal atmospheric pressure that is a part of a sound
wave is characterized by the rate at which the variation occurs and the extent
of the variation. Thus, the standard tone "A" occurs when the pressure
changes through a complete cycle 440 times per second. The frequency of
this tone is then said to be 440 hertz, or 440 cycles per second (abbreviated
"Hz" and " c/s", respectively). " Hertz" and "cycles per second" are synony-
mous terms, but most standardizing agencies have adopted "hertz" as the
preferred unit of frequency.
Many prefixes are used with the unit of frequency, but the one that is
common in acoustics and vibrations is "kilo-," abbreviated "k", which stands
for afactor of 1000. Thus, 8000 Hz or 8000 c/s becomes 8kHz or 8kc/s.
The extent of the variation in pressure is measured in terms of aunit called
the " microbar*", which is approximately one-millionth of the normal atmos-
pheric pressure (standard atmospheric pressure = 1,013,250 microbars), or in
terms of newtons per square meter, which is 10 microbars. Actually, these
units are not often mentioned in noise measurement. Results are stated in
decibels.

'Here, the prefix " micro" stands for a factor of one- millionth, and that prefix is
abbreviated by the use of the Greek letter "le ( mu). Thus "µN" stands for 0.000001
newton.

3
2.2 THE DECIBEL - WHAT IS IT?
Although to many laymen the decibel (abbreviated " dB") is uniquely
associated with noise measurements, it is a term borrowed from electrical-
communication engineering, and it represents arelative quantity. When it is
used to express noise level, a reference quantity is implied. Usually, this
reference value is a sound pressure of 20 rnicronewtons per square meter
(abbreviated 20 111%1/m 2 ). For the present, the reference level can be referred
to as "0 decibels," the starting point of the scale of noise levels. This starting
point is about the level of the weakest sound that can be heard by aperson
with very good hearing in an extremely quiet location. Other typical points
on this scale of noise levels are shown in Figure 2-1. For example, the noise

TYPICAL A- WEIGHTED SOUND LEVELS

AT A GIVEN DISTANCE FROM NOISE SOURCE ENV.RONRENT AL
DECIBELS

RE no .2

140

50 HP SIREN 1100)
130

JET TAKEOFF 1200

120

RIVETING MACHINE HO CASTING SHAKEOUT AREA

CUT-OFF SAY/
•PNEUMATIC PEEN HAmmER
100 ELECTRIC FURNACE AREA

•TEXTILE WEAVING PLANT

SUBWAY TRAIN ( 201 90 BOILER ROOM
PRINTING PRESS PLANT

PNEUMATIC DRILL (
se)
80 TABULATING ROOM
INSIDE SPORT CAR ISO 0APHI

FREIGHT TRAIN 11001

VACUUM CLEANER 1101 70
SPEECH II',

NEAR FREEWAY ( AUTO TRAFFIC)

60 LARGE STORE
ACCOUNTING OFFICE

LARGE TRANSFORMER - 202 PRIVATE BUSINESS OFFICE

50 LIGHT TRAFFIC 11031

AVERAGE RESIDENCE

40 RIN LEVELS • RESIDENTIAL AREAS IN

CHICAGO AT NIGHT

SOFT WHISPER IDI

30 STUDIO ( SPEECH)

20 STUDIO FOR SOUND PICTURES

10

THRESHOLD OF HEARING
YOUTHS WOO 4000 Ht o
OPE RAT OP S POSIT ON ...INN

Figure 2-1. Typical A- weighted sound levels measured with a sound-level meter. These
values are taken from the literature. Sound-level measurements give only part of the
information usually necessary to handle noise problems, and are often supplemented by
analysis of the noise spectra.

4
 level in alarge office usually is between 50 and 60 decibels. Among the very
loud sounds are those produced by nearby airplanes, railroad trains, riveting
machines, thunder, and so on, which are in the range near 100 decibels. These
typical values should help the newcomer to develop afeeling for this term
"decibel" as applied to sound level.
For some purposes it is not essential to know more about decibels than the
above general statements. But when we need to modify or to manipulate the
measured decibels, it is desirable to know more specifically what the term
means. There is then less danger of misusing the measured values. From a
strictly technical standpoint, the decibel is a logarithm of a ratio of two
values of power, and equal changes in decibels represent equal ratios.
Although we shall use decibels for giving the results of power-level calcula-
tions, the decibel is most often used in acoustics for expressing the sound-
pressure level and the sound level. These are extensions of the original use of
the term, and all three expressions will be discussed in the following sections.
First, however, it is worthwhile to notice that the above quantities include
the word "level." Whenever level is included in the name of the quantity, it
can be expected that the value of this level will be given in decibels or in some
related term and that areference power, pressure, or other quantity is stated
or implied.

2.3 POWER LEVEL.

Because the range of acoustic powers that are of interest in noise measure-
ments is about one-billion-billion to one ( 10 18 : 1), it is convenient to relate
these powers on the decibel scale, which is logarithmic. The correspondingly
smaller range of numerical values is easier to use and, at the same time, some
calculations are simplified.
The decibel scale can be used for expressing the ratio between any two
powers; and tables for converting from a power ratio to decibels and vice
versa are given in Appendix Iof this book. For example, if one power is four
times another, the number of decibels is 6; if one power is 10,000 times
another, the number is 40 decibels.
It is also convenient to express the power as apower level with respect to a
reference power. Throughout this book the reference power will be 10 -12
watt. Then the power level ( PWL) is defined as

W
PWL = 10 log dB re 10 -12 watt
10 -12

where W is the acoustic power in watts, the logarithm is to the base 10, and re
means referred to. This power level is conveniently computed from

PWL = 10 log W + 120

since 10 -1 2 as apower ratio corresponds to — 120 dB. The quantity 10 log W,

which is the number of decibels corresponding to the numerical value of
watts, can be readily obtained from the decibel tables in the Appendix. For
example, 0.02 watt corresponds to apower level of

—17 + 120 = 103 dB.

5
 ACOUSTIC POWER
POWER POWER LEVEL
SOURCE
(WATTS , ,OB RE 10 - 12 WATT I

25 TO 40 MILLION 195 SATURN ROCKET

100,000 170
RAM JET
TURBO- JET ENGINE WITH AFTERBURNER

10,000
160 TURBO- JET ENGINE, 3000- LB THRUST

1,000
150
4- PROPELLER AIRLINER

100
140

75- PIECE ORCHESTRA } PEAK RUS LEVELS IN

10
130 PIPE ORGAN I 8- SECOND INTERVALS

SMALL AIRCRAFT ENGINE

120 LARGE CHIPPING HAMMER

J
>
1
PIANO
PEAK RMS LEVELS IN
TIIP 8- SECOND INTERVALS
0.1 110 BLARING RADIO

CENTRIFUGAL VENTILATING FAN ( 13,000 CFM1

0.01
100 4'LOOM
AUTO ON HIGHWAY

0.021 90 VANEAXIAL VENTILATING FAN ( 1506 CFM)

VOICE - SHOUTING ( AVERAGE LONG-TIME RUS)

0.0001 80

0.00001 70 VOICE - CONVERSATIONAL LEVEL

(AVERAGE LONG-TIME RMS)

0 «10001
60

0 0000001 50

0 000,000,01
40

0.0O0,000, 001 30 VOICE - VERY SOFT WHISPER

Figure 2-2. Typical power levels for various acoustic sources. These levels bear no simple
relation to the sound levels of Figure 2-1.

6
 Some typical power levels for various acoustic sources are shown in Figure
2-2.
No instrument for directly measuring the power level of a source is
available. Power levels can be computed from sound-pressure measurements.
2.4 SOUND-PRESSURE LEVEL.
It is also convenient to use the decibel scale to express the ratio between
any two sound pressures; tables for converting from a pressure ratio to
decibels and vice versa are given in the Appendix. Since sound pressure is
usually proportional to the square root of the sound power, the sound-pres-
sure ratio for agiven number of decibels is the square root of the correspond-
ing power ratio. For example, if one sound pressure is twice another, the
number of decibels is 6; if one sound pressure is 100 times another, the
number is 40 decibels.
The sound pressure can also be expressed as asound-pressure level with
respect to a reference sound pressure. For airborne sounds this reference
sound pressure is generally 20 pN/m 2 .For some purposes areference pressure
of one microbar (0.1 N/m 2 )has been used, but throughout this book the
value of 20 pN/m 2 will always be used as the reference for sound-pressure
level. Then the definition of sound-pressure level ( SPL) is

SPL = 20 log 00002 dB re 20 micronewtons/m et

er square d
.
where P is the root-mean-square sound pressure in newtons/meter squared for
the sound in question. For example, if the sound pressure is 1N/m 2 ,then the
corresponding sound-pressure ratio is

1
.00002 or 50000.
From the tables, we find that the pressure level is 94 dB re 20 iiN/m 2 .If
decibel tables are not available, the level can, of course, be determined from a
table of logarithms.
The instrument used to measure sound-pressure level consists of amicro-
phone, attenuator, amplifier, and indicating meter. This instrument must have
an over-all response that is uniform (" flat") as afunction of frequency, and
the instrument is calibrated in decibels according to the above equation.
The position of the selector switch of the instrument for this measurement
is often called " FLAT" or " 20-kHz" to indicate the wide frequency range
that is covered. The result of a measurement of this type is also called
"over-all sou nd -pressure level."

2.5 SOUND LEVEL.

The apparent loudness that we attribute to asound varies not only with
the sound pressure but also with the frequency ( or pitch) of the sound. In
addition, the way it varies with frequency depends on the sound pressure. If
this effect is taken into account to some extent for pure tones, by "weight-
ing" networks included in an instrument designed to measure sound-pressure
level, then the instrument is called asound-level meter. In order to assist in
obtaining reasonable uniformity among different instruments of this type, the
American National Standards Institute (formerly, USA Standards Institute
and American Standards Association), in collaboration with scientific and
engineering societies, has established astandard to which sound-level meters
should conform.

7
 The current American National Standard Specification for Sound-Level
Meters ( ANSI S1.4-1971) requires that three alternate frequency-response
characteristics be provided in instruments designed for general use (see Figure
2-3)*. These three responses are obtained by weighting networks designated
as A, B, and C. Responses A, B, and C selectively discriminate against low and
high frequencies in accordance with certain equal-loudness contours, which
will be described in alater section.

1.5
A
o

5 8 AND C
RESPONSE - DECIBELS

10

15

FREQUENCY RESPONSES
20
FOR SLM WEIGHTING CHARACTERISTICS
25

30
A
RELATIVE

35

40

45

5020 50 100 200 500 1000 2000 5000 10,000

FREQUENCY- Hz

Figure 2-3. Frequency-response characteristics in the American National Standard Spec-

ification for Sound-Level Meters, ANSI-S1.4-1971.

Whenever one of these networks is used, the reading obtained should be

described as in the following examples: the "A-weighted sound level is 45
dB" "sound level ( A) = 45 dB," or "SLA = 45 dB". In a table, the
abbreviated form " LA "with the unit " dB" is suggested, or where exceptional
compactness is necessary, "dB(A)." The form " dBA" has also been used, but
this notation implies that anew unit has been introduced and is therefore not
recommended. Note that when aweighting characteristic is used, the reading
obtained is said to be the "sound level.**" Only when the over-all frequency
response of the instrument is flat are sound-pressure levels measured. Since
the reading obtained depends on the weighting characteristic used, the charac-
teristic that was used must be specified or the recorded level may be useless.
A common practice is to assume A-weighting if not otherwise specified.
It is often recommended that readings on all noises be taken with all three
weighting positions. The three readings provide some indication of the fre-
quency distribution of the noise. If the level is essentially the same on all
three networks, the sound probably predominates in frequencies above 600
Hz. If the level is greater on the C network than on the A and B networks by
several decibels, much of the noise is probably below 600 Hz.
In the measurement of the noise produced by distribution and power
transformers, the difference in readings of level with C-weighting and A-

•The current international standards and most national standards on sound-level meters
specify these same three responses.
**It was customary, if a single sound-level reading was desired, to select the weighting
position according to level, as follows: for levels below 55 dB, A weighting, for levels
from 55 dB to 85 dB, B weighting; and for levels above 85 dB. C weighting. Now,
however, the A-weighted sound level is the one most widely used regardless of level. See
paragraph 3.16.3.

8
weighting networks ( Lc- LA )is frequently noted. (This difference in decibels
is called the " harmonic index" in that application only.) It serves, as indi-
cated above, to give some idea of the frequency distribution of the noise. This
difference is also used in other noise-rating techniques in conjunction with
the A-weighted sound level.

2.6 COMBINING DECIBELS.

A number of possible situations require the combining of several noise
levels stated in decibels. For example, we may want to predict the effect of
adding a noisy machine in an office where there is already asignificant noise
level, to correct a noise measurement for some existing background noise, to
predict the combined noise level of several different noise sources, or to
obtain acombined total of several levels in different frequency bands.
In none of these situations should the numbers of decibels be added
directly. The method that is usually correct is to combine them on an energy
basis. The procedure for doing this is to convert the numbers of decibels to
relative powers, to add or subtract them, as the situation may require, and
then convert back to the corresponding decibels. By this procedure it is easy
to see that a noise level of 80 decibels combined with a noise level of 80
decibels yields 83 decibels and not 160 dB. A table showing the relation
between power ratio and decibels appears in Appendix I. A chart for combin-
ing or subtracting different decibel levels is shown in Appendix II.
The single line chart of Figure 24 is particularly convenient for adding
noise levels. For example, anoisy factory space has apresent A-weighted level
at a given location of 82 dB. Another machine is to be added 5 feet away.
Assume it's known from measurements on the machine, that at that location
in that space, it alone will produce an A-weighted level of about 78 dB. What
will the over-all level be when it is added? The difference in levels is 4 dB. If
this value is entered on the line chart, one finds that 1.5 dB should be added
to the higher level to yield 83.5 dB as the resultant level.

2.7 VIBRATION.
Vibration is the term used to describe continuing or steady-state periodic
motion. The motion may be simple harmonic motion like that of a pendu-
lum, or it may be complex like aride in the " whip" at an amusement park.

N
'------„--„
2 3 4 5 6 7 8 9 10 II 12 13 14 15
DIFFERENCE IN DECIBELS BETWEEN TWO LEVELS
BEING ADDED

Figure 2-4. Chart for combining noise levels.

9
The motion may involve tiny air particles that produce sound when the rate
of vibration is in the audible frequency range ( 20 to 20,000 Hz), or it may
involve, wholly or in part, structures found in machinery, bridges, or battle-
ships. Usually the word vibration is used to describe motions of the latter
types, and is classed as solid-borne, or mechanical, vibration.
Many important mechanical vibrations lie in the frequency range of 1to
2,000 Hz ( corresponding to rotational speeds of 60 to 120,000 rpm). In some
specialized fields, however, both lower and higher frequencies are important.
For example, in seismological work, vibration studies may extend down to a
small fraction of a Hz, while in loudspeaker-cone design, vibrations up to
20,000 Hz must be studied.

2.7.1 Nature of Vibratory Motion. Vibration problems occur in so many

devices and operations that alisting of these would be impractical. Rather, we
shall give aclassification on the basis of the vibratory motion, together with
numerous examples of where that motion occurs, to show the practical
application. The classes of vibratory motion that have been selected are given
in Table 2-1. They are not mutually exclusive and, furthermore, most devices
and operations involve more than one class of vibratory motion.

Table 21
NATURE OF VIBRATORY MOTION

Torsional or twisting vibration Flexural and plate- mode Belts

vibration Punch presses
Examples:
Examples. Tamping machines
Reciprocating devices
Gasoline and diesel engines Aircraft Intermittent vibration
Valves Circular saws (mechanical shock)
Compressors Loudspeaker cones
Examples:
Pumps Sounding boards
Blasting
Rotating devices Ship hulls and decks
Gun shots
Electric motors Turbine blades
Earthquakes
Fans Gears
Drop forges
Turbines Bridges
Heels impacting floors
Gears Floors
Typewriters
Turntables Walls
Ratchets
Pulleys Geneva mechanisms
Propellers Translational, axial, or Stepping motors
rigid- body vibration
Autos
Bending vibration Examples: Catapults
Examples. Reciprocating devices Planers
Shafts in motors, engines Gasoline and diesel engines Shapers
String instruments Compressors Chipping hammers
Springs Air hammers Riveters
Belts Tamping machines Impact wrenches
Chains Shakers
Tape in recorders Punch presses Random and
Autos miscellaneous motions
Pipes
Bridges Motors Examples:
Propellers Devices on vibration mounts Combustion
Transmission lines Ocean waves
Aircraft wings Extensional and shear vibration Tides
Reeds on reed instruments Examples: Tumblers
Rails Transformer hum Turbulence
Washing machines Earthquakes
Hum in electric motors
Gas and fluid motion
and generators
and their interaction
Moving tapes
with mechanisms

10
 2.7.2 Vibration Terms. Vibration can be measured in terms of displace-
ment, velocity, acceleration and jerk. The easiest measurement to understand
is that of displacement, or the magnitude of motion of the body being
studied. When the rate of motion (frequency of vibration) is low enough, the
displacement can be measured directly with the dial-gauge micrometer. When
the motion of the body is great enough, its displacement can be measured
with the common scale.
In its simplest case, displacement may be considered as simple harmonic
motion, like that of the bob of a pendulum, that is, asinusoidal function
having the form

X = A sin cot (1)

where A is a constant, co is 2/r times the frequency, and tis the time, as
shown in Figure 2-5. The maximum peak-to-peak displacement, also called
double amplitude, (a quantity indicated by a dial gauge) is 2A, and the
root-mean-square (rms) displacement is A/Vi(=0.707A). The average (full-
wave rectified average) value of the displacement is 2A//r (=0.636A), while
the "average double amplitude" (a term occasionally encountered) would be
4A//r (= 1.272A). Displacement measurements are significant in the study of
deformation and bending of structures.

Figure 2-5. A simple sinusoidal function.

When a pure tone is propagated in air, the air particles oscillate about their
normal position in a sinusoidal fashion. We could then think of sound in
terms of the instantaneous particle displacement and specify its peak and rms
value. But these displacements are so very small that they are very difficult to
measure directly.
In many practical problems displaCement is not the important property of
the vibration. A vibrating mechanical part will radiate sound in much the
same way as does a loudspeaker. In general, velocities of the radiating part
(which corresponds to the cone of the loudspeaker) and the air next to it will
be the same, and if the distance from the front of the part to the back is large
compared with one-half the wavelength of the sound in air, the actual sound
pressure in air will be porportional to the velocity of the vibration. The sound
energy radiated by the vibrating surface is the product of the velocity squared
and the resistive component of the air load. Under these conditions it is the
velocity of the vibrating part and not its displacement that is of greater
importance.
Velocity has also been shown by practical experience to be the best single
criterion for use in preventive maintenance of rotating machinery. Peak-to-
peak displacement has been widely used for this purpose, but then the

11
amplitude selected as adesirable upper limit varies markedly with rotational
speed.
Velocity is the time rate of change of displacement, so that for the
sinusoidal vibration of equation ( 1) the velocity is:

= co A cos cot ( 2)

Thus velocity is proportional to displacement and to frequency of vibration.

The analogy cited above covers the case where a loudspeaker cone or
baffle is large compared with the wavelength of the sound involved. In most
machines this relation does not hold, since relatively small parts are vibrating
at relatively low frequencies. This situation can be compared to a small
loudspeaker without abaffle. At low frequencies the air may be pumped back
and forth from one side of the cone to the other with ahigh velocity, but
without building up much of a pressure or radiating much sound energy
because of the very low air load, which has areactive mechanical impedance.
Under these conditions an acceleration measurement provides abetter meas-
ure of the amount of noise radiated than does avelocity measurement.
In many cases of mechanical vibration, and especially where mechanical
failure is a consideration, the actual forces set up in the vibrating parts are
important factors. The acceleration of a given mass is proportional to the
applied force, and a reacting force equal but opposite in direction results.
Members of a vibrating structure, therefore, exert forces on the total struc-
ture that are afunction of the masses and the accelerations of the vibrating
parts. For this reason, acceleration measurements are important when vibra-
tions are severe enough to cause actual mechanical failure.
Acceleration is the time rate of change of velocity, so that for asinusoidal
vibration,

a = sin cot ( 3)

It is proportional to the displacement and to the square of the frequency or

the velocity and the frequency.
Jerk is the time rate of change of acceleration. At low frequencies this
change is related to riding comfort of autos and elevators and to bodily
injury. It is also important for determining load tiedown in planes, trains, and
trucks.

2.7.3 Acceleration and Velocity Level Some use is now being made of
"acceleration level" and "velocity level," which, as the names imply, express
the acceleration and velocity in decibels with respect to areference accelera-
tion and velocity. The reference value of 10-8 m/s ( 10-6 cm/s) for velocity
and 10-5 m/s 2 ( 10-3 cm/s 2 )for acceleration are now used, although other
references have been proposed.

2.7.4 Nonsinusoidal Vibrations. Equations ( 1), ( 2), and ( 3) represent only

sinusoidal vibrations but, as with other complex waves, complex periodic
vibrations can also be represented as acombination of sinusoidal vibrations
often called aFourier series. The simple equations may, therefore, be expand-
ed to include as many terms as desirable in order to express any particular
type of vibration. For agiven sinusoidal displacement, velocity is proportion-

12
 al to frequency and acceleration is proportional to the square of the frequen-
cy, so that the higher-frequency components in a vibration are progressively
more important in velocity and acceleration measurements than in displace-
ment readings.

2.8 SUMMARY.
2.8.1 Sound. Reference quantities (ANSI S1.8-1969) and relations pre-
sented in this chapter included the following:
Reference sound pressure: 20 micronewtons/square meter ( 20µN/m 2 )*
Reference power: 10-12 watt.**

Power level PWL = 10 log dB re 10 -12 watt.

10-12
where W is the acoustic power in watts.

Sound pressure level: SPL = 20 log dB re 20 el/m 2

.00002

where Pis the root-mean-square sound pressure in newtons/square meter.

(Logarithms are taken to the base 10 in both PWL and SPL calculations.)
Important concepts that aid in interpreting noise measurement results can
be summarized as follows:
To measure sound level, use asound-level meter with one or more of its
frequency-response weightings ( A, B, and C).
To measure sound-pressure level, use asound-level meter with the controls
set for as uniform afrequency response as possible.
Decibels are usually combined on an energy basis, not added directly.
Speed of sound in air:
at 0°C is 1087 ft/s or 331.4 m/s
at 20 °C is 1127 ft/s or 343.4 m/s
Pressure Level
Pressure re 20 µN1m 2
1Newton/m 2 94 dB
1microbar 74 dB
1pound/ft. 2 127.6 dB
1pound/in. 2 170.8 dB
1atmosphere 194.1 dB

NOTE: The reference pressure and the reference power have been selected
independently because they are not uniquely related.
2.8.2 Vibration. Displacement is magnitude of the motion.
Velocity is the time rate of change of displacement.
Acceleration is the time rate of change of velocity.
Jerk is the time rate of change of acceleration.
Reference quantities:

Velocity: 10-8 meters/second ( 10-8 cm/s)

Acceleration: 10-5 meters/second/second (l0 CM/S 2 )

*At one time the reference for a sound-level meter was taken as 161 2 watt/square
meter. For most practical purposes, this reference is equivalent to the presently used
pressure. This earlier reference value is not a reference for power, since it is power
divided by an area. The pressure 20 phl/m s is also expressed as 2 x 10' s newton/square
meter, 0.0002 microbar, or 0.0092 dyne/cms.
•"A reference power of 10'" watt is also used in the USA, and has been used in very
early editions of this handbook, bu t th e re f
erence power of 10 -1 ' watt is preferred
(ANSI SI.8-1969).

13
tEFERENCES

itandards
‘NSI S1.1-1960, Acoustical Terminology.
‘NSI S1.8-1969, Preferred Reference Quantities for Acoustical Levels.
‘NSI S1.4-1971, Sound-Level Meters.
‘NSI S1.13-1971, Methods for the Measurement of Sound Pressure Levels.

Dther
L. Beranek ( 1949), Acoustic Measurements, John Wiley & Sons, Inc., New York.
-. L. Beranek ( 1954), Acoustics, McGraw-Hill Book Company, Inc., New York.
L. Beranek, ed. ( 1971), Noise and Vibration Control, McGraw-Hill Book Company,
Inc., New York.
1E.D. Bishop ( 1965), Vibration, Cambridge University Press, Cambridge, England.
Z. E. Crede ( 1951), Vibration and Shock Isolation, John Wiley 8c Sons, Inc., New York.
f. P. Den Hartog ( 1956), Mechanical Vibrations, McGraw-Hill Book Company Inc., New
York.
Z. M. Harris, ed. ( 1957), Handbook of Noise Control, McGraw-Hill Book Company, Inc.,
New York.
2. M. Harris and C. E. Crede, ed. ( 1961), Shock and Vibration Handbook, McGraw-Hill
Book Company, Inc., New York.
L. E. Kinsler and A. R. Frey ( 1962), Fundamentals of Acoustics, John Wiley & Sons,
Inc., New York.
S. T. Morrow ( 1963), Shock and Vibration Engineering, John Wiley & Sons, Inc., New
York.
?. M. Morse and K. U. lngard ( 1968), Theoretical Acoustics, McGraw-Hill Book Com-
pany, Inc., New York. ( A graduate text.)
Fl. F. Olson ( 1957), Acoustical Engineering, D. Van Nostrand Company, Inc., New York,
3rd Edition.
I. R. Pierce and E. E. David, Jr.,(1958), Man's World of Sound, Doubleday 8c Company,
Inc., Garden City, New York.
R. H. Randall ( 1951), An Introduction to Acoustics, Addison-Wesley Press, Cambridge,
Mass.
R. W. B. Stephens and A. E. Bate ( 1966), Acoustics and Vibrational Physics, Arnold,
London, St. Martin's Press, New York, 2nd Edition.
G. W. Swenson, Jr. ( 1953), Principles of Modern Acoustics, D. Van Nostrand Company,
Inc., New York.
G. W. Van Santen ( 1953), Mechanical Vibration, Elsevier, Houston.
W. Wilson ( 1959), Vibration Engineering, Charles Griffin, London.
L. F. Yerges ( 1969), Sound, Noise and Vibration Control, Van Nostrand-Reinhold
Company, New York.
R. W. Young ( 1955), "
A Brief Guide to Noise Measurement and Analysis," Research and
Development Report 609, 16 May 1955, US Navy Electronics Laboratory,
PB118036.

14
 Chapter 3

What Noise and Vibration

Do and How Much Is
Acceptable
3.1 WHY WE MEASURE NOISE.
\That very intense noise may cause hearing loss, that we are annoyed by a
noisy device and anoisy environment, or that noise may interfere with our
sleep, our work, and our recreation is frequently the basic fact that leads to
noise measurements and attempts at quietinèln order to make the most
significant measurements and to do the job of quieting most efficiently, it is
clearly necessary to learn about these effects of noise. We seek to estimate
from these effects what levels of noise are acceptable, and thus establish
suitable noise criteria. Then if we measure the existing noise level, the
difference between this level and the acceptable level is the noise reduction
necessary.
Unfortunately, not all the factors involved in annoyance, interference, and
hearing loss are known at present. Nor are we yet sure how the known factors
can best be used. But abrief discussion of our reactions to sounds will serve
to show some of the factors and their relative significance. This information
will be useful as aguide for selecting electronic equipment to make the most
significant measurements for the problem at hand.

3.2 PSYCHOACOUSTICAL EXPERIMENTS.

Scientists and engineers have investigated many aspects of man's reactions
to sounds ( Stevens, 1951). For example, they have measured the levels of the
weakest sounds that various observers could just hear in avery quiet room
(threshold of hearing), they have measured the levels of the sounds that are
sufficiently high in level to cause pain ( threshold of pain), and they have
measured the least change in level and in frequency that various observers
could detect (differential threshold). These experimenters have also asked
various observers to set the levels of some sounds so that they are judged
equal in loudness to reference sounds (equal loudness), and they have asked
the observers to rate sounds for loudness on anumerical scale.
In order to get reliable measures of these reactions, the experimenters have
to simplify the conditions under which people react to sounds. This simplifi-
cation is mainly one of maintaining unchanged as many conditions as possible
while a relatively few characteristics of the sound are varied. Some of the
conditions that have to be controlled and specified are the following: the
physical environment of the observer, particularly the background or ambient
level; the method of presenting the changing signals, including the order of
presentation, duration, frequency, and intensity; the selection of the observ-
ers; the instructions to the observers; the experience of the observers in the
specific test procedure; the normal hearing characteristics of the observers ;
the responses; and the method of handling the data.

15
 Variations in the conditions of the measurement will affect the result.
Such interaction is the reason for requiring controlled and specified condi-
tions. It is desirable to know, however, how much the various conditions do
affect the result. For example, small changes in room temperature are usually
of little significance. But if the observer is exposed to a noise of even
moderate level just before athreshold measurement, the measured threshold
level will, temporarily, be significantly higher than normal.
The basic method used by the observer to present his reaction to the
signals is also important in the end result. Numerous methods have been
developed for this presentation. Three of these psychophysical methods are as
follows:
1. In the method of adjustment, the observer sets an adjustable control to
the level he judges suitable for the test.
2. In the method of the just-noticeable difference, the observer states
when two signals differ sufficiently, so that he can tell they are different.
3. In the method of constant stimuli, the observer states whether two
signals are the same, or which is the greater, if they seem to differ.
The approach an observer takes in making adecision is significant. If an
observer attempts to detect asignal that is sometimes present in abackground
of noise, four possible conditions exist. With the signal present or absent, he
may respond that it is or is not present. The choice he makes can be
influenced by the instructions. On the one hand, he may be told that false
alarms are serious errors and that he should respond that the signal is present
only if he is very certain of it. Or he may be told that occasional false alarms
are unimportant. These different instructions will produce different ap-
proaches to the decision problem and will affect the results of the experi-
ment. These factors have been organized in modern detection theory (Green
and Swets, 1966) to permit aquantitative approach to such psychoacoustic
problems by the use of a " receiver-operating characteristic," usually called
"ROC." Experiments based on this theory have also shown that earlier
concepts of a "threshold" are oversimplified. We shall, however, use the term
threshold here without attempting to define it accurately, since it is areadily
accepted concept, and it is adequate for the present discussion.
When psychoacoustical experiments are performed, the resultant data
show variability in the judgments of agiven observer as well as variability in
the judgments of a group of observers. The data must then be handled by
statistical methods, to obtain an average result as well as ameasure of the
deviations from the average. In general, it is the average result that is of most
interest but the extent of the deviations is also of value, and in some
experiments these deviations are of major interest.
The deviations are not usually shown on graphs of averaged psychoacous-
tical data, but they should be kept in mind. To picture these deviations one
might think of the curves as if they were drawn with awide brush instead of a
fine pen.
The measured psychoacoustical responses also have a certain degree of
stability, although it is not the degree of stability that we find in physical
measurements. In the normal course of events, if one's threshold of hearing is
measured today, a similar measurement tomorrow should give the same
threshold level within afew decibels.
In the process of standardizing the measurement conditions for the sake of
reliability and stability, the experiments have been controlled to the point

16
 3.5 RATING THE LOUDNESS OF A SOUND.
Many psychoacoustical experiments have been made in which listeners
have been asked to rate the loudness of a sound. As a result of these
experiments, involving all sorts of sounds in various arrangements, much has
been learned about the concept of loudness in laboratory situations. The way
in which the judgment of loudness is obtained seems to affect the results
sufficiently, however, so that we cannot reliably scale all the sounds of
everyday life on an absolute basis. In particular, it does not seem possible to
give anumerical value to the loudness ratio of two sounds and have this ratio
be reasonably independent of the conditions of comparison .</tdoes seem
possible, however, to rank asound with satisfactory reliability according to
its loudness> For example, if sound A is judged louder than sound B and if
sound B is judged louder than sound C, then, in general, sound A will also be
judged louder than sound C.
3.5.1 Equal- Loudness Contours. One step in the direction of rating the
loudness of asound has been to determine the sound-pressure levels of simple
tones of various frequencies that sound just as loud to an observer as a
1000-Hz tone of agiven sound-pressure level. The results of this determina-
tion by Robinson and Dadson based on the averages of many observations,
are given as equal-loudness contours in F.gure 3-3. The number on each curve

120
6 hail 20 LOUDNESS
EVEL (
PHOM /
il
ballimMIIIIIIiiin7qujneniira
usamirommunup%I. menu
mmancimaiiiiiiimsivzon.
"E l00

1.
o
w 80 IIIMSIIIMEHIliiiill niii
Mk\ IMP.talifflialiiill
rx

-I 60 11111MIMME1111M11 MEMEL'
IIIIIMMIKolliiill 50 'nullMillil
Lai
11111110INIMMillijill --nujOrg--
IIIIEWIMMiiiiiill
e 40
t
U)
U)

a_
MinilliiialMiiil 20 Nall
IIMIMMIllelllini cMae' 7,1
rAl 20

MEMII IICtide
amain
20
II100 1000 5000 10,000

!!!ri
FREQUENCY IN CYCLES PER SECOND ( Hz

~2 "''e n e?;!*37e MfXaren-aElFee.11:;>.t mik,,oef

Figure 3-3. Free-field equal-loudness contours for pure tones (observer facing source),
determined by Robinson and Dadson 1956 at the National Physical Laboratory, Ted-
dington, England. (ISO/R226-1961) Piano keyboard helps identify the frequency scale.
Only the fundamental frequency of each piano key is indicated.

21
is the sound-pressure level of the 1000-Hz tone used for comparison for that
curve. To use the contours for determining the equally loud levels at other
frequencies, we find the point on the curve corresponding to the desired
frequency and read off the corresponding sound-pressure level as the ordi-
nate. For example, the 60-dB contour line shows that a67-dB level at 100 Hz
is just as loud as a60-dB 1000-Hz tone. We can also interpolate to find that a
60-dB 100-Hz tone is equal in loudness to a 51-dB 1000-Hz tone.(The
corresponding sound-pressure level in dB for the 1000-Hz tone has been
defined as the loundess level in eons. Therefore, a 100 -Hz tone at a
sound-pressure level of 60 dB has aloudness level of 51 phons.
The weighting networks for the sta ndard sound-level me er are based on
similar contours developed much earlier by Fletcher and Munson ( 1933)<the
"A" and " B" weighting characteristics are in accordance with the 40- and
70-phon Fletcher-Munson contours, but with modifications to take into
account the usually random nature of the sound field in a room and to
simplify their simulation with electrical networks>
A set of equal-loudness contours (Pollack, 1952) for bands of random
noise is shown in Figure 3-4. Random noise is acommon type of noise that
occurs in ventilating systems, jets, blowers, combustion chambers, etc. It does
not have a well-defined pitch, such as characterizes atone with the energy
concentrated in components of definite frequencies. Rather, random noise
has energy distributed over a band of frequencies. If the noise energy is
uniform over a wide range, it is called " white noise," being analogous in
spectrum characteristics to white light. When the energy is distributed over a
very wide band, it is asort of " hishing" sound. When the broadband noise has
little energy at low frequencies, it is more of a hissing sound. When it is
concentrated in narrower bands, the sound takes on some aspects of pitch.
For example, low-frequency random noise may be asort of roar.
The contours shown in Figure 3-4 are for relatively narrow bands ooisil,
such that 11 bands cover the range from 60 to 5800 Hz. The are distributed
uniformly on ascale of pitch for simple tones (see 3.15.2 . The numbers on
the curves are phons, that is, the sound-pressure levels of equally loud
1000-Hz tones, and the levels are plotted according to the centers of the
bandFor example, one band covers the range from 350 to 700 Hz. From
the curves we can see that when the sound-pressure level of the noise in that
band is 43 dB re 20 µN/m 2 ,the indicated loudness level is about 34 phons.

140

,zo

Figure 3-4. Equal-loud-

ness contours for rela-
tively narrow bands of
random noise. The cen- 20
ter frequency of the
band is shown as the o
5000
abscissa, and the num- 100 500 1000

bers on the curves are FREQUENCY IN Hz MOS. 1,11

phons (Pollack, 1952).

22
 mu

7
50

20

10

e
2
.,..s,
.2".
,

1 . 0

o2

Figure 3-5. Loudness vs

sound-pressure level for a
20 40 50 80 00
pure tone of 1000 Hz.
SOUND PRESSURE LEVEL- DECIBELS re 20µN/m 2

3.5.2 Loudness and Loudness Level. Although we may remark that some
sounds are louder than others, we do not ordinarily rate sounds for loudness
on anumerical basis. Experimenters have asked observers to make judgments
of the loudness ratio of sounds, that is, to state when one sound is twice, four
times, one-half, etc., as loud as another. The resultant judgments depend to a
considerable extent on how the problem is presented to the observer. But on
the basis of such judgments<several scales of loudness have been devised,
which rate sounds from "soft" to "loud" in units of soneAs areference, the
loudness of a1000-Hz tone with asound-pressure level of 40 dB re 201.41•1/m 2
(a loudness level of 40 phons) is taken to be 1sonè)(A tone that sounds twice
as loud has aloudness of 2 sonej This scale is shown on the vertical axis of
Figure 3-5, and the horizontal scale is the sound-pressure level of the sound in
decibels. The curve shown in this figure relates the loudness in sones to the
sound-pressure level for a 1000-Hz simple tone. This relation was developed
as auseful engineering approximation by Stevens as aresult of his analysis of
the data reported by many experimenters, who used a wide variety of
techniques. He also performed aseries of experiments in which the loudness
estimates were made on an unusually direct basis, and these experiments
confirmed the relation shown. Robinson has also suggested this relation,
which is published as a Recommendation of the International Standards
Organization ( ISO R131-1959).
Incidentally, the relation shown in Figure 3-5 tends to refute the point of
view that the decibel is used in acoustics because we respond to sound
pressure in a logarithmic manner.KActually, the loudness is approximately
proportional to the sound pressure raised to the 0.6 power

23
them here. But to remind one that aging and noise are only two of the many
possible factors, here are some of the more obvious contributors — congenital
defects, anatomical injuries, and disease.

3.4 "WHAT NOISE ANNOYS AN OYSTER?"

No adequate measures of the annoyance levels of noises have yet been
devised. Various aspects of the problem have been investigated, but the
psychological difficulties in making these investigations are very great. For
example, the extent of our annoyance depends greatly on what we are trying
to do at the moment, it depends on our previous conditioning, and it depends
on the character of the noise.
The annoyance level of anoise is sometimes assumed to be related directly
to the loudness level of the noise. Although not completely justifiable, this
assumption is sometimes helpful because a loud sound is usually more
annoying than one of similar character that is not so loud.
This approach is one of the reasons that many experiments have been
made on judged loudness of various sounds, and procedures have been
developed for predicting the loudness of noise from physical measurements.
Some of the results of these experiments will be reviewed in the next section.
In addition other experiments have been made in which listeners have been
asked to judge noises for their "noisiness," " unacceptability," " objectionabil-
ity," " annoyingness," or on how " disturbing" they were. Some of these
experiments have led to the concept of " perceived noise level" and " noisi-
ness." Since perceived noise level has been used widely, it too is described in
more detail below.
In acomprehensive review of such experiments Stevens ( 1972) shows how
remarkably similar most of the results are. One is led to the conclusions that
these distinct terms do not produce really significant differences in judgment
at least for the controlled experimental conditions. He has used the available
evidence to produce a new, but related, procedure for predicting the " per-
ceived level" and the " perceived magnitude" from physical measurements of
anoise. This procedure, called "Mark VII," is also described below.
<One conclusion that can be drawn from these experiments is that high-fre-
quency sounds (in the vicinity of 5000 Hz) are usually louder, more annoying
or disturbing than are lower-frequency sounds of the same sound-pressure
levelTherefore, when it is determined, by methods to be explained later,
that a significant portion of the noise is in this higher-frequency region,
considerable effort at reducing these levels from the viewpoint of annoyance
may be justified.
A rather different effect that may determine some of the annoying quality
of asound concerns its localization. When alarge office has acoustically hard
walls, floor, and ceiling, the room is "live," reverberant. The noise from any
office machinery then is reflected back and forth, and the workers are
immersed in the noise with the feeling that it comes from everywherel'\If the
office is heavily treated with absorbing material, the reflected sound is
reduced, and the workers then feel that the noise is coming directly from the
machine. This localized noise seems to be less annoying While no adequate
measures of this effect have been developed, the general principle discussed
here seems to be accepted by many who are experienced in noise problems.

20
 3.5 RATING THE LOUDNESS OF A SOUND.
Many psychoacoustical experiments have been made in which listeners
have been asked to rate the loudness of a sound. As a result of these
experiments, involving all sorts of sounds in various arrangements, much has
been learned about the concept of loudness in laboratory situations. The way
in which the judgment of loudness is obtained seems to affect the results
sufficiently, however, so that we cannot reliably scale all the sounds of
everyday life on an absolute basis. In particular, it does not seem possible to
give anumerical value to the loudness ratio of two sounds and have this ratio
be reasonably independent of the conditions of comparison .</tdoes seem
possible, however, to rank asound with satisfactory reliability according to
its loudness> For example, if sound A is judged louder than sound B and if
sound B is judged louder than sound C, then, in general, sound A will also be
judged louder than sound C.
3.5.1 Equal- Loudness Contours. One step in the direction of rating the
loudness of asound has been to determine the sound-pressure levels of simple
tones of various frequencies that sound just as loud to an observer as a
1000-Hz tone of agiven sound-pressure level. The results of this determina-
tion by Robinson and Dadson based on the averages of many observations,
are given as equal-loudness contours in F.gure 3-3. The number on each curve

120
6 hail 20 LOUDNESS
EVEL (
PHOM /
il
ballimMIIIIIIiiin7qujneniira
usamirommunup%I. menu
mmancimaiiiiiiimsivzon.
"E l00

1.
o
w 80 IIIMSIIIMEHIliiiill niii
Mk\ IMP.talifflialiiill
rx

-I 60 11111MIMME1111M11 MEMEL'
IIIIIMMIKolliiill 50 'nullMillil
Lai
11111110INIMMillijill --nujOrg--
IIIIEWIMMiiiiiill
e 40
t
U)
U)

a_
MinilliiialMiiil 20 Nall
IIMIMMIllelllini cMae' 7,1
rAl 20

MEMII IICtide
amain
20
II100 1000 5000 10,000

!!!ri
FREQUENCY IN CYCLES PER SECOND ( Hz

~2 "''e n e?;!*37e MfXaren-aElFee.11:;>.t mik,,oef

Figure 3-3. Free-field equal-loudness contours for pure tones (observer facing source),
determined by Robinson and Dadson 1956 at the National Physical Laboratory, Ted-
dington, England. (ISO/R226-1961) Piano keyboard helps identify the frequency scale.
Only the fundamental frequency of each piano key is indicated.

21
is the sound-pressure level of the 1000-Hz tone used for comparison for that
curve. To use the contours for determining the equally loud levels at other
frequencies, we find the point on the curve corresponding to the desired
frequency and read off the corresponding sound-pressure level as the ordi-
nate. For example, the 60-dB contour line shows that a67-dB level at 100 Hz
is just as loud as a60-dB 1000-Hz tone. We can also interpolate to find that a
60-dB 100-Hz tone is equal in loudness to a 51-dB 1000-Hz tone.(The
corresponding sound-pressure level in dB for the 1000-Hz tone has been
defined as the loundess level in eons. Therefore, a 100 -Hz tone at a
sound-pressure level of 60 dB has aloudness level of 51 phons.
The weighting networks for the sta ndard sound-level me er are based on
similar contours developed much earlier by Fletcher and Munson ( 1933)<the
"A" and " B" weighting characteristics are in accordance with the 40- and
70-phon Fletcher-Munson contours, but with modifications to take into
account the usually random nature of the sound field in a room and to
simplify their simulation with electrical networks>
A set of equal-loudness contours (Pollack, 1952) for bands of random
noise is shown in Figure 3-4. Random noise is acommon type of noise that
occurs in ventilating systems, jets, blowers, combustion chambers, etc. It does
not have a well-defined pitch, such as characterizes atone with the energy
concentrated in components of definite frequencies. Rather, random noise
has energy distributed over a band of frequencies. If the noise energy is
uniform over a wide range, it is called " white noise," being analogous in
spectrum characteristics to white light. When the energy is distributed over a
very wide band, it is asort of " hishing" sound. When the broadband noise has
little energy at low frequencies, it is more of a hissing sound. When it is
concentrated in narrower bands, the sound takes on some aspects of pitch.
For example, low-frequency random noise may be asort of roar.
The contours shown in Figure 3-4 are for relatively narrow bands ooisil,
such that 11 bands cover the range from 60 to 5800 Hz. The are distributed
uniformly on ascale of pitch for simple tones (see 3.15.2 . The numbers on
the curves are phons, that is, the sound-pressure levels of equally loud
1000-Hz tones, and the levels are plotted according to the centers of the
bandFor example, one band covers the range from 350 to 700 Hz. From
the curves we can see that when the sound-pressure level of the noise in that
band is 43 dB re 20 µN/m 2 ,the indicated loudness level is about 34 phons.

140

,zo

Figure 3-4. Equal-loud-

ness contours for rela-
tively narrow bands of
random noise. The cen- 20
ter frequency of the
band is shown as the o
5000
abscissa, and the num- 100 500 1000

bers on the curves are FREQUENCY IN Hz MOS. 1,11

phons (Pollack, 1952).

22
 mu

7
50

20

10

e
2
.,..s,
.2".
,

1 . 0

o2

Figure 3-5. Loudness vs

sound-pressure level for a
20 40 50 80 00
pure tone of 1000 Hz.
SOUND PRESSURE LEVEL- DECIBELS re 20µN/m 2

3.5.2 Loudness and Loudness Level. Although we may remark that some
sounds are louder than others, we do not ordinarily rate sounds for loudness
on anumerical basis. Experimenters have asked observers to make judgments
of the loudness ratio of sounds, that is, to state when one sound is twice, four
times, one-half, etc., as loud as another. The resultant judgments depend to a
considerable extent on how the problem is presented to the observer. But on
the basis of such judgments<several scales of loudness have been devised,
which rate sounds from "soft" to "loud" in units of soneAs areference, the
loudness of a1000-Hz tone with asound-pressure level of 40 dB re 201.41•1/m 2
(a loudness level of 40 phons) is taken to be 1sonè)(A tone that sounds twice
as loud has aloudness of 2 sonej This scale is shown on the vertical axis of
Figure 3-5, and the horizontal scale is the sound-pressure level of the sound in
decibels. The curve shown in this figure relates the loudness in sones to the
sound-pressure level for a 1000-Hz simple tone. This relation was developed
as auseful engineering approximation by Stevens as aresult of his analysis of
the data reported by many experimenters, who used a wide variety of
techniques. He also performed aseries of experiments in which the loudness
estimates were made on an unusually direct basis, and these experiments
confirmed the relation shown. Robinson has also suggested this relation,
which is published as a Recommendation of the International Standards
Organization ( ISO R131-1959).
Incidentally, the relation shown in Figure 3-5 tends to refute the point of
view that the decibel is used in acoustics because we respond to sound
pressure in a logarithmic manner.KActually, the loudness is approximately
proportional to the sound pressure raised to the 0.6 power

23
 3.5.3 Loudness- Level Calculations. If the sound to be measured is known
to be a simple tone, the procedure for determination of loudness level is
relatively easy. The sound-Pressure level and the frequency of the tone are
determined, and the equal-loudness contours of Figure 3-3 then indicate the
loudness level. Since the weighting networks on asound-level meter approxi-
mate two of the equal-loudness contours, a determination of the weighted
level (sound level) can be used to give an estimate of the loudness level of a

simple
For tone.
any other type of sound, however, the measured sound level will be
lower than the loudness levelKThe error in estimating loudness level will
depend on the type of sound and for many noises will be more than 10
phons,›For example, if we have auniform wide-band noise from 20 to 6000
Hz of 80-dB sound-Pressure level, the B-weighted sound level would be about
79 dB and the A-weighted sound level would be about 80 dB, whereas the
actual loudness level of such anoise is about 95 phons. Here we see that the
sound level is not only misleading, but is no nearer the loudness level than is
the sound-pressure level. This result, for most noises, illustrates the fact that
we need to know more about asound than just its sound-pressure level or its
sound level. If we know how the energy in asound is distributed as afunction
of frequency, we can make amore useful estimate of its probable subjective
effect than we can by knowing just its sound-pressure level. One of the ways
such knowledge is used is the calculation of loudness level.
<A number of workers in noise measurements have found it useful to
translate their noise measurements into such loudness terms. Then they can
say the measured sound is, for example, about equal in loudness to another,
more familiar, sound›To some groups, such as executive and lay clients, this
type of statement is seemingly more meaningful than levels quoted in deci-

bels.
For steady, wide-band noises, atechnique developed by Stevens has been
found to give good results. The sound is divided by an analyzer into frequen -
cy bands covering the audio spectrum. The loudness level is then calculated
according to the procedure given in the next section.
A set of 8 or 9 octave bands is most often used for this purpose. These
have center frequencies of 31.5, 63, 125, 250, 500, 1000, 2000, 4000 and
8000 Hz, with each band actually covering a2:1 frequency range. A more
detailed division provided by athird-octave analysis is also widely used. Both
of these band divisions are described in more detail in Chapter 5.

• 3.5.4 Procedure for Calculating Loudness* .able 3-1 is used to calculate

the loudness for octave-band levels of the preferred serie The procedure is

as follows:
Ti. From the table find the proper loudness index for each band level.
2. Add all the loudness indexes ( ES).
3. Multiply this sum by 0.3.
4. Add this product to 0.7 of the index for that band that has the largest
index (0.3 E S + 0.7 Smax ). This value is the total loudness in sones.
5. This total loudness is then converted to loudness level in phons by the
relation shown in the two columns at the right of the table.
'The method used here is that standardized in ANSI S3.4-1968 and originally given by
S. S. Stevens ( 1961). Chart paper No. 31460-A (Codex Book Company, Norwood,
Massachusetts 02062) is available for this calculation when the older series of octave

bands is used.

24
 Table 3-1.
BAND LEVEL CONVERSION TO LOUDNE
Band
Loudness
Level Band Loudness Index Loudness Lev‘..1
dB 31.5 63 125 250 500 1000 2000 4000 8000 Sones Phons
20 .18 .30 .45 .61 .25 20
21 .22 .35 .50 .67 .27 21
22 .07 .26 .40 .55 .73 .29 22
23 .12 .30 .45 .61 .80 .31 23
24 .16 .35 .50 .67 .87 .33 24
25 .21 .40 .55 .73 .94 .35 25
26 .26 .45 .61 .80 1.02 .38 26
27 .31 .50 .67 .87 1.10 .41 27
28 .07 .37 .55 .73 .94 1.18 .44 28
29 .12 .43 .61 .80 1.02 1.27 .47 29
30 .16 .49 .67 .87 1.10 1.35 .50 30
31 .21 .55 .73 .94 1.18 1.44 .54 31
32 .26 .61 .80 1.02 1.27 1.54 .57 32
33 .31 .67 .87 1.10 1.35 1.64 .62 33
34 .07 .37 .73 .94 1.18 1.44 1.75 .66 34
35 .12 .43 .80 1.02 1.27 1.54 1.87 .71 35
36 .16 .49 .87 1.10 1.35 1.64 1.99 .76 36
37 .21 .55 .94 1.18 1.44 1.75 2.11 .81 37
38 .26 .62 1.02 1.27 1.54 1.87 2.24 .87 38
39 .31 .69 1.10 1.35 1.64 1.99 2.38 .93 39
40 .07 .37 .77 1.18 1.44 1.75 2.11 2.53 1.00 40
41 .12 .43 .85 1.27 1.54 1.87 2.24 2.68 1.07 41
42 .16 .49 .94 1.35 1.64 1.99 2.38 2.84 1.15 42
43 .21 .55 1.04 1.44 1.75 2.11 2.53 3.0 1.23 43
44 .26 .62 1.13 1.54 1.87 2.24 2.68 3.2 1.32 44
45 .31 .69 1.23 1.64 1.99 2.38 2.84 3.4 1.41 45
46 .07 .37 .77 1.33 1.75 2.11 2.53 3.0 3.6 1.52 48
47 .12 .43 .85 1.44 1.87 2.24 2.68 3.2 3.8 1.62 47
48 .16 .49 .94 1.56 1.99 2.38 2.84 3.4 4.1 1.74 4H
49 .21 .55 1.04 1.69 2.11 2.53 3.0 3.6 4.3 1.87 49
50 .26 .62 1.13 1.82 2.24 2.68 3.2 3.8 4.6 2.00 sn
51 .31 .69 1.23 1.96 2.38 2.84 3.4 4.1 4.9 2.14 51
52 .37 .77 1.33 2.11 2.53 3.0 3.6 4.3 5.2 2.30 52
53 .43 .85 1.44 2.24 2.68 3.2 3.8 4.6 5.5 2.46 53
54 .49 .94 1.56 2.38 2.84 3.4 4.1 4.9 5.8 2.64 54
55 .55 1.04 1.69 2.53 3.0 3.6 4.3 5.2 6.2 2.33 55
56 .62 1.13 1.82 2.68 3.2 3.8 4.6 5.5 6.6 3.03 56
57 .69 1.23 1.96 2.84 3.4 4.1 4.9 5.8 7.0 3.25 57
58 .77 1.33 2.11 3.0 3.6 4.3 5.2 6.2 7.4 3.48 58
59 .85 1.44 2.27 3.2 3.8 4.6 5.5 6.6 7.8 3.73 59
60 .94 1.56 2.44 3.4 4.1 4.9 5.8 7.0 8.3 4.00 60
61 1.04 1.69 2.62 3.6 4.3 5.2 6.2 7.4 8.8 4.29 61
62 1.13 1.82 2.81 3.8 4.6 5.5 6.6 7.8 9.3 4.59 62
63 1.23_ 1.96 3.0 4.1 4.9 5.8 7.0 8.3 9.9 4.92 63
64 1.33 2.11 3.2 4.3 5.2 6.2 7.4 8.8 10.5 5.28 64
65 1.44 2.27 3.5 4.6 5.5 6.6 7.8 9.3 11.1 5.66 65
66 1.56 2.44 3.7 4.9 5.8 7.0 8.3 9.9 11.8 6.06 66
67 1.69 2.62 4.0 5.2 6.2 7.4 8.8 10.5 12.6 6.50 67
68 1.82 2.81 4.3 5.5 6.6 7.8 9.3 11.1 13.5 6.96 68
69 1.96 3.0 4.7 5.8 7.0 8.3 9.9 11.8 14.4 7.46 69
70 2.11 3.2 5.0 6.2 7.4 8.8 10.5 12.6 15.3 8.00 70
71 2.27 3.5 5.4 6.6 7.8 9.3 11.1 13.5 16.4 8.6 71
72 2.44 3.7 5.8 7.0 8.3 9.9 11.8 14.4 17.5 9.2 72
73 2.62 4.0 6.2 7.4 8.8 10.5 12.6 15.3 18.7 9.8 73
74 2.81 4.3 6.6 7.8 9.3 11.1 13.5 16.4 20.0 10.6 74
75 3.0 4.7 7.0 8.3 9.9 11.8 14.4 17.5 21.4 11.3 75
76 3.2 5.0 7.4 8.8 10.5 12.6 15.3 18.7 23.0 12.1 76
77 3.5 5.4 7.8 9.3 11.1 13.5 16.4 20.0 24.7 13.0 77
78 3.7 5.8 8.3 9.9 11.8 14.4 17.5 21.4 26.5 13.9 78
79 4.0 6.2 8.8 10.5 12.6 15.3 18.7 23.0 28.5 14.9 79
80 4.3 6.7 9.3 11.1 13.5 16.4 20.0 24.7 30.5 16.0 80
81 4.7 7.2 9.9 11.8 14.4 17.5 21.4 26.5 32.9 17.1 81
82 5.0 7.7 10.5 12.6 15.3 18.7 23.0 28.5 35.3 18.4 82
83 5.4 8.2 11.1 13.5 16.4 20.0 24.7 30.5 38 19.7 83
84 5.8 8.8 11.8 14.4 17.5 21.4 26.5 32.9 41 21.1 84
85 6.2 9.4 12.6 15.3 18.7 23.0 28.5 35.3 44 22.6 85

25
 Table 3-1 ( Continued)
Band
Loudness
Level Band Loudness Index Loudness Level
dB 31.5 63 125 250 500 1000 2000 4000 8000 Sones Phons
86 6.7 10.1 13.5 16.4 20.0 24.7 30.5 38 48 24.3 86
87 7.2 10.9 14.4 17.5 21.4 26.5 32.9 41 52 26.0 87
88 7.7 11.7 15.3 18.7 23.0 28.5 35.3 44 56 27.9 88
89 8.2 12.6 16.4 20.0 24.7 30.5 38 48 61 29.9 89
90 8.8 13.6 17.5 21.4 26.5 32.9 41 52 66 32.0 90
91 9.4 14.8 18.7 23.0 28.5 35.3 44 56 71 34.3 91
92 10.1 16.0 20.0 24.7 30.5 38 48 61 77 36.8 92
93 10.9 17.3 21.4 26.5 32.9 41 52 66 83 39.4 93
94 11.7 18.7 23.0 28.5 35.3 44 56 71 90 42.2 94
95 12.6 20.0 24.7 30.5 38 48 61 77 97 45.3 95
96 13.6 21.4 26.5 32.9 41 52 66 83 105 48.5 96
97 14.8 23.0 28.5 35.3 44 56 71 90 113 52.0 97
98 16.0 24.7 30.5 38 48 61 77 97 121 55.7 98
99 17.3 26.5 32.9 41 52 66 83 105 130 59.7 99
100 18.7 28.5 35.3 44 56 71 90 113 139 64.0 100
101 20.3 30.5 38 48 61 77 97 121 149 68.6 101
102 22.1 32.9 41 52 66 83 105 130 160 73.5 102
103 24.0 35.3 44 56 71 90 113 139 171 78.8 103
104 26.1 38 48 61 77 97 121 149 184 84.4 104
105 28.5 41 52 66 83 105 130 160 197 90.5 105
106 31.0 44 56 71 90 113 139 171 211 97 106
107 33.9 48 61 77 97 121 149 184 226 104 107
108 36.9 52 66 83 105 130 160 197 242 Ill 108
109 40.3 56 71 90 113 139 171 211 260 119 109
110 44 61 77 97 121 149 184 226 278 128 110
111 49 66 83 105 130 160 197 242 298 137 Ill
112 54 71 90 113 139 171 211 260 320 147 112
113 59 77 97 121 149 184 226 278 343 158 113
114 65 83 105 130 160 197 242 298 367 169 114
115 71 90 113 139 171 211 260 320 181 115
116 77 97 121 149 184 226 278 343 194 116
117 83 105 130 160 197 242 298 367 208 117
118 90 113 139 171 211 260 320 223 118
119 97 121 149 184 226 278 343 239 119
120 105 130 160 197 242 298 367 256 120
121 113 139 171 211 260 320 274 121
122 121 149 184 226 278 343 294 122
123 130 160 197 242 298 367 315 123
124 139 171 211 260 320 338 124
125 149 184 226 278 343 362 125

The calculated loudness is labeled sones (OD) and the loudness level is
abeled phons (OD) to designate that they have been calculated from octave-
)and levels (0) and for adiffuse field ( D).
A similar calculation can be made for third-octave bands, and they are
labeled (TD).
For steady noises having abroad frequency spectrum, the loudness calcu-
ated by means of the tables, which are based on Steven's* method agrees
Seasonably well with direct assessments made by loudness balances against a
1000-Hz tone.
To illustrate this procedure, consider the calculations based on octave-band
leasurements of the noise in afactory (Table 3-2).
For a quick check to find which band contributes most to the loudness,
tdd 3dB to the band level in the second octave, 6 dB to the third, 9 dB to
he fourth, and so on. Then the highest shifted level is usually the dominant
)and. This check will often be all that is needed to tell where to start in a
toise-
reduction program, if one doesn't have the loudness calculation charts
thand. This check is not reliable if the levels are low and the low-frequency
sands dominate.
*Loc. cit.

26
 Table 3 2

SAMPLE BAND LEVEL-TO- LOUDNESS-

INDEX CONVERSIONS
Octave Octave Band Band
Band Band Level Loudness
No. (Hz) (dB) Index
15 31.5 78 4
18 63 76 5
21 125 78 8
24 250 82 13
27 500 81 14
30 1000 80 16
33 2000 80 20
36 4000 73 15
39 8000 65 11

ES = Sum of Band Loudness Indexes = 106

Sm = Maximum Band Loudness Index = 20
0.3 ES = 31.8
0.7 Sm = 14

0.3 ES + 0.7 Sm = 46 sones ( OM*

or computed loudness level = 95 phons ( OD)*

*OD = Octave Diffuse ( an octave-band analysis for adiffuse field).

Another and more elaborate loudness calculation procedure has been

developed by Zwicker ( 1960) for third-octave analysis. It is not at all clear,
however, that this more difficult calculation results in a calculated loudness
that is in better agreement with subjective data.
Bauer and his associates ( 1971) have developed a simpler loudness meter
that has been applied to broadcast program monitoring.

3.6 PERCEIVED- NOISE LEVEL.

Kryter ( 1970) and his co-workers ( Kryter and Pearsons, 1963) have
followed a procedure similar to that used for loudness, but they asked the
observer to compare noises on the basis of their acceptability or their
"noisiness." The resulting judgments were found to be similar to those for
loudness, but enough difference was noticed to give a somewhat different
rating for various sounds. On the basis of these resultsKryter has set up a
calculation procedure for "perceived noise level," PNL in dB, alsQ called
"eNdB." The corresponding "noisiness" is given in units called "noys.
Ratings in terms of perceived noise level are now widely used for aircraft
noise particularly for aircraft flying overhead. The calculations for aircraft
noise are based on levels in third-octave bands and the detailed procedures
used are given in FAA regulations and in ISO recommendations.
A number of versions of the perceived-noise-level-calculation procedure
have been proposed. The one used here is for octave-band levels ( Kryter,
19x0). Proceed as follows:
.Combine on apower basis (use the chart of Figure 2-4), the levels in the
63-Hz and 125-Hz bands. Replace the 125-Hz level by this new level
and ignore the 63-Hz level, unless the 63-Hz level was greater than the
original 125-Hz level. In the latter case, replace the 63-Hz level by the
combined level, and ignore the 125-Hz level.

27
 In the frequency range from 500 Hz and up, note if any band level pro-
jects above the level in adjacent bands. If the difference between this pro-
jecting level and the average level in the two adjacent bands is 3 dB or
„ greater, acorrection is required, often called atone correction.
j. Find the noy values from the appropriate table given by Kryter ( 1970),
table 78 in his book, for the summed bands, the corrected band levels,
and the remaining band levels.
4. Add all these noy values ( E N).
5. Multiply the sum by 0.3.
6. Add this product to 0.7 of the noy value for that band that has the
largest noy value (0.3 E N + 0.7 N max ).
,.7.Convert this summed noy value to PNL in dB, by the use of the
1000-Hz noy-to-dB column.
Here is a sample calculation for the factory noise used previously for a
loudness calculation:

Octave Band Band

Center Band Level Noisiness
(Hz) ( dB) ( noys)

63 761
125 78 t80.1 11
250 82 16
500 81 18
1000 80 16
2000 80 28
4000 73 21
8000 65 9
EN = 119

0.3 X 119 = 35.7

0.7 X 28 = 19.6
55.3

55.3 noys correspond to 98-dB perceived noise level.

3.7 PERCEIVED LEVEL — STEVENS'S MARK VII.

As aresult of his extensive review of the available evidence on "loudness,"
"annoyance," "noisiness," "acceptability," "objectionability," etc of noise.
Stevens ( 1972) has revised his earlier calculation procedure in a number of
important respects.
1/3-octave band of noise centered at 3150 Hz is used as the reference
sound instead of a 1000-Hz tone, and this sound at a level of 32 dB re
20 eilsl/m 2 is assigned aperceived magnitude of 1soi.
An increase in level of 9 dB (it was 10 dB before) in the reference tone
doubles the perceived magnitude in son .
e).
The contours of equal perceived magnitude have been modified and the
masking factor in the calculation procedure now varies with level.

28
 Table 3-3*
PERCEIVED MAGNITUDE IN SONES AS A FUNCTION OF BAND PRESSURE LEVEL.

26-31 IS 16
Band 17 1
8 4
0 21 21 25 400- 32 II 14 1150- 40 41
Freq 50 63 12; 1611 260 (I; 1250 1600 2000 25)6 141100 10 000 12 000

1 313 0.078
2 0.087
3 0.078 0.097
4 0.087 0.107

5 0.078 0.097 0.118 0.078

6 0.087 0.107 0.129 0.087
0.078 0.097 0.118 0.141 0.097
8
' 0.087 0.107 0.129 0.153 0.107
0.078 0.097 0.118 0.141 0.166 0.118 0.078
10
4 0.087 0.107 0.129 0.153 0.181 0.129 0.087
II 0.097 0.118 0.141 0.166 0.196 0.141 0.097
12 0.107 0.129 0.10.4 0.181 0.212 0.153 0.107
13 0.077 0.118 0.141 0.166 0.196 0.230 0.166 0.118
14 0.087 0.129 0.153 0.181 0.212 0.248 0.181 0.129

15 0.097 0.141 0.166 0.196 0.230 0.269 0.196 0.141

16 0.107 0.153 0.181 0.212 0.248 0.290 0.212 0.153
17 0.076 0.119 0.166 0.196 0.2.40 0.269 0.314 0.230 0.166
18 0.086 0.130 0.181 0.212 0.248 0.290 0.319 0.248 0.181
19 0.097 0.143 0.196 0.230 0.269 0.314 0.367 0.269 0.196

20 0.108 0.156 0.212 0.248 0.290 0.3.49 0.396 0.290 0.212

21 0.075 0.120 0.169 0.230 0.269 0.314 0.367 0.428 0.314 0.140
22 0.086 0.131 0.185 0.248 0.290 0.339 0.396 0.463 0.339 0.248
23 0.097 0.144 0.201 0.269 0.314 0.367 0.428 0.500 0.367 0.269
24 0.108 0.158 0.219 0.290 0.339 0.396 0.46.4 0.540 0.396 0.290

25 0.074 0.121 0.17.3 0.237 0.314 0.367 0.428 0.500 0.583 0.428 0.314
26 0.085 0.134 0.190 0.256 0.339 0.396 0.463 0.540 11.6.30 11.46.1 0.3.49
27 0.097 0.147 0.207 0.279 0.367 0.428 0.500 0.583 0.680 0.500 0.367
28 0.110 0.162 0.224 0.102 0.396 0.463 0.040 0.6.30 0.7.30 0.540 0.396
24 0.073 0.122 0.178 0.244 0..429 11.428 0.500 0.583 0.680 0.794 0.583 0.428

30 0.085 0.136 0.194 0.267 0.356 0.46.3 0.540 0.630 0.735 0.857 0.630 0.463
31 0.097 0.149 0.212 0.290 11.384 0.500 0.583 0.680 0.794 0.926 0.680 0.500
32 0.110 0.165 0.233 0.316 0.418 0.540 0.630 0.735 0.857 1.00 0.735 0.540
33 0.072 0.123 0.182 0.254 0.345 0.452 0.583 0.680 0.794 0.926 1.08 0.794 0.583
34 0.084 0.137 0.201 0.277 0.375 0.490 0.630 0.735 0.857 1.00 1.17 0.857 0.630

35 0.097 0.153 0.221 0.304 0.406 0.531 0.680 0.794 0.926 1.08 1.26 0.926 0.680
36 0.111 0.169 0.241 0.332 0.442 0.576 0.7.45 0.857 1.00 1.17 1.36 1.00 0.7.45
37 0.070 0.125 0.187 0.264 0.361 0.481 0.624 0.794 0.926 1.08 1.26 1.47 1.08 0.794
38 0.084 0.140 0.207 0.290 0.396 0.523 0.676 0.857 1.00 1.17 1.37 1.59 1.17 0.857
39 0.097 0.156 0.228 0.319 0.431 0.570 0.732 0.926 1.08 1.26 1.47 1.72 1.26 0.926
40 0.112 0.173 0.250 0.350 0.470 0.618 0.794 1.00 1.17 1.36 1.59 1.85 1.36 1.00
41 0.126 0.193 0.277 0.381 0.511 0.672 0.860 1.08 1.26 1.47 1.71 2.00 1.47 1.08
42 0.142 0.214 0.304 0.418 0.561 0.729 0.933 1.17 1.36 1.59 1.85 2.16 1.59 1.17
43 0.160 0.237 0.337 0.459 0.611 0.794 1.01 1.26 1.47 1.71 2.011 2.3; 1.71 1.26
44 0.079 0.178 0.262 0.370 0.504 0.665 0.864 1.10 1.36 1.59 1.85 2.16 2.52 (. 05 (. 66

From S. S. Stevens, - Perceived Level of Noise by Mark VII and Decibels ( E)", The Journal of Acoustical Society of America, Vol. 51 No. 2 ( Part 2) Feb,
1972, pp 594-596. Reprinted with permission.
 Table 3-3 ( Continued)
PERCEIVED MAGNITUDE IN SONES AS A FUNCTION OF BAND PRESSURE LEVEL.

26-31 35-39
Band 17 18 19 20 21 22 13 24 25 400- 32 33 34 .4150- 40 41
Freq 50 6.4 30 100 125 ISO 2,1,) 250 .
115 1250 1600 2000 2500 8000 10 000 12 500

45 0.092 0.199 0.290 0.406 0.542 0.727 (

1.9 48 1.18 1.47 1.71 2.00 2.33 2.72 2.00 1.47
46 0.107 0.222 0.321 0.448 0.606 0.794 1.02 1.28 (. 59 1.85 2.16 2.52 2.94 2.16 1.59
47 0.121 0.246 0.336 0.492 0.664) 0.866 1.1(1 1..49 1.71 2.00 2.33 1.71 3.18 1.33 1.71
48 0.138 0.275 0.393 0.540 0.724 0.945 1.20 1.50 1.85 2.16 2.52 2.94 3.4 1 2.52 1.85
49 0.156 0.307 0.435 0.597 0.794 1.03 1.31 1.64 2.00 2.33 2.72 3. 18 4.70 1.71 2.00

50 0.072 0.176 0.341 0.481 0.655 0.871 1.12 1.42 1.77 2.16 2.52 2.94 1.43 4.00 2.94 2.16
51 0.086 0.197 0.378 0.531 0.724 0.955 1.23 1.55 1.91 2.33 1.71 .
4.18 3.70 4.32 .1.18 2.33
52 0.101 0.222 0.422 0.588 0.794 1.04 1.34 1.69 2.08 2.52 2.94 3.43 4.00 4.67 3.43 2.52
53 0.117 0.250 0.468 0.649 0.871 1.14 1.46 1.82 1.26 2.72 3.18 .4.70 4.32 5.04 3.70 2.72
54 0.134 0.279 0.519 0.718 0.962 1.25 1.59 1.98 2.44 2.94 .4.43 4.00 4.67 5.44 4.00 2.94

55 0.152 0.314 0.579 0.794 1.06 1.37 1.74 2.16 2.64 .4.18 .1.70 4.32 5.04 5.88 4.32 .3.18
56 0.175 0.347 0.643 0.877 1.17 1.50 (. 90 2..45 2.85 3.43 4.0(1 4.67 5.44 6.35 4.67 3.43
57 0.197 0.390 0.714 0.970 1.28 1.65 2.06 2.56 .4.10 .1.70 4.32 5.04 5.88 6.86 5.04 3.70
58 0.222 0.435 0.794 1.07 1.40 1.80 2.26 2.78 3.35 4.00 4.67 5.44 6.35 7.41 5.44 4.00
59 0.250 0.488 0.882 1.18 1.55 1.97 2.46 .4.01 3.65 4.32 5.04 5.88 6.86 8.00 5.88 4.32

60 0.282 0.544 0.977 1.31 1.70 2.16 2.68 3.27 3.94 4.67 5.44 6.35 7.41 8.64 6.35 4.67
61 0.319 0.611 1.09 1.45 1.87 2.37 2.94 3.56 4.27 5.04 5.88 6.86 8.00 9.33 6.86 5.04
62 0.358 0.686 1.21 1.60 2.06 2.60 3.20 3.88 4.63 5.44 6.35 7.41 8.64 10.1 7.41 5.44
63 0.402 0.762 1.34 1.77 2.26 2.8.4 .4.48 4.22 5.00 5.88 6.86 8.00 9.3.1 10.9 8.00 5.88
64 0.454 0.851 1.49 1.95 2.50 3.10 .1.79 4.58 5.44 6.35 7.41 8.64 10.1 11.8 9.64 6.35

65 0.511 0.952 1.66 2.16 2.74 3.40 4.16 4.98 5.88 6.86 8.00 9.33 10.9 12.7 9.33 6.86
66 0.574 1.06 1.84 2.39 3.01 3.73 4.52 5.40 6.37 7.41 8.64 10.1 11.8 13.7 10.1 7.41
67 0.649 1.18 2.05 2.64 3.32 4.09 4.94 5.88 6.91 8.00 9.3.4 10.9 12.7 14.8 10.9 8.00
68 0.729 1.33 2.28 2.92 3.65 4.47 5.40 6.40 7.48 8.64 10.1 11.8 I3.7 16.0 11.8 8.64
69 0.818 1.48 2.54 .
1.22 4.02 4.89 5.88 6.96 8.10 9.33 10.9 12.7 14.8 17..3 12.7 9.33

70 0.921 1.66 2.81 3.56 4.42 5.36 6.40 7.55 8.78 10.1 11.8 13.7 16.0 18.7 13.7 10.1
71 1.03 1.87 3.13 3.94 4.85 5.88 7.00 8.21 9.51 10.9 12.7 14.8 17.3 20.2 14.8 10.9
72 1.16 2.08 .3.48 4.35 5.34 6.45 7.64 8.91 10.3 11.8 13.7 16.0 18.7 21.8 16.0 11.8
73 1.32 2.3.3 3.85 4.81 5.88 7.07 8.33 9.70 11.1 12.7 14.8 17.3 20.2 23.5 17.3 12.7
74 1.48 2.58 4.29 5.32 6.47 7.70 9.09 10.6 12.1 13.7 16.0 18.7 21.8 25.4 18.7 13.7

75 1.66 2.90 4.76 5.88 7.13 8.46 9.92 11.5 13.1 14.8 17.3 20.2 23.5 27.4 20.2 14.8
76 1.87 3.24 5.28 6.50 7.82 9.26 10.8 12.5 14.1 16.0 18.7 21.8 25.4 29.6 21.8 16.0
77 2.10 3.62 5.88 7.18 8.61 10.2 11.8 13.5 15.4 17.3 20.2 2.3.5 27.4 32.0 13.5 17.3
78 2.37 4.03 6.53 7.94 9.48 11.1 12.9 14.7 16.6 18.7 21.8 25.4 29.6 34.6 25.4 18.7
79 2.66 4.52 7.26 8.78 10.4 12.2 14.0 16.0 18.0 20.2 23.5 27.4 32.0 37.3 27.4 20.2

80 2.99 5.05 8.06 9.70 11..5 13.3 15.3 17.3 19.4 21.8 25.4 29.6 34.6 40.3 29.6 21.8
81 3.35 5.64 8.95 10.7 12.6 14.6 16.6 18.7 21.0 23.5 27.4 32.0 37.3 4.4.5 32.0 23.5
82 3.79 6.31 9.96 11.8 13.8 16.0 18.0 20.2 22.6 25.4 29.6 34.6 40.3 47.0 34.6 25.4
83 4.25 7.05 11.1 111 15.3 17.3 19.4 21.8 24.4 27.4 .12.0 37.3 43.5 50.8 37.3 27.4
84 4.79 7.88 12.3 14.5 16.6 18.7 21.0 23.5 26.4 29.6 34.6 40.3 47.0 54.9 40.3 29.6
 Table 3-3 ( Continued)
PERCEIVED MAGNITUDE IN SONES AS A FUNCTION OF BAND PRESSURE LEVEL.

26-31 35-39
Band 17 18 19 20 21 22 23 24 25 400- 32 33 34 3150- 40 41
Fren 50 63 80 100 125 150 200 250 315 1250 1600 2000 2500 8000 10 000 12 500

85 5.40 8.81 13.7 16.0 18.0 20.2 22.5 25.4 28.5 32.0 37.3 43.5 50.8 59.3 43.5 32.0
86 6.06 9.85 15.2 17.3 19.4 21.8 24.4 27.4 30.8 34.6 40.3 47.0 54.9 64.0 47.0 34.6
87 6.82 11.0 16.6 18.7 21.0 23.5 26.4 29.6 33.3 .47.3 43.5 50.8 59.3 69.1 50.8 37.3
88 7.68 12.3 18.0 20.2 22.6 25.4 28.5 32.0 35.9 40.3 47.0 54.9 64.0 74.7 54.9 40.3
89 8.64 13.8 19.4 21.8 24.4 27.4 30.8 .44.6 38.8 43.5 50.8 59.3 69.1 80.6 59.3 43.5

90 9.71 15.4 21.0 23.5 26.4 29.6 33.3 37.3 41.9 47.0 54.9 64.0 74.9 87.1 64.0 47.0
91 10.9 16.8 22.6 25.4 28.5 32.0 35.9 40.3 45.2 50.8 59.3 69.1 80.6 94.1 69.1 50.8
92 12.3 18.3 24.4 27.4 30.8 34.6 38.8 43.5 48.9 54.9 64.0 74.7 87.1 102 74.7 54.9
93 13.8 19.8 26.4 29.6 33.3 37.3 41.9 47.0 52.8 59.3 69.1 80.6 94.1 110 80.6 59.3
94 15.6 21.5 28.5 32.0 35.9 40.3 45.2 50.8 57.1 64.0 74.7 87.1 102 119 87.1 64.0

95 17.1 23.3 30.8 34.6 38.8 43.5 48.9 54.9 61.6 69.1 80.6 94.1 110 128 94.1 69.1
96 18.6 25.3 33.3 37.3 41.9 47.0 52.8 59.3 66.6 74.7 87.1 102 119 138 102 74.7
97 20.3 27.4 35.9 40.3 45.3 50.8 57.1 64.0 71.9 80.6 94.1 110 128 149 110 80.6
98 22.1 29.8 38.8 43.5 48.9 54.9 61.6 69.1 77.6 87.1 102 119 138 161 119 87.1
99 24.1 .32.3 41.9 47.0 52.8 59.3 66.6 74.7 83.8 94.1 110 128 149 174 128 94.1

100 26.3 35.1 45.3 50.8 57.1 64.0 71.9 80.6 90.6 102 119 138 161 188 138 102
‘3.1 101 28.6 38.0 48.9 54.9 61.6 69.1 77.6 87.1 98.0 110 128 149 174 203 149 110
102 31.2 41.2 52.8 59.3 66.6 74.7 83.8 94.1 106 119 138 161 188 219 161 119
103 34.0 44.7 57.0 64.0 71.9 80.6 90.6 102 114 128 149 174 203 237 174 128
104 37.0 48.5 61.6 69.1 77.6 87.1 98.0 110 124 138 161 188 219 256 188 138

105 40.4 52.4 66.5 74.7 83.8 94.1 106 119 1.43 149 174 203 237 276 203 149
106 44.0 57.0 71.8 80.6 90.6 102 114 128 144 161 188 219 256 299 219 161
107 48.0 61.8 77.6 87.1 98.0 110 124 138 155 174 203 2.47 276 323 .
237 174
108 52.3 67.1 83.8 94.1 106 119 1.43 149 168 188 219 256 299 228 256 188
109 57.0 72.8 90.5 102 114 128 144 161 181 203 237 276 323 376 276 203

110 62.1 78.9 97.8 110 124 138 155 174 196 219 256 299 348 406 299 219
III 67.5 85.6 106 119 133 149 168 188 211 2.47 276 32.4 .476 430 323 237
112 73.8 92.9 114 128 144 161 181 203 228 256 299 348 406 474 348 256
113 80.5 101 123 138 155 174 196 219 246 276 323 376 4.49 512 376 276
114 87.8 109 133 149 168 188 211 237 266 299 348 406 474 553 406 299

115 95.6 119 144 161 181 203 228 256 288 323 376 439 512 597 439 323
116 104 129 155 174 196 219 246 276 311 348 406 474 553 645 474 348
117 114 139 168 188 211 237 266 299 336 376 439 512 597 697 512 376
118 124 152 181 203 228 256 288 323 362 406 474 553 645 752 553 406
119 13$ 164 196 219 246 276 311 348 391 439 512 597 697 813 597 439

120 147 178 211 237 266 299 336 .476 422 474 55.1 645 752 878 645 474
121 160 193 228 256 288 323 362 406 456 512 597 697 813 948 697 512
122 175 209 246 276 311 348 391 439 493 55.4 645 752 878 1024 752 553
123 190 227 266 299 336 376 422 474 532 597 697 813 948 1106 81.4 597
124 207 246 287 323 362 406 450 512 575 645 752 878 1024 1194 878 045
 Table 3-4't
F vs SONES IN ONE-THIRD 0.B.

Son, Sones

0.181 0.10 8.64 0.230

0.196 0.122 9.33 0.226
0.212 0.140 10.1 0.222
0.230 0.158 10.9 0.217
0.248 0.174 11.8 0.212

0.269 0.187 11.7 0.208

0.290 0.200 13.7 0.204
0.314 0.212 14.8 0.200
0.339 0.222 16.0 0.197
0.367 0.232 17.3 0.195

0.396 0.241 18.7 0.194

1/.428 0.250 20.2 0.193
0.463 0.259 21.8 0.192
1E500 0.267 23.5 0.191
0.540 0.274 25.4 0.190

0.58.1 1/.281 27.4 0.190

0.630 0.287 29.6 0.190
0.680 0.293 32.0 0.190
0.735 0.298 .14.6 0.190
0.794 0.303 37.3 0.190

1/.857 0.308 40.3 0.191

0.926 0.312 43.5 0.191
1.00 0.316 47.0 0.192
1.08 0.319 50.8 0.193
1.17 0.320 54.9 0.194

1.26 0.322 59.3 0.195

1.36 0.322 64.0 0.197
1.47 0.320 69,1 0.199
1.59 0.319 74.7 0.201
1.72 0.317 80.6 0.203

1.85 1/.314 87.1 0.205

2.00 11.311 94.1 0.208
2.16 0.308 102 0.210
2..43 0.304 110 0.212
1.52 0.300 119 0.215

2.72 0.296 128 0.217

2.94 0.292 148 0.219
3.18 0.288 149 0.221
3.43 0.284 161 0.223
.3.70 0.279 174 0.224

4.1110 0.275 188 0.225

4.32 0.270 203 0.226
4.67 0.266 219 0.227
SA/4 0.262 217 0.227
5.44 0.258 256 0.227

5.88 0.253
6.35 0.248
6.86 0.244
7.41 0.240
8.00 0.235

•The factor F as a function of the number of sones in the 1/3- octave band that is
maximally loud or noisy. The value of F remains constant above 219 sones.

tFrom S. S. Stevens, " Perceived Level of Noise by Mark VII and Decibels ( E),"
The Journal of Acoustical Society of America, Vol 51 No. 2 ( Part 2) Feb 1972,
p 597. Reprinted with permission.

32
 To calculate the perceived level and magnitude of anoise by the Mark VII
procedure, proceed as follows:
1. From Table 3-3 find the proper perceived magnitude in sones for each
band level.
2. From the maximum of these perceived magnitudes, Sm, find the factor,
F, from Table 3-4. If octave-band levels are used, subtract 4.9 dB from
the level of the loudest band; find the corresponding sone value; use
this value for finding the factor F; double the value found in the table,
and use it as F.
3. Add all the perceived magnitudes ( ES); subtract the maximum, Sm .
4. Multiply the sum by the factor F.
5. Add this product to the maximum perceived magnitude, St = ( 1-F) •
Sm + F E S. This value is the total perceived magnitude in sones.
6. Convert this magnitude to perceived level in dB from Table 3-3 by
the use of the 3150-Hz column.
This new perceived level will be about 8 dB less than that obtained for the
loudness level in the Mark VI calculation procedure. This shift is aresult of
the use of areference signal at 3150 Hz.
Here is a sample calculation for the factory noise used previously for the
loudness calculation:

Octave Band Perceived = 144.0

Center Band Level Magnitude ES - Sm = 114.4
(Hz) ( dB) ( sones)
Adjusted band level = 80 - 4.9 = 75.1 dB
31.5 78 Corresponding perceived
63 76 3.2 magnitude at 2000 Hz = 20.4 sones
125 78 9.5
2 X F = 2 X . 193 = . 386
250 82 20.2
500 81 23.5 .386 X 114.4 = 44.2
1000 80 21.8 Sm = 29.6
2000 80 29.6
4000 73 23.5 St = 73.8 sones
8000 65 12.7 PL = 87.8 dB

3.8 EFFECT OF TONAL COMPONENTS.

When a noise is a mixture of random noise and audible tonal components,
the loudness or annoyance may be somewhat greater than expectefrom the
direct analysis and calculation schemes ( Kryter and Pearsons, 1963; Wells,
1969). The effect is usually taken into account by a correction of the
calculated level, but it normally requires a detailed analysis of the sound at
least equivalent to that of athird-octave analysis.

3.9 EFFECT OF DURATION.

\If we are talking with someone and we are interrupted by a noise that
interferes seriously with speech, it is more annoying if it lasts for along time
than if it is very brieln order to take this effect into account anumber of
procedures have been suggested ( Pietrasanta and Stevens, 1958; Young, 1968;
Kryter, 1968). Most of them are, in effect, an integration (on apower basis)
of the loudness, perceived noise, or A-weighted sound pressure over time.

33
These procedures are still not well validated. Some qualifying term, for
example,<Effective Perceived Noise Level" is often used to indicate that a
correction for duration has been made

3.10 NC CURVES.
\Another rating procedure, which uses "noise-criterion curves," was devel-
oped by Beranek ( 1957) for design goals for satisfactory background noise
inside office buildings and in rooms and halls of various type.1(it is helpful in
deciding where in the spectrum additional effort is required in noise reduc-
tion, in order to make the noise acceptable>

NC 70
N.
N..
.....
\ CA- 60
dB re 20µN/m2

NC -60

o N
N.
IN

...... NCA-40

o
OCTAVE- BAND LEVEL

C 30

-eo

NC ZO

I
Figure 3-6. Noise- criteria curves.
THRESHOLD

OCTAVE- BAND CENTER FREQUENCIES IN Hz .

fflue

A set of these NC Curves ( Schultz, 1968; ASHRAE, 1967) are shown in

Figure 3-6. A threshold curve for octave bands of noise ( Robinson and
Whittle, 1964), is also shown for reference.
In use, the measured spectrum is plotted on the chart. ach band level is
then compared with the NC curves to find the one that penetrates to the
highest NC level. The corresponding value on the NC curve is the NC rating of
the noise>
As an example, the measured background noise level of an office is shown
on the figure as encircled crosses. This noise would have arating of NC- 38.
ince a recommended range of NC- 30 to NC-40 has been suggested for an
executive office) ( ASHRAE, 1967, p. 379), it would be considered acceptable
for that purpose. But if one were to turn it into a conference room, as is
sometimes done with large offices, it would not be as acceptable. Here, the

34
 recommended range is NC-25 to NC- 35. If one were to try to reduce the noise
level to make it more acceptable it is clear from the chart that one should try
to find the source of noise in the prominent 250-Hz band and work to reduce
that level.
The dashed NCA curves on the figure indicate the direction in which a
compromise should be made if economic considerations preclude achieving
the normal criterion given by the NC curve.

3.11 NOISE AND NUMBER INDEX.

\Another rating for aircraft noise, called NNI, noise and number index, is
based on perceived noise level. It was developed in Great Britain ( Committee
on the Problem of Noise, 1963)(a.nd takes into account the effect of the
number of aircraft per day on the annoyancèy It is defined by the following
relation:
<NNI = (Average Peak Perceived Noise Level) + 15 ( log i0 N) —8(?.>
where N is the number of aircraft per day or night. The value 80 is subtracted
to bring the index to about 0for conditions of no annoyance.
The " Average Peak Perceived Noise Level" is obtained in the following
way. The maximum perceived noise level that occurs during the passage of
each airplane is noted. These maximum levels are then converted into equiva-
lent power and averaged ( Section 2.6). This average value is then converted
back into alevel and used in the equation.
If the perceived noise level is approximated by the use of A-weighted
sound levels, the average A-level is obtained in a similar fashion, the 80 is
reduced to about 67, and we have
NNI •.•_•• ( Average Peak A-Level) + 15 (log i0 N) — 67.

3.12 NOISE-POLLUTION LEVEL.

Robinson ( 1969 and 1971) reviews a number of the measures derived in
various countries for rating a composite noise history. He lists the following
measures: Noise and number index, Composite noise rating, Storindex, Indice
de Classification, Aircraft Noise Exposure index, Noisiness index, Aircraft
Exposure level, Annoyance index, Traffic Noise index, Equivalent Disturb-
ance level, Office Noise Acceptability Scale, and Noise ' mission level. Most
of these are closely related.
egeneralizes these measures into anew measure called " Noise Pollution
Level," which is expressed by the relation

Lnp = Leg + 2.56 cr>

<where L eq is the noise level over a specified period averaged on an energy

basis, and a is the standard deviation (rms, see paragraph 5.5.1) of the
instantaneous level about that average value over the same perio (The
coefficient of a is not as accurate as the precision shown, but it was selected
by Robinson from a range of possible values to yield asimple relation for
certain noise-level distributions.)
l he noise level used in the expression can be the A-weighted level, the
loudness level, or some other similar level>
<In the calculation of noise-pollution level, the time period is to be one in
which similar conditions prevail. Thus, for example, night and day would be
treated separately

35
 This noise-pollution level is another way of looking at anoise history and
it yields results that are similar to the others. The variety of these measures
reflects the considerable activity in this area and the fact that many factors
enter into the effects that are to be predicted. Different measures are now
standardized and used by different groups. Since it is unlikely that a close
correspondence will be found between the effects and a combination of
physical measurements ( Hazard, 1971), some general agreement on arelative-
ly simple relation is urgently needed.

3.13 MASKING — " ICAN'T HEAR YOU WHEN THE WATER'S RUNNING."
It is common experience to have one sound completely drowned out when
another louder noise occurs. For example, during the early evening when a
fluorescent light is on, the ballast noise may not be heard, because of the
usual background noise level in the evening. But late at night when there is
much less activity and correspondingly less noise, the ballast noise may
become relatively very loud and annoying. Actually, the noise level produced
by the ballast may be the same in the two instances. But psychologically the
noise is louder at night, because there is less of the masking noise that reduces
its apparent loudness.
Experimenters have found that the masking effect of asound is greatest
upon those sounds close to it in frequency ( Egan and Hake, 1950; Fletcher,
1953). At low levels the masking effect covers arelatively narrow region of
frequencies. At higher levels, above 60 dB, say, the masking effect spreads out
to cover a wide range, mainly for frequencies above the frequencies of the
dominating components. In other words, the masking effect is asymmetrical
with respect to frequency. Noises that include a wide range of frequencies
will correspondingly be effective in masking over awide-frequency range.

3.13.1 Speech- Interference Level. Most of us have been in locations where

it was impossible to hear over a telephone because the noise level was too
high; and, in order to hear, production machinery had to be turned off,
resulting in time and money lost. Even direct discussions can be difficult and
tiring because of excessive noise. Excessive noise may make it impossible to
give danger warnings by shouting or to give directions to workerSerious
problems may occur because of speech interference from noisy machinery
while training employees to operate the machineri.,-,
In a large classroom with eavy acoustical treatment, particularly in the
ceiling, the attenuation may be so great that the teacher at one end can be
but poorly hearci)through the background noise at the other end, even though
the noise is not very great.
Incidentally, other factors also affect speech intelligibility. In alive room,
çveech syllables are smeared by reflected sound and the intelligibility is
consequently reduced.
Because of the annoyance of interference with speech and also because
noise interferes with work where speech communication is necessary, anoise
rating based on the speech-interference level is frequently useful! We should
know how to improve speech communication in anoisy place. In order to
effect this improvement we shall find it useful to evaluate the speech-inter-
ference level of anoise. How this can be done will appear from aconsidera-
tion of how noise interferes with speech>
Noise interference with speech is usually a masking process. The back-
ground noise increases our threshold of hearing, and, as aresult, we may hear

36
only afew or perhaps none of the sounds necessary for satisfactory intelligi-
bility.
<The consonants contain most of the information in speech, but, unfor-
tunately, they are more readily masked than vowels, because they are weaker
than vowels> Noise of acertain levelinay mask some speech sounds and not
others, depending on the talking level, the particular sound, and the relative
frequency distribution of the sound and of the noise
The energy of the various speech sounds is distributed over the frequency
range from below 100 to above 10,000 Hz. The actual instantaneous distribu-
tion depends on the particular speech sound. For example, the "s" sound has
its energy broadly distributed in the range from 4000 to beyond 8000 Hz. In
contrast, most of the energy in the "ce" sound of "speech" is distributed in
fairly definite groups ( called "formants") below 4000 Hz. All the frequency
range of speech sounds is not necessary, however, for complete intelligibility.
number of experimenters have shown that nearly all the information in
speech is contained in the frequency region from 200 to 6000 Hz?
In any frequency subdivision that we may make of this range, the sound-
pressure levels vary over arange of about 30 dB, as successive sounds occur.
Tests on the intelligibility of speech show thate we can hear the full 30-dB
range in each of the frequency bands into which speech is divided, the
contribution to intelligibility by that band will be 100 percent. If, however,
noise limits the range that can be heard to only 15 dB, the contribution will
be about 50%, and so forth≥ Furthermore<ii the range between 200 to 6000
Hz is divided into alarge number of frequèncy bands of equal importance to
speech intelligibility, the total contribution to speech intelligibility is equal to
the average of the contributions from the individual bands. This quantity is
called the articulation index, because it is ameasure of the percentage of the
total possible . information that we might have perceived of importance to
speech intelligibilit9French et al, 1947; Beranek, 1947; Kryter, 1962; ANSI
S3.5-1969).
For many noises, the measurement and calculation can be simplified even
further by the use of a three-band analysis ( Beranek, 1947).< The bands
chosen are the octave bands centered on 500, 1000, and 2000 Hz.* The
arithmetic average of the sound-pressure levels in these three bands gives the
quantity called the three-band preferred octave speech-interference level
(PSIOWne can use this level for determining when speech communication or
telephone use is easy, difficult, or impossible, and one can determine what
changes in level are necessary to shift from one order of difficulty to alower
order.
Face- To-Face Communication. For satisfactory intelligibility of difficult
speech material, maximum permissible values of speech-interference levels for
men with average voice strengths are given in Figure 3-7, which is an
extension by Webster of Beranek's work (Webster, 1969).
It is assumed in this chart that there are no reflecting surfaces nearby, that
the speaker is facing the listener, and that the spoken material is not already
familiar to the listener. For example, the speech-interference level of the
factory noise in paragraph 3.5.4 is 80 dB, which is high, and the chart

•The bands used before the shift to the currently preferred series were 600-1200,
1200-2400, and 2400-4800 Hz, or those three bands plus the band from 300 to 600 Hz.
The results of the two measures are similar, but some shift in the reference values is
necessary (Webster, 1965, 1969).

37
 +
e,. +
-4
e 4s

%
\ s„„

POSSIBLE \ e 414,4 DIFFICULT IMPOSSIBLE

, C4 7,»
WITH "NORMAL VOICE 4.

' ee0 - k
\ /

\ el,
< \ ...
-..
\
\
0.5
PSIL 40 50 60 70 80 90 100 110 I20
47 57 67 77 87 97 107
SIL 37
67 77 87 97 107 117 17
127
LA 47 57

:igure 3-7. Rating chart for determining speech communication capability from speech
nterference levels. (By permission, Webster, 1969.)

ndicates that the two people must ordinarily be no more than two feet apart
n order to be understood satisfactorily. If the words spoken are carefully
elected and limited in number, intelligible speech will be possible at greater
listances.
If anumber of conversations are to be held in the same reverberant room,
he procedure is more complicated. This chart cannot be used on the basis of
he background-noise level before the conversations are in progress, because a
:iven conversation will be subject to interference from the noise produced by
.11 the other conversations. The general procedure for calculating aspeech-in-
erference level under those conditions has not been completely worked out.
Telephone Usability in Noisy Areas The speech-interference level can also
)e used to predict the expected usability of atelephone under given noise
•onditions. The following schedule has been found generally satisfactory,
vhen the F-1 Western Electric handset is used for long-distance or suburban
•alls.

Speech- Interference Level Telephone Use

less than 60 dB Satisfactory

60 to 75 dB Difficult
above 80 dB Impossible

For calls within a single exchange, the permissible speech-interference

evels are 5dB greater than those shown.
Criteria for Indoor Noise Levels. A suggested rating system for offices,
ased on anumber of psychological and acoustical tests, is shown in Figure
-8. The curves on this graph relate the measured speech-interference level of
he background noise and the subjective rating of the noise ranging from
'very quiet" to " intolerably noisy." The two different rating curves illustrate
hat the environment influences the subjective rating. In order to be rated
'noisy" the noise level must be appreciably higher in alarge office than in a
nivate office.
It can be expected that the probability of receiving complaints about noise
vill be high for subjective ratings above " moderately noisy" and low for

38
 TELEPHONE USE SATISFACTORY TELEPHONE TELEPHONE USE
USE DIFFICULT UNSATISFACTORY
1 I 1 I I T
I

INTOLERAB LY -
NOISY

VERY NOI

0 Ce
S e
t+C,'

,k65- 60
NOISE -
c,e ,ee.
<,e• ,•-•
e
ee-
t •
/
1
4
1,- CRITERION POINTS DESIGNATE DE-
MODERATE LY _
fSe Ce SIRABLE UPPER LIA TS OF NOISE _
NOISY ,,t B
FOR INTELLIGIBLE DISCUSSION
A .
(A) NORMAL VOICE AT 9FEET
(B) SLIGHTLY RAISED VOICE AT 3FEET

OulET
-

VERY
QUIET _

i I I I 1
.. i I I
41 51 61 71 81 91 101
AVERAGE OF SOUND LEVELS IN BANDS AT SOD 1000.2000 11:
(PSIL)
DECIBELS RE 20 pN/rn'

Figure 3-8. Rating chart for office noises. Data were determined by an octave-band
analysis and correlated with subjective tests. (Courtesy Beranek and Newman, but
modified for preferred bands).

subjective ratings below " moderately noisy." Furthermore, because of direct

interference with transferring information, efficiency may be reduced for
levels appreciably above the criterion points marked A and B.
Suggested criteria for noise control in terms of maximum permissible
speech-interference level ( PSIL), measured when the room is not in use, are
given in Table 3-5.
The purpose of these criteria will be shown by the following example.
Assume that we are to put asmall conference room in afactory space. We
measure the speech-interference level at that location and find it to be 69 dB,
whereas the suggested speech-interference level criterion for asmall confer-
ence room is 35 dB. The room must then be designed to attenuate the noise
Table 3 5
CRITERIA FOR NOISE CONTROL

Maximum Permissible PSIL

(measured when room
Type of Room is not in use)
Small Private Office 45
Conference Room for 20 35
Conference Room for 50 30
Movie Theatre 35
Theatres for Drama
(500 seats, no amplification) 30
Coliseum for Sports Only ( Amplification) 55
Concert Halls ( No amplification) 25
Secretarial Offices ( Typing) 60
Homes ( Sleeping Areas) 30
Assembly Halls ( No amplification) 30
School Rooms 30

39
 from the factory space by about 34 dB, in order to have aconference room
that will be satisfactory as far as background noise level is concerned (such an
attenuation is provided by adouble-plastered, three- or four-inch thick stud
wall, or by ahollow-tile wall plastered on one side).

A similar but more extensive set of such criteria for noise control, based
on A-weighted sound levels, is given in the Guide and Data Book Systems and
Equipment, 1967, of the American Society of Heating, Refrigerating and
Air-Conditioning Engineers. (ASHRAE, 1967).
Privacy. Privacy of conversation is often desired both in the home or
apartment and in business. The use of extensive and carefully constructed
sound
sive, isolation is the safest way to ensure privacy. This approach is expen-
however.

If the noise level outside an executive office is relatively high, only a

moderate amount of isolation may be needed to bring the speech level from
the office to the point outside the office where it is masked by the back-
ground noise. It is important then that the executive office have a back-
ground PSIL below 45 dB, in order to avoid encouraging araised voice level.
An exceptionally low background noise level, however, may make it possible
for the one in the office to hear those outside, and he will then feel that his
true.
office is not private even though it may be so in fact. The inverse may also be

If his air conditioner is exceptionally noisy, he may feel that his speech
will be covered by the noise. But if the adjacent space is relatively quiet, he
may be overheard. In fact, privacy in offices depends on some background
noise as well as isolation and distance (Cavanaugh, et al, 1962; Young, 1965),
and mutual privacy is often essential. This approach to privacy sometimes
requires that noise be introduced, often conveniently by way of turbulent
noise from aventilator grill (Waller, 1969).

3.14 CRITICAL RATIO AND CRITICAL BANDWIDTH.

Early studies of masking led Fletcher ( 1953) to define acritical bandwidth

for hearing. He measured the threshold of pure tones masked by wide bands
of noise whose frequency range spread about that of the tone. In the
comparison of the levels of the tone and the noise, he used the spectrum level
of the noise, which is the level that would be obtained if the noise were
filtered through an ideal filter 1-Hz wide (see Chapter 5).(i-le found that, for a
wide range of levels, the difference between the threshold level of the masked
tone and the spectrum level of the noise was aconstant. This constant is now
dten called a critical ratio. It varies with the frequency of the pure tone,
'rom about 17 dB at 300 Hz to about 28 dB at 8000 Hz›Hawkins et al,
950).
iith At frequencies below about 500 Hz, the ratio does not change much
frequency.

Fletcher tried narrowing the bandwidth of the noise but with its frequency
Inge centered on the frequency of the tone. He found that above a certain
ideal bandwidth the critical ratio was reasonably independent of the band-
idth.\Below this critical bandwidth the critical ratio decreased, that is, the
ne could be more readily heard in the noise. This experiment led to the
ncept that
•ertain masking
frequency by anoise is mainly aresult of the noise energy within
band

The early values quoted for this critical bandwidth are small, being about
-Hz wide at 250 Hz and 500 Hz, increasing to about 600 Hz at 8000 Hz.

40
 Since then, the results of many experiments have shown large enough varia-
tions to cast doubt on the reliability of the critical-band measurement by this
technique (de Boer, 1962; Green and Swets, 1966).
Other psychoacoustical measurements have led to specifications of critical
bands that are appreciably wider than those quoted above. For example, the
loudness of aband of noise is observed as afunction of the bandwidth of the
noise with constant overall level. The experiments show that up to acertain
critical bandwidth the loudness is essentially independent of the bandwidth.
Beyond that point the loudness increases with constant over-all sound-pres-
sure level (Zwicker et al, 1957). This critical bandwidth is relatively independ-
ent of the level.
This loudness critical band is found to be about 90 Hz wide, centered at
100 Hz, 110 Hz at 500 Hz, and thereafter increasing to about 2300 Hz at
10,000 Hz. Over much of the range it can be reasonably well approximated
by aone-third octave. The Zwicker method of loudness calculation is based in
part on use of these critical bands ( Zwicker, 1960).

3.15 ADDITIONAL HEARING CHARACTERISTICS.

In addition to the characteristics already described, numerous others have
been investigated, and a few of these are of interest in noise-measurement
problems. Therefore, we shall discuss briefly differential sensitivity for inten-
sity and the pitch scale.

3.15.1 Differential Sensitivity for Intensity. One question that comes up in

quieting anoisy place or device is: "Just how little achange in level is worth
bothering with? Is a one-decibel change significant, or does it need to be
twenty decibels?" This question is partially answered in the section on
loudness, but there is additional help in the following psychoacoustical
evidence. Psychologists have devised various experiments to determine what
change in level will usually be noticed ( Stevens, 1951).'When two different
levels are presented to the observer under laboratory conditions with little
delay between them, the observer can notice as small a difference as 1/4 dB
for a 1000-Hz tone at high levelsjpis sensitivity to change varies with level
and the frequency, but over the range of most interest this differential
sensitivity is about 1/4 to 1d13. For a wide-band random noise (a " hishing"
sound) asimilar test gives avalue of about 1/2 dB for sound-pressure levels of
30 to 100 dB (re 20 µN/m 2 )<Under everyday conditions, a 1-dB change in
level is likely to be the minimum detectable by an average observir. On the
basis of these tests, we can conclude that 1-dB total change in level is hardly
worth much, althougK6 is usually significanèt should be remembered,
however, that many noise problems are solved by a number of small reduc-
tions in levelThere is also the importance of a change in character of the
noise. For example, the high-frequency level of a noise may be reduced
markedly by acoustic treatment, but, because of strong low-frequency com-
ponents, the over-all level may not change appreciably. Nevertheless, the
resultant effect may be very much worthwhile. This example illustrates one
reason for making a frequency analysis of anoise before drawing conclusions
about the noise.

3.15.2 Pitch and MeIs. Just as they have done for loudness, psychologists
have experimentally determined ascale for pitch (Stevens et al, 1937). The
unit for this scale is the ‹mel" (from " melody"), and a 1000-Hz tone at a

41
level of 40 dB is said to have apitch of 1000 melln terms of frequency, this
pitch scale is found to be approximately linear bi
elow 1000-Hz and approxi-
mately logarithmic above 1000-Hz. Some people have suggested that afre-
quency analysis with bands of equal width in mels would be more efficient
for some types of noise analysis than would one with bands of other widths.
At present no commercial analyzers of this type are available, but some work
has been done using such an analysis. In addition, the pitch scale has been
found useful for some types of charts.

3.16 WEIGHTED SOUND LEVELS.

Rating noise by loudness level, perceived-noise level, perceived level,
speech-interference level, or a noise-criterion curve is sufficiently complex
that simpler techniques have been sought. Often the simpler approach is to
return to the earlier procedures, that is, to use one or two of the presently
available weightings on the sound-level meter when rating noises of similar
character. The A-weighted sound level has been the most successful of these
measures (Young, 1958, 1964; Parkin, 1965; Botsford, 1969).
Another suggested approach is to use anew weighting characteristic, and
some of these will be described briefly before reviewing the relative success of
weighting networks and their limitations.

3.16.1 Other Weighting Characteristics. Among a number of weighting

characteristics that have been suggested recently, N, D or DI,and D2 have
been proposed for estimating perceived-noise level ( Kryter, 1970, Kryter and
Pearsons, 1963) E (for Ear) has been suggested for perceived level ( Stevens,
1972) and SI is proposed for speech interference (Webster, 1969). These
weighting characteristics are shown plotted in Figure 3-9 along with the

IL Da
10
,-. -

.....•
/
..,..
.. .: A
0
RELATIVE RESPONSE IN dB

..
D

•
•
.'xi,
20 \
/ .
///
/

/
\--
SO

/
'

40 1
25 50 100 250 500 iK 2.5K 5K IC
FREQUENCY IN HERTZ

.11MM 9

Figure 3-9. A weighting and other proposed weightings.

42
standard A-weighting. The N weighting is not shown, since it is the same as
the D weighting but raised in level by 7dB ( Batchelder, 1968).
The A-weighting raised by 5-dB is also shown for comparison. It is easy to
see that the trends of the D, D2, and E weightings are similar to that of the A.
As a result, if we are concerned mainly with rank ordering noises whose
energy is widely distributed over the frequency range, we would not expect
marked differences in usefulness among these weightings.

3.16.2 Comparison of Calculation Schemes and Weighted Levels. We shall

discuss briefly anumber of techniques for the following tasks:
1. Predict a subjective effect of any of a variety of noises including a
reference tone or narrow band of noise.
2. Rank order any of a variety of common noises for aparticular subjec-
tive effect.
3. Rank order noises of similar character, for example, automobiles, for
their subjective effects.
4. Predict from aweighted level the answer that acalculation scheme gives
on avariety of noises.
We must recognize, as discussed earlier, that subjective effects are not
consistent to start with, and many factors beyond the physical measurements
can enter into the result. Even as far as physical measurements are concerned,
however, the effect of the duration of the noise, for example, is an important
factor that cannot yet be adequately taken into account.
If we ignore these points and concentrate on the relative behavior of the
various weighting and calculation schemes, we find that the calculation
schemes tend to be more consistent than asimple weighting for predicting
results if they are to be referenced to a tone or a narrow band of noise
(Fletcher and Munson, 1933; Churcher and King, 1937; Beranek et al, 1951;
Quietzsch, 1955; Stevens, 1956; Kryter, and Pearsons 1963; Bauer et al,
1971; but see Corliss and Winzer, 1965 for an exception). Some relatively
large discrepancies in loudness, for example, appear in comparing wide-band
noise and a pure tone if a weighted level is used but the loudness predicted
from a calculation scheme can be much more nearly in agreement with the
subjective effect.
Such errors made in predictions from weighted levels led to many of the
studies of loudness summation and to various calculation schemes.
The situation is somewhat different if we merely need to rank order a
variety of common noises for their loudness or perceived noise level. Then, if
A-weighting or a similar one is used, the consistency is fairly good ( Young,
1964; Klumpp et al, 1963). But a C-weighted level almost always appears to
be significantly poorer than an A-weighted level in consistency (Wells, 1969).
The consistency for interference with speech for avariety of noises ( Webster,
1969), for example, is particularly poor for a C-weighted level (standard
deviation ( a) = 7.4 dB), better for an A-weighted level ( a = 4.7 dB), and still
better for PSIL ( a = 2.8 dB).
•I3ecause of the nature of the speech-interference effect, the proposed SI
weighting should be better than A-weighting for predicting speech interfer-
ence, but it should not be as good for predicting other effects, such as
loudnes
When we rank order noises of similar character, for example, automobiles
or aircraft, we find still less significance in the difference in behavior between
the calculation schemes and an A-weighted or similar level ( Hillquist, 1967;

43
Young, 1964; Young and Peterson 1969; Lavender, 1971). But again the
C-weighted level is generally poor.
Because the calculation schemes are used for specific effects, anumber of
studies have been made of how well the various weighted levels can be used to
predict the results of the calculation schemes. As ageneral rule the weightings
that are closely related to the calculation scheme tend to predict the related
result slightly better than other weightings ( e.g. Parkin, 1965; Stevens 1972).
Since the A weighting has the same general trend as the weightings used in the
zalculation schemes, it tends to work reasonably well for any of them ( Loye,
1956; Botsford, 1969). It should be recognized that this predictability is only
at* limited value, since it i removed by one additional variability from the
subjective effect.
NThe relative uniformity of the basic data used for loudness and perceived

wise (Stevens, 1972), and the success, however modest, of the A-weighted
evel in comparison with the C-weighted level leads one to conclude that some
mprovement could probably be obtained for rating many noises by the use
)f a weighting that is more like the E or the D2 weightink>(Kryter, 1970;
itevens, 1972).
The speech-interference level appears to be in adifferent category, how-
:ver. Because the speech-frequency range is more limited, it is likely that a
weighting such as the D2 or E would be no better, if not less satisfactory,
:han the A weighting. PSIL is still the approach to use in this application.

3.16.3 A-Weighted Sound Level as a Single- Number Rating. For simple

ratings or screenings of similar devices, the A-weighted sound level at a
specified distance is now widely used. This measurement is mainly useful for
relatively nondirectional sources that are outdoors and where the effect of
the noise also occurs outdoors and nearby. It is also useful in preliminary
ratings of similar ambient noises for the human reactions that may occur.
Measurement of A-weighted sound-level has been adopted for checking corn-
plia9ce with many ordinances and regulations.
"---Because of its widespread use, a number of investigators have determined
the approximate relation between the A-weighted sound-level of anoise and
:he calculated loudness level, perceived noise level, and speech-interference
evel of the noise. Table 3-6 shows the results compiled from various sources..,.
Robinson et al, 1963; Parkin, 1965; Young, 1964, Young and Peterson
1969; Hillquist, 1967; Nakano, 1966; Klumpp and Webster, 1963; ASHRAE,
1967; Jahn, 1965 --- Because the calculation schemes changed somewhat over
:he years, complete uniformity in procedures was not maintained; but the
effects of the changes were small>
In order to get more consistent results in these relations, Botsford ( 1969)
las used the difference in the C-weighted and A-weighted levels as an
tdditional parameter. He has compiled an extensive set of charts for the
rarious quantities and their relations to the A and C-A level. For the 953
wises he uses, the correlation and standard deviation are relatively good,
-

)eing poorest for speech-interference level, which had astandard deviation of

≥.9 dB (see also Shimizu, 1969).

3.16.4 Some Limitations of aWeighted Sound Leve When only a single

weighted sound level is measured, the usefulness of the measurement is
;everely restricted. One should almost always try to measure the spectrum
ilso. The spectrum is needed for efficient noise control, because the effects of

44
 Table 3 6

COMPARATIVE NOISE RATINGS

LL- LA PNL-LA LA- PSI L

Noise Type (dB) ( dB) (dB)
Office 13 13 E)
Truck 10 13
Pneumatic Machines 13 14
Ship Compartment 14 15
Urban 16
Aircraft — Jet 9 12
Prop 14
Prop ( approach) 16
Helicopter ( piston) 14
Aircraft — General q
— Flyover 1
,
Diverse 12 10
Airflow 12

LL = Loudness level by Stevens Mark VI

PNL = Perceived noise level — 1963-65
PSI L = Three- band preferred-octave speech- interference level
LA = A-weighted sound level

sound isolation, acoustic treatment, vibration reduction and other forms of

noise control are frequency depende nt In addition, the reaction to the no i
se

is frequency dependent, and the spect. um can show us the frequency region
where the noise energy is most important in determining the effects.
We almost always want to know the reason for the noise rating. The
spectrum often provides the most important clues for tracking down and
reducing the noise.
If anoisy machine is to be used in aroom, we need to know the acoustic
characteristics of the room as a function of frequency and the radiated-
sound-power level in octave or third-octave bands, in order to estimate the
noise level at some distance from the machine.
The spectra help in the long run in providing data for later comparisons
when conditions change or if better evaluation techniques are developed.
The limitations of the simple, weighted measurement should be recognized
when plans for sound measurements are being made.

3.17 RESIDENTIAL NOISE LEVELS.

Some factories, recreation halls, electrical substations, trucks, and air-
planes are so noisy that they annoy people living near them. The reactions of
those that are annoyed may range from mild remarks to legal action. Those
that are responsible for the noise would naturally like to avoid the expense of
court action; in order to maintain the good will of the neighborhood, they are
often willing to put considerable effort into controlling the noise, so as to
avoid anything but mild annoyance.
In order to put this noise control on a systematic basis, a number of
engineering groups have analyzed the experiences obtained in many different
situations. They have found that reactions of annoyance cannot be success-
fully predicted on the basis of a single measurement, or even of computed
loudness ratings, but that many factors enter into the problemn addition to
the range of reactions to be expected from different individuals, some other
factors are the following: The level and spectrum of the noise; whether or not

45
.here are strong, pure-tone components; the time pattern of the noise,
ncluding the rate of repetition and the actual time of occurence during t hç
lay; and the general background noise level in the residential area affected.
3o far, the data that is available is limited primarily to the reactions of people
n residential areas of single-family houses surrounding industrial plants. We
:an expect that, because of the conditioning to noises that occur in multiple-
imily dwellings, the reactions of the people there would be modified.
KStu dies have also shown that people vary markedly in their susceptibility
:o noise: In asurvey in central London (McKennel and Hunt, 1966), of those
nterviewed about general factors affecting living, 25% were classified as
-elatively insusceptible to noise while 10% were regarded as extremely suscep-
:ible. Since no relation was found between the existing noise environment
where they lived and the degree of susceptibility of these adults, one can
expect to find some highly susceptible individuals in almost any neighbor-
lood.
The work of Rosenblith and Stevens in 1953 (see also Stevens et al, 1955;
Beranek, 1954) and Parrack ( 1957) forms the basis of many of the rating
ichemes used for residential noise. This work has been modified and verified
by others ( Kosten and Van Os, 1962). To illustrate the nature of the
procedure, here is essentially the techniques developed by those authors but
modified to use A-weighted sound levels.
The A-weighted sound level in the residential area can be measured and the
corrections applied. ( Table 3-7).
Compare the corrected level with Table 3-8 to estimate the average public
reaction.
The overlap in these ranges indicate to some extent the variable nature of
the reaction.

3.18 EFFECT OF NOISE ON WORK OUTPUT.

Noise can influence work output in many ways; it can interfere with
:ommunication ( paragraph 3.13.1), and it can cause adecrease in the quality

Table 3-7

CORRECTIONS REQUIRED FOR VARIOUS

NOISE SUSCEPTIBILITIES
Correction
Noise Characteristic (dB)

Pure tone easily perceptible +5

Impulsive and/or intermittent character +5
Noise only during working hours -5
Noise duration
continuous o
< 30 min -5
10 min -10
< 5 min -15
< 1min -20
< 15 sec -25
Neighborhood
Very quiet suburban +5
Suburban o
Residential urban -5
Urban near some industry -10
Area of heavy industry -15

46
 Table 3 8

PUBLIC REACTIONS TO CORRECTED NOISE LEVELS

Corrected Level-d13(A) Expected Reaction

< 45 No observed reaction
45-55 Sporadic complaints
50-60 Widespread complaints
55-65 Threats of community action
> 65 Vigorous community action

of work output when the background noise level is above 90 dB, but noise is
occasionally useful as ameans of masking distracting conversations.
Broadbent ( 1958) and others have found that the effects of noise on work
output depend greatly upon the nature of the work; along-term job requiring
constant vigilance is especially susceptible. Noise is more likely to cause a
higher rate of errors and accidents than an actual reduction in total output.
This result and other findings lead to the interpretation that attention
wanders from the work at hand more often as the noise level increases.
From the standpoint of noise reduction, two findings are worth noting:
first, noise is more likely to lead to increased errors in susceptible tasks if it is
above 90 dB; and second, high-frequency audible noise seems more harmful
in this respect than does low-frequency noise.

3.19 NON- AUDITORY EFFECTS OF NOISE ON MAN.

Recently, Glorig ( 1971) and Kryter ( 1970) have summarized the present
knowledge of nonauditory effects of noise exposure. Very high levels ( 120 to
150 dB), at certain resonant frequencies of the body structure, can produce
noticeable symptomatic reactions. Even moderate noise levels produce tem-
porary changes in the size of some blood vessels, but it is not clear that these
effects eventually produce permanent changes. The production of stress and
fatigue by noise exposure is difficult to verify in ameaningful way.
Much more research is necessary, and arecent report (Welch and Welch,
1970), of a symposium on the physiological effects of noise, shows that
considerable effort is directed toward trying to establish the extent of such
effects.

3.20 WHAT VIBRATION DOES.

Vibration related problems can be classified as in Table 3-9.
These problems will be discussed in the following sections on the effects
on man, maintenance, vibration specifications, and other effects. The prob-
lem of excessive noise has already been reviewed.

3.20.1 Effects of Vibration on Man. The subjective effects of vibration are

important to those concerned with passenger or operator comfort in automo-
biles, planes, boats, trains, and other vehicles. Vibration levels that are
structurally safe for a vehicle are often uncomfortable, annoying or even
dangerous for the occupant. Some machinery and hand tools vibrate all or
parts of the body, and this vibration may affect performance as well as
comfort. Sometimes buildings and floors vibrate enough to be alarming or to
affect the performance of fine tasks.
Such effects have led to extensive studies, which have been reviewed
comprehensively by Goldman and von Gierke ( 1961) and Guignard ( 1965).
These excellent reviews, which cover the injurious levels of vibration as well

47
 Table 39
VIBRATION- RELATED PROBLEMS

Effect on man
Injury
Fatigue
Annoyance
Interference with performance

Mechanical failure
Excessive stress
Fatigue
Destructive impacts

Other
Excessive wear
Excessive noise
Inadequate performance
Failure to satisfy vibration specifications

as the subjective aspects, are recommended to those concerned with these

problems.
The sensation of vibration is not localized as it is for hearing, since
vibration can be felt throughout the body and different mechanisms operate
to provide the sensation. Curves that present human responses to vibration
cannot, therefore, be as complete as are the equal-loudness curves for simple
tones of sound.
As an example of information that is available, Figure 3-10 shows results
reported by Parks ( 1962) for vertical whole-body vibration that was classed
by the subjects as "mildly annoying". Another contour is given as the
approximate mean threshold at which subjects found the vibration unpleasant
(Goldman and von Gierke, 1961). The variability of this determination is
large, with astandard deviation of about + 4dB, —6dB.
Some comfort criteria and tolerance criteria are now in use (Goldman and
von Gierke, 1961; Guignard, 1965). As an example, Janeway (SAE J6a,
1965) has prepared a chart giving recommended limits of vertical vibration
for passenger comfort in automobiles. Janeway limited his analysis to data
obtained for vertical sinusoidal vibration at asingle frequency, with subjects
standing or sitting on ahard seat. The recommended characteristic consists of
three simple relations, each of which covers aportion of the frequency range.
In the low-frequency range from 1to 6 Hz the recommended limit is afixed
value of jerk. The corresponding maximum comfortable displacement at any
frequency between 1and 6Hz is 2divided by the frequency cubed (f 3 ). Over

the frequency range from 6 to 20 Hz the recommended limit is a constant

acceleration. The corresponding displacement is 1/3 f 2. From 20 to 60 Hz the
recommended limit is a constant velocity, and the corresponding displace-
ment is 1/60 f. In each instance, the amplitude calculated from these
formulas is the maximum displacement from the static positions, expressed in
inches. The limits are plotted in Figure 3-11, in terms of the rms acceleration
and vibratory-acceleration level in dB re 10 -5 m/s2 ,rms.
Resonance effects of the internal organs and their supports, and the upper
torso and the shoulder-girdle structures, probably account for the marked
sensitivity to vibration in the range from 4 to 10 Hz (Goldman and von
Gierke, 1961; Guignard, 1965). Many other resonances occur, however,
because the body structure is so varied (Guignard, 1965). The resonances that

48
 110

2 80

MILDLY ANNOYING ( PARKS 1962)

VERTICAL VIBRATION

_ 1.0 40

E MEAN THRESHOLD OF UNPLEASANTNESS

• AVERAGE-STANDING,SITTING,LYING
(GOLDMAN et al, 1961) o,
VI BR ATORY

.05 20

10
.•••

VERTICAL VIBRATION LIMIT-( SAE J6a)

02 — e

80 4
2 20 40 80 100
FREQUENCY IN Hz

Figure 3-10. Subjective response of the human body

to vibratory motion as a function of frequency.

are observed depend on the mode of excitation and the place the vibration is
applied.
One of the effects of vibration occurs in some who have worked for two or
more years with certain hand-held power tools. They may exhibit Raynaud's
phenomenon, in which the fingers become white and numb when the person
is chilled. The incidence of this effect seems to be closely related to the
vibrational energy in the frequency range from 40 to 125 Hz (Guignard,
1965).
For a visual task, large-amplitude vibration at frequencies between 2and
20 Hz is particularly disturbing (Guignard, 1966).

3.20.2 Maintenance. It is widely recognized that excessive vibration leads

to high costs for machinery maintenance. Conversely, gradual deterioration of
machinery, for example, bearings going bad or rotors becoming more un-
balanced, leads to increased vibration and noise. Recognition of this latter fact
has led some groups to institute periodic vibration measurements of machin-
ery as an important preventive maintenance procedure ( Bowen and Graham,
1967; Maten, 1970; Schiff, 1970; Glew and Watson, 1971).
The analysis of vibration permits one to estimate the probable condition
of the machine, to schedule downtime for maintenance usually before the
condition gets too serious, and to tell what to look at when the machine is
shut down.
If a program of this type is pursued, some acceptable limits of vibration
must be set to make possible adecision as to when corrective measures must
be taken. One approach is to analyze the vibration velocity at all bearing
housings, when the machine is newly installed and periodically thereafter.

49
When an appreciable change in vibration level is noted, the amoun I. of change,
the frequency region where it occurred, and the measurement location are
used to decide what action if any is necessary.
These early measurements should include the vibration at the various
bearing housings in all three directions, vertical and the two horizontal axes.
They should be measured for the different operating conditions made possi-
ble by the various clutches and speed-changing systems on the machine.
Incidentally, these early checks may occasionally reveal afaulty new machine
that should be rejected and returned to the manufacturer.
Various degrees of refinement are used in spectrum analysis of these
vibrations. A separation of the spectrum into the standard 8- to 10-octave
bands (paragraph 5.2.1) is often adequate (Glew and Watson, 1971), but the
finer divisions of the 1/3-octave ( paragraph 5.2.2) ( Bowen and Graham,
1967), and even an analysis into hundreds of bands by an FFT analyzer (see
paragraph 5.2.3), are also used.
In addition to ahistory of the vibration levels for each machine, it is useful
to have atype of absolute criterion, of which anumber have been proposed.
Among those who have proposed criteria, T. C. Rathbone ( 1939) was a
pioneer in synthesizing the available experience in this area. The chart that he
prepared in 1939 has been the basis for many subsequent specifications. This
chart showed the maximum allowable peak-to-peak displacement as afunc-
tion of rotation speed, with ratings varying from "Very Smooth" to "Too
Rough to Operate."
One of the important points to be gained from such charts is that asimple
specification of displacement or even of acceleration is not adequate for a
rating, although many have assumed from physical reasoning that one of
those parameters should be specified. Actually, velocity happens to be a
better parameter to use for a relatively wide range of shaft speeds. For
example, Rathbone has recommended some simplified upper limits of vibra-
tion that can be specified in terms of velocity for vibration frequencies above
20 Hz ( 1200 cycles per minute). The limits that he recommended ( Rathbone,
1963) are: For power machinery, electric motors, large fans, turbines, pumps,
dishwashers, dryers, vacuum cleaners, mixers, etc., the velocity should be less
than 0.13 in./s, peak ( 110 dB re 10 -8 m/s peak *). For hand tools, small fans,
and room air conditioning equipment, the velocity, should be less than 0.1
in./s peak ( 108 dB re 10 -8 m/s, peak)**. For precision machinery and busi-
ness machines, the velocity should be less than .063 in./s, peak ( 104 dB re
10 -8 m/s, peak).
These values should be used only as a guide. Considerable variation in
significance can be expected for several reasons. For example, the relation
between the actual spindle or shaft vibration and the vibration measured on
the associated bearing housings is complex and would not necessarily be the
same for machines of the same type but of different design.
Furthermore, the vibration at a bearing housing may vary significantly
around the housing because of components of different phase being intro-
duced external to the bearing. The nature of the vibration, that is, if it is

•Fqr the equivalent rms value (re 10 -8 m/s, rms) subtract 3 dB; for average values (re
10 m/s avg), subtract 4 dB,

The ratings in terms of rms values of sinusoidal vibration, as measured on some

vibration meters, will be about 0.7 of these peak values; for average values ( actually
"average absolute"), use 0.6 of the peak values.

50
rough or random or of an impact type rather than if it is sinusoidal motion,
also affects the value that is significant.
Even if no element of human reaction is involved, different criteria can be
set up for the same application. Thus, the manufacturer of acompressor may
select avelocity of 0.5 in./s, peak ( 122 dB re 10 -8 m/s, peak) measured on
the bearing housings, as asafe upper limit, but the user may prefer to have
the vibration kept to 0.1 in./s, peak ( 108 dB re 10 -8 m/s, peak) or less, for
best performance and low maintenance costs ( cf. Power, Vol 109, May 1965,
pp 162-164.)
The manufacturer is influenced by what can be competitively produced
and still have areasonable life, but the user should be willing to pay more for
a unit with the reduced maintenance costs that usually accompany lower
vibration levels.
It is important to recognize that resilient mounting of amachine will not
ordinarily reduce the vibration levels at the machine unless the vibration is
coming from the foundation. Resilient mounting may be helpful in preven-
tive maintenance, however, since it can reduce the effects of extraneous vi-
brations on the machine being measured. The vibration data from any given
machine are then more representative of the condition of the machine.

3.20.3 Vibration Specifications. Limits on vibration on many machines

have been set for avariety of reasons, generally on the basis of experience.
For example, on agood lathe one may find aspecification such as:
'Vibration to 1200 rpm (20 Hz) should not exceed 0.0005 in. on bed and
0.0003 in. at spindle.
These are peak-to-peak measurements and the corresponding peak-velocity
measurements at 1200 rpm are .03 in./s and .018 in./s. Such aspecification
should help to insure both high quality of work and low maintenance. But
it is strange to find that many manufacturers and users of precision rotating
machinery neglect such an important specification.

3.20.4 Other Effects. Many of the useful effects of vibration in chemical,

biological, and physical procedures are discussed by Hueter and Bolt ( 1955),
Crawford, ( 1955), Bergman ( 1954), Frederick ( 1965), and Brown and Good-
man ( 1965). The effects of machine-tool vibration have been reviewed by S.
A. Tobias ( 1961), and metallic fatigue has been covered by Harris ( 1961).
Many of the effects of vibration are discussed briefly in books and trade
journals for the particular specialty in which the effect occurs. The handbook
edited by Harris and Crede ( 1961) is, however, remarkably comprehensive in
its coverage of the many problem areas of shock and vibration.
REFERENCES

Standards
ANSI S3.4-1968, Computation of the Loudness of Noise
ANSI S3.5-1969, Calculation of the Articulation Index
ISO/R131-1959, Expression of the Physical and Subjective Magnitudes of Sound or
Noise
ISO/R226-1967, Normal Equal-Loudness Contours for Pure Tones

Other
AIHA ( 1966), Industrial Noise Manual, Second Edition, American Industrial Hygiene
Association, 14125 Prevost, Detroit, Michigan 48227.
ASHRAE ( 1967), ASHRAE Guide and Data Book, Systems and Equipment for 1967,
American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc., 345
East 47th St., New York, NY 10017, Chapter 31, pp 377 and 379.

51
 The helmet, can be as simple as a heavy-duty cloth covering the upper
head and ears, or some variation of the hard hat. Generally, the helmet is not
an effective protector, having attenuation characteristics less efficient than
those of ear plugs.
In summary, we can say that accepted devices, such as ear plugs and muffs,
potentially provide virtually the same amount of attenuation. The actual
amount of attenuation afforded by these protective devices is primarily
dependent upon how well they are fitted, and how consistently they are
worn. A recent study ( Heftier, 1967) on the effectiveness of ear plugs
revealed that, when loosely inserted, plastic plugs provided only 10-dB
attenuation at low frequencies, and 35-dB attenuation or better at other test
frequencies. It is a fact that the attenuation efficiency of other types of
protectors—not merely plugs—falls off as their positioning and fit become
poorer. But, even a poorly fitted protective device is better than none at all,
and workers jn hazardous environments, who use such devices inconsistently
or not at all, risk permanent hearing damage.
The percentage of employees misusing protective devices is alarmingly
high. Maas ( 1969), Audiologist and Hearing Consultant for Employers, Mu-
tual of Wausau, indicated at one point that little more than 20% of personal-
protection programs could be termed in any way successful. Evidence in
support of this statement lies in the fact that, in spite of the institution of
protective programs, individual hearing records taken across several years
continue to show increasing hearing loss. And the reason protective devices
are not working is not that the devices are inadequate, but that they are being
used improperly, if at all.
There are a number of steps that can be taken to correct this situation.
Most important, the worker must be convinced of the desirability of using
these devices. This conviction can be most effectively instilled by a good
indoctrination program, which clearly spells out the hazards of unrestricted
noise and the utility of personal protective devices ‘k/orn properly.
Second, the worker should be allowed some choice in the selection of the
device he will be wearing. It is essential that for him the device be: comfort-
able, capable of being worn up to 8 hours per day; non-toxic, causing no
irritation or skin reaction when worn; easy to use, store, fit, clean and carry;
and durable, capable of surviving for several years of daily use. If these
conditions are met, personal-protection programs will have a much higher
likelihood of success.

4.5 HEARING MONITORING.

çl'he third and perhaps the most important phase of the hearing-conserva-
tion program involves the ongoing monitoring of employees exposed to noisy
environmenHosey and Powell, 1967). If noise is indeed having an adverse
effect upon the ear,<hen ameaningful record of each worker's hearing across
time should reflect this change.) The measurement of the human hearing
function is called "audiometry," and basically requires that the subject make
judgments about specified auditory stimuli. These judgments can relate to the
least-intense sound that can be detected (absolute threshold), the minimal
detectable difference between two sounds (differential threshold), or the
matching of one sound characteristic to another. Similarly, the stimuli can be
pure tones, white noise, narrow-band noise, or avariety of speech modes. The
choice of which judgment is requested and which stimulus is presented is a

63
D. M. Green and J. A. Swets ( 1966), Signal Detection Theory and Psychophysics, John
Wiley & Sons, Inc., New York.
W. F. Grether ( 1971), " Vibration and Human Performance," Human Factors, Vol. 13,
#3, pp. 203-216.
J. C. Guignard ( 1965), " Vibration," Chapter 29, pp 813-894, in J.A. Gillies, ed., A
Textbook of Aviation Physiology, Pergamon Press, Oxford.
J. C. Guignard ( 1966), " Effects of Vibration on Man," J Env Sci, Vol 9, #4, August
1966, pp 29-32.
J. C. Guignard ( 1971), "Human Sensitivity to Vibration" JSound and Vib, Vol 15, # 1,
8March 1971, pp 11-16.
C. M. Harris, ed. ( 1957), Handbook of Noise Control, McGraw-Hill, New York.
C. M. Harris and C. E. Crede ( 1961), Shock and Vibration Handbook, 3 Volumes,
McGraw-Hill, New York.
W. J. Harris ( 1961), Metallic Fatigue, Pergamon Press, New York.
J. E. Hawkins, Jr. and S. S. Stevens ( 1950), "The Masking of Pure Tones and of Speech
by White Noise,") Acoust Soc Am, Vol 22, # 1, January 1950, pp 6-13.
W. R. Hazard ( 1971), " Predictions of Noise Disturbance Near Large Airports," J Sound
and Vib, Vol 15, #4, 22 April 1971, pp 425-446.
R. K. Hillquist ( 1967), " Objective and Subjective Measurement of Truck Noise," Sound
and Vib, Vol 1, #4, April 1967, pp 8-13.
I. J. Hirsh ( 1952), "The Measurement of Hearinz," McGraw-Hill, New York.
T. F. Hueter and R. H. Bolt ( 1955),"Sonics", John Wiley & Sons, Inc., New York
M. Jahn ( 1965), " Subjektive und Objektive Bewertung von Maschinen Gerauschen,"
Acustica, Vol 16, # 3, 1965/66, pp 175-186.
R. G. Klumpp and J. C. Webster ( 1963), " Physical Measurements of Equal Speech-Inter-
fering Navy Noises," J Acoust Soc Am, Vol 35, #9, September 1963, pp 1328-1338.
C. W. Kosten and G. J. Van Os ( 1962), "Community Reaction Criteria for External
Noises," National Physical Laboratory Symposium No. 12, The Control of Noise, Her
Majesty's Stationery Office, London, pp 373-387.
K. D. Kryter ( 1962), "Methods for the Calculation and Use of the Articulation Index," J
Acoust Soc Am, Vol 34, # 11, November 1962, pp 1689-1697. (" Validation of the
Articulation Index," pp 1698-1702.)
K. D. Kryter ( 1968), "Concepts of Perceived Noisiness, Their Implementation and
Application," JAcoust Soc Am, Vol 43, #2, Feb 1968, pp 344-361.
K. D. Kryter ( 1970), The Effects of Noise on Man, Academic Press, Inc., New York,
Chapter 11 and Part III.
K. D. Kryter and K. S. Pearsons ( 1963), " Some Effects of Spectral Content and
Duration on Perceived Noise Level," ) Acoust Soc Am, Vol 35, #6, June 1963, pp
866-883.
D. C. Lavender ( 1971), " Interpretation of Noise Measurements," J Sound and Vib, Vol
15, # 1, 8March 1971, pp 1-9.
D. P. Loye ( 1956), "The Legal Aspects of Noise Control," Noise Control, Vol 2, #4,
July 1956, pp 56-60.
A. C. McKennel and E. A. Hunt ( 1966), " Noise Annoyance in Central London," Report
S 332. U.K. Government Social Survey, Atlantic House, Holburn Viaduct, London,
E.C.1., U.K.
S. Maten ( 1970), " Program Machine Maintenance by Measuring Vibration Velocity,"
Hydrocarbon Processing, Vol 49, September 1970, pp 291-296.
A. Nakano ( 1966), " Relation between Noise Rating Number and Objective Measurement
of Machine Noise," ) Acoust Soc Ja, Vol 22, #2, March 1966, pp 62-76.
R. H. Nittinger ( 1964), "Vibration Analysis Can Keep Your Plant Humming," Chemical
Engineering, Vol 71, August 17, 1964, pp 152-158.
P. H. Parkin ( 1965), " On the Accuracy of Simple Weighting Networks for Loudness
Estimates of Some Urban Noises," J Sound and Vib, Vol 2, # 1, pp 86-88.
D. L. Parks ( 1962), " Defining Human Reaction to Whole-Body Vibration," Human
Factors, Oct, pp 305-314.
H. O. Parrack ( 1957), "Community Reaction to Noise," Chapter 36 in Harris ( 1957).
A. C. Pietrasanta and K. N. Stevens ( 1958), " Noise Exposure in Communities Near Jet
Air Bases," Noise Control, Vol 4, #2, March 1958, pp 93-100.
I. Pollack ( 1952), " The Loudness of Bands of Noise," J Acoust Soc Am, Vol 24, # 5,
September 1952, pp 533-538.

53
 G. Quietzsch ( 1955), " Objektive und Subjektive Lautstarkemessungen," Acustica, Vol 5,
Akustische.Beihefte #1, pp 49-66.
T. C. Rathbone ( 1939), Power Plant Engineering, Vol 43, # 11, Nov, pp 721-724.
T. C. Rathbone ( 1963), " A Proposal for Standard Vibration Limits," Product Engineer-
ing, Vol 34, Mar 4, pp 68f.
D. W. Robinson ( 1958), "A New Determination of the Equal-Loudness Contours," IRE
Trans on Audio, Vol AU-6, January-February 1958, pp 6-13.
D. W. Robinson ( 1969), "The Concept of Noise Pollution Level," NPL Aero Report AC
38, and " An Outline Guide to Criteria for the Limitation of Urban Noise," NPL Aero
Report AC 39, March 1969, National Physical Laboratory, Aerodynamics Division.
D. W. Robinson ( 1971), "Towards aUnified System of Noise Assessment,"J Sound and
Vib, Vol 14, # 3, Feb, pp 279-298.
D. W. Robinson and R. S. Dadson ( 1956), "A Re-determination of the Equal Loudness
Relations for Pure Tones," BriJ of Appl Phys, Vol 7, May, pp 166-181.
D. W. Robinson, J. M. Bowsher, and W. C. Copeland ( 1963), "On Judging the Noise
from Aircraft in Flight," Acustica, Vol 13, #5, pp 324-336.
D. W. Robinson and L. S. Whittle ( 1964), "The Loudness of Octave Bands of Noise,"
Acustica, Vol 14, # 1, pp 24-35.
W. Rosenblith and K. N. Stevens ( 1953), "Handbook of Acoustic Noise Control" Vol II,
"Noise and Man" WADC Technical Report 52-204, PB111274, June 1953, pp
181-200.
J. Sataloff ( 1957), Industrial Deafness , Blakiston Division, McGraw-Hill, New York.
R. W. Shoenberger and C. S. Harris ( 1971), " Psychophysical Assessment of Whole-Body
Vibration," Human Factors, Vol. 13, # 1, pp 41-50.
I. Schiff ( 1970), "Vibration Analysis - Downtime Insurance," Foundry, Vol 98, Dec, pp
90f.
T. J. Schultz ( 1968), " Noise Criterion Curves for Use with the USASI Preferred
Frequencies,") Acoust Soc Am, Vol 43, # 3, Mar, pp 637f.
M. Shimizu ( 1969), "A Simple Method of Measuring Loudness Level - On Automobile
Noises,"J Acoust Soc Jap, Vol 25, #6, pp 319-324.
S. R. Silverman ( 1947), "Tolerances for Pure Tones and Speech in Normal and Defective
Hearing," Ann of Otology, Rhinology, and Laryngology, Vol 56, September 1947, pp
658-677.
L. J. Sivian and S. D. White ( 1933), "On Minimum Audible Sound Fields," JAcoust Soc
Am, Vol 4, #4, April 1933, pp 288-321.
SAE J6a ( 1965), "Ride and Vibration Data Manual", Society of Automotive Engineers,
New York.
A. Spoor ( 1967), "Presbycusis Values in Relation to Noise Induced Hearing Loss,"
International Audiology, Vol 6, # 1, July 1967, pp 48-57.
K. N. Stevens, W. A. Rosenblith, and R. H. Bolt ( 1955), " A Community's Reaction to
Noise: Can it Be Forecast?," Noise Control, Vol 1, # 1, January 1955, pp 63-71.
S. S. Stevens, editor ( 1951), Handbook of Experimental Psychology, John Wiley 8c Sons,
Inc., New York.
S. S. Stevens ( 1956), "The Calculations of the Loudness of Complex Noise," J Acoust
Soc Am, Vol 28, #5, September 1956, pp 807-832.
S. S. Stevens ( 1961), " Procedure for Calculating Loudness: Mark VI," JAcoust Soc Am,
Vol 33, # 11, November 1961, pp 1577-1585.
S. S. Stevens ( 1972), " Perceived Level of Noise by Mark VII and Decibels ( E)",
J Acoust Soc Am, Vol 51 #2 Part 2, Feb, pp 575-601.
S. S. Stevens, J. Volkmann and E. B. Newman ( 1937), "A Scale for the Measurement of
the Psychological Magnitude Pitch," JAcoust Soc Am, Vol 8, #3, Jan, pp 185-190.
S. S. Stevens and H. Davis ( 1938), Hearing, John Wiley & Sons, Inc., New York.
S. S. Stevens and F. Warshofsky ( 1965), Sound and Hearing, Time, Inc., New York.
S. A. Tobias ( 1961), " Machine-Tool Vibration," Chapter 40 in Harris and Crede ( 1961).
R. A. Waller ( 1969), " Office Acoustics- Effect of Background Noise," Applied Acous-
tics, Vol 2, #2, April, pp 121-130.
J. C. Webster ( 1965), " Speech Communications as Limited by Ambient Noise," JAcoust
Soc Am, Vol 37, #4, April, pp 692-699.
J. C. Webster ( 1969), " Effects of Noise on Speech Intelligibility,"Proc Conference Noise
as a Public Health Hazard, Washington, DC, June 13-14, 1968, ASHA Reports 4, The
American Speech and Hearing Association, Washington, DC, 1969, pp 49-73.

54
B. L. Welch and A. S. Welch, ed. ( 1970), Physiological Effects of Noise, Plenum Press,
New York.
R. J. Wells ( 1969), "A New Method for Computing the Annoyance of Steady State
Noise versus Perceived Noise Level and Other Subjective Measures," J Acoust Soc
Am, Vol 46, # 1 ( Part 1), July, p 85.
R. W. Young ( 1958), " Don't Forget the Simple Sound-Level Meter," Noise Control, Vol
4, # 3, May, pp 42f.
R. W. Young ( 1964), " Single-Number Criteria for Room Noise, J Acoust Soc Am, Vol
36, #2, Feb, pp 289-295.
R. W. Young ( 1965), " Re-Vision of the Speech- Privacy Calculation,"J Acoust Soc Am,
Vol 38, #4, Oct, pp 524-530.
R. W. Young ( 1968) " Effective Duration of an Aircraft Flyover," Reports of the 6th
International Congress on Acoustics, Paper F-3-11, pp F-109-112, Tokyo.
R. W. Young and A. P. G. Peterson ( 1969), "On Estimating Noisiness of Aircraft
Sounds,"J Acoust Soc Am, Vol 45, #4, April, pp 834-838.
E. Zwicker ( 1960), " Ein Verfahren zur Berechnung der Lautstarke," Acustica, Vol 10,
Akustische Beihefte #1, pp 304-308.
E. Zwicker, G. Flottorp, and S. S. Stevens ( 1957), "Critical Bandwidth in Loudness
Summation,"J Acoust Soc Am, Vol 29, # 5, May, pp 548-557.

55
t•
Chapter 4

Hearing- Conservation
Programs in Industry
By
Rufus L. Grason and Carol Hetzel
(Grason-Stadler Company, Inc.)
4.1 INTRODUCTION.
Noise has been recognized as a contributing factor in hearing loss for
hundreds of years. Chadwick ( 1963) notes, from about the 12th century,
references citing noise-induced hearing losses. Fosbroke ( 1831), for example,
states: "The blacksmiths' deafness is a consequence of their employment; it
creeps on them gradually, in general at about forty or fifty years of age. At
first the patient is insensible of weak impressions of sound; the deafness
increases with aringing and noise in the ears, ..."
In 1831 it was also recognized that these comments were not new.
Fosbroke continues with: " It has been imputed to a paralytic state of the
nerve, occasioned by the noise of forging, by certain modern writers, and by
the old writers, to permanent over-tension of the membrane, ..." Not until
about a century later, however, was this awareness substantiated by data that
defined specific characteristics of hazardous noise and indicated how these
characteristics might be controlled. The consequence of such careful defini-
tion has been twofold:
1. the increasingly common establishment of industrial hearing-conserva-
tion programs, and
2. the introduction of federal and state legislation that provides financial
compensation for noise-induced occupational deafness.
This chapter defines the hazardous properties of noise, as understood on
the basis of current data, their effect on the physiology of the ear, and the
manner in which both hearing conservation and legislative programs are
working to protect industrial workers.

4.2 THE HUMAN EAR.

4.2.1 Anatomy. Anatomically ( Davis and Silverman, 1970), the ear is

divided into three sections—the outer, the middle, and the inner ears—through
which air-conducted sound waves must travel in order for " hearing" to occur.
The outer ear consists of the fleshy appendage attached to the head and the
ear canal, both of which serve to channel sound waves toward the elastic
tympanic membrane commonly known as the eardrum. The conically shaped
tympanic membrane transforms the energy of sound waves into mechanical
energy of the middle-ear ossicles, a set of small bones. The ossicle chain,
acting as an impedance transformer, transmits the vibrations of the tympanic
membrane to the oval window. This window moves in and out, much like a
piston, generating pressure waves in the perilymph, anearly incompressible
fluid in the inner ear. The pressure differential that results moves the basilar

57
membrane and the organ of Corti. The hair cells in the organ of Corti
transform the mechanical motions into nerve impulses, which are transmitted
through the eighth nerve into higher centers in the brain, where they are
decoded and interpreted as sound.
4.2.2 Effects of Noise. The description above indicates that sound is
transmitted to the brain first by conductive and then by neurological means.
Conductive mechanisms—movement of membranes, bones, and fluid—propa-
gate the sound waves from the external through the middle and inner ears. In
the inner ear proper, direct conductive stimulation of the tiny nerve receptors
translates the previously mechanical activity to an electrical, or neurological,
activity.
<The importance of making this distinction between conductive and neuro-
logical transmission is simply that noise tends zincipally to damage the
neurological auditory mechanisms—the hair cells/ Most frequently, noise-
induced injury first occurs to the outer- and inner-hair-cell structure; then, if
hazardous noise conditions persist, the organ of Corti itself is destroyed. The
consequence of such damage is that nerve cells that would have transmitted
the auditory signal degenerate; they will never regrow and cannot be replaced.
Severe damage from noise exposure, then, is permanent. The one exception
to this rule is that short, intense blasts of noise can rupture an eardrum,
dislodge a bone in the middle-ear chain, or otherwise damage a conductive
mechanism. This acoustic trauma is often temporary and in many cases can
be repaired by surgery or, if simple enough, will heal itself in time.

4.3 NOISE- INDUCED HEARING LOSS.

4.3.1 Data Sources.

Field Studies. In recent years, a great deal of research has been carried
out defining properties of noise hazardous to human hearing. Many studies
have been made in actual industrial environments where the hearing
of the workers was assessed over the course of many months and after
exposure for many years ( Rosenblith, 1942; Ruedi and Furrer, 1946; Cox et
al, 1953; ANSI Z24-X-2-1954; Rudmose, 1957; Kylin, 1960; Glorig et al,
1961; Nixon et al, 1961; Schneider et al, 1961; Yaffe and Jones, 1961; Burns
et al, 1964; Harris, 1965; Taylor et al, 1965; Baughn, 1966; Atherley et al,
1967; La Benz et al, 1967; Noweir et al, 1968; Botsford, 1969; Cohen et al,
1970; Kronoveter and Somerville, 1970). These studies by their very nature,
could not be controlled to yield well-defined statements regarding the exact
relations between various noise parameters and hearing loss. What such
studies did yield, however, were aseries of general principles that laboratory
experiments later were to define in detail:
. 1. Low-frequency noise energy tends to be less damaging to hearing than
mid-frequency noise.
2. Beyond certain levels, increased intensity and increased exposure time
produce increased hearing loss, and an increased fraction of the exposed
group will have significant losses.
3. Individuals show a differential susceptibility to noise-induced hearing
loss.
4. Hearing loss due to noise is most pronounced in the region near 4000
Hz, but spreads over the frequency range as exposure time and level
increase.

58
 Laboratory Studies. A laboratory-simulated industrial environment, unlike
actual industrial situations, has the advantage of permitting the specification
and control of noise parameters. The measurement of the effects of noise,
however, is potentially difficult, as it would seem to require that apermanent
hearing loss be inducpd in the subject. Fortunately, an alternative procedure
has been developed: 4t is now generally accepted that there are two types of
hearing losses—temporary and permanent—both indicated as an increase in an
individual's threshold>(i.e., the intensity level at which he just detects the
presence or absence of an audio signal). When a hearing loss occurs for a
relatively brief period of time following exposure to noise, a temporary
hearing loss or temporary threshold shift ( TTS) is said to have occurred. When
the effect of the exposure is to permanently lessen an individual's sensitivity
for hearing, a permanent threshold shift ( PTS) has occurred. On the basis of
recent evidence, many investigators assume certain similarities between TTS
and PTS ( Glorig, 1967; Kryter et al, 1966). Such an assumption has the
practical advantage of permitting laboratory studies to be carried out on
various parameters of noise, using TTS as an indication of the long-term
effects of noise. Typical experiments set up along these lines commonly
involve one parameter of noise—its level, duration, or spectrum, etc— being
varied while all other parameters are held constant. The resulting characteris-
tics of the TTS, its duration, extent, rate of growth and rate of recovery, are
used as indicators of the character of PTS that would result from prolonged
exposure to the same variable.
On the assumption that TTS and PTS are related, a measurement of TTS
for any of the varied and complex noises that occur in practice could be a
useful guide to the potentially hazardous nature of these noises. Rather than
exposing astatistically useful number of subjects to these noises and observe
their TTS's, one can simulate TTS behavior by an electrical model that
responds in accordance with the experimentally observed TTS behavior
(Keeler, 1968). Such devices have been built ( Benson et al, 1964; Botsford
and Laks, 1970; Moser, 1970) and used on asmall scale.

4.3.2 Hazardous Properties of Noise. Using data from studies on PTS and
TTS, investigators ( see below) have isolated several characteristics of noise as
contributors to the destruction of neurological elements in the ear, and
consequent loss of hearing. Of these characteristics, at least the following are
critical: over-all noise level, the spectrum of the noise, total exposure dura-
tion, and the temporal distribution of the noise.
Over-all Sound Level of the Noise Spectrum. Since noise level is only one
parameter involved in hearing loss produced by noise exposure, only a
preliminary rating of the potential danger is possible with asingle noise-level
reading .'As ageneral rule, however, it can be stated that noise whose over-all
A-weighted sound level is below 80 dB ( Ward, 1969) is probably reasonably
safe> Since many industrial noise levels are greater than 80 dB ( A), additional
information about the character of the noise is required before any statement
regarding its effects on hearing can be made.
Shape of the Noise Spectrum. Laboratou studies have indicated that TTS
is afunction of the spectrum of the noisee indicated earlier that the ear is
most sensitive to frequencies above 1kHz, and that hearing losses occur more
readily at these than at other frequencies'>:Ñoise containing concentrated
energy in the mid-to-high frequency regions (octave bands 600-1200 Hz and
above) is more hazardous to hearing than noise containing energy concen-

59
trated below 600 Hz)(Ward et al, 1959)(Evidence also indicates apure tone
at agiven level is more damaging to hearing than aband of noise at the same
level, centered about that frequenc,(Cohen and Baumann, 1964; Ward,
1962).
Cumulative Duration of Noise Exposure. PTS data, as well as data gathered
in TTS studies, indicate that up to a point, as the total duration of noise
exposure increases for any given population, the incidence and magnitude of
the resultant hearing loss also increases.
The noise-induced hearing loss may not continue until total deafness is
induced.<Glorig ( 1961) found that the permanent hearing loss at 4 kHz
stemming from daily noise exposures of 5-8 hours tended to reach a maxi-
mum at about twelve years of exposure.?
Temporal Distribution of Noise. The relation between intermittent noise
and hearing loss is not clearly defined. One simple rule that is approximately
correct over a limited range is based on the premise thatlegardless of how
energy in noise is distributed across time, its net effect upon threshold shift is
afunction of total energy)This relation is reflected in the recommendations
of the American Academy of Ophthalmology and Otolaryngology ( AA00)
Committee on Conservation of Hearing ( 1969).
In their recommendations, the partial noise exposures are related to the
equivalent continuous A-weighted noise level by equivalent energy summa-
tions. Thus, 90 dB ( A) for 12 hours/week, 95 dB ( A) for 4hours/week, and
100 dB ( A) for 80 minutes/week are each rated as equivalent to 85 dB ( A) for
40 hours/weeknhis relation can also be expressed as that twice the energy ( 3
dB ( A) increase in level) is permissible for every halving of exposure duration,
without increasing the risl¡>
number of laboratory studies indicate that the ear can tolerate signifi-
cantly higher noise levels, when the duration is shortened, than this simple rule
allows (Ward et al, 1959)›ttle Department of Labor for example, permits a
5-dB tradeoff in level for each halving of the duration in its Occupational
Safety and Health StandardeFedera/ Register, 36, 105, May 29, 1971).

4.4 HEARING- CONSERVATION PROGRAMS.

Industrial hearing-conservation programs are designed to protect workers
from the hazardous effects of noise. 4,n order to be effective, a program
should include three areas of concentration: noise assessment, noise reduc-
tion, and hearing assessment>

4.4.1 Assessment of Noise K

A
common method of assessing noise involves
using a sound-level meter (General Radio Type 1565 or similar) set to A
weighting> At this setting, the electrical equivalent of the noise is passed
through a weighting network that has afrequency characteristic like that of
the human ear at threshold, i.e., both are more sensitive at middle frequencies
than at very high or very low frequencies. A single reading on the "A" scale
gives an approximate indication of the noise level as it affects the human ear.
Under these measurement conditions<an SPL of 80 dB ( A) is often regarded
as completely acceptable as the level of continuous noise that should not be
exceeded in the course of an 8-hour day. For enforcement purposes, however,
the Department of Labor has set 90 dB ( A) as the limit for 8 hours of
exposure per day>

60
 As already noted, however, most industrial noise is not uniform and
continuous during the course of an entire dayStvlore commonly, the noise will
be intermittent, or fluctuating in level, or botl> The regulations of the
Department of Labor include rules for determining thquivalent exposure>
The present limits are as follows (
Federal Register, May 29, 1971).
910.95 Occupational noise exposure>
(a) Protection against the effects of noise exposure shall be provided when the sound
levels exceed those shown in Table G-16 when measured on the A scale of astandard
sound level meter at aslow response. When noise levels are determined by octave band
analysis, the equivalent A-weighted sound level may be determined as follows:

140
OCTAVE BAND SOUND PRESSURE LEVEL IN DEC:BELS

A- WEIGHTED SOUND LEVEL

130 125

120

115

110

105

100
100

95

EQUIVALENT
90
90

80 r1 g
) us 1 ' I TS
100 200 500 1000 2000 4000 8000

BAND CENTER FREQUENCY IN CYCLES PER SECOND

Equivalent sound level contours. Octave band sound pressure levels may be converted to
the equivalent A-weighted sound level by plotting them on this graph and noting the
A-weighted sound level corresponding to the point of highest penetration into the sound
level contours. This equivalent A-weighted sound level which may differ from the actual
A-weighted sound level of the noise, is used to determine exposure limits from Table
G-16.

Table G- 16— Permissible Noise Exposures 1 —

Sound level
dBA slow
Duration per day, hours response
8 90
6 92
4 95
3 97
2 --- 100
11
/
2 102
1 105
/
2
1 110
or less 115

When the daily noise exposure is composed of two or more periods of noise exposure
of different levels, their combined effect should be considered, rather than the individual
effect of each. If the sum of the following fractions: CI/T1 + C2/T2 Cn1Tn exceeds
unity, then, the mixed exposure should be considered to exceed the limit value,Cn
indicates the total time of exposure at aspecified noise level, and Tn indicates the total
time of exposure permitted at that level.
Exposure to impulsive or impact noise should not exceed 140 dB peak sound pressure
level.

61
 (b41) When employees are subjected to sound exceeding those listed in Table G-16,
tsible administrative or engineering controls shall be utilized.>f such controls fail to
iuce sound levels within the levels of Table G-16, personal protective equipment shall
provided and used to reduce sound levels within the levels of the table.
(2) If the variations in noise level involve maxima at intervals of 1second or less, it is

6
considered continuous.
In all cases where the sound levels exceed the values shown herein, acontinuing,
Fective hearing conservation program shall be administered.
When the noise level is varying in a simple fashion during the period of
:posurev-5 simple chart record can be prepared to determine the equivalent
:posure from the sound-level meter readingp ( See Primer of Plant Noise
easurement and Hearing Testing, available free from General Radio Corn-
.ny). But if the exposure history is complicatedçfhe GR 1934 Noise-Expo-
re Monitor, specifically designed to determine the percent of noise expo-
re automatically, is particularly helpful in checking compliance with the
gulations,\(See description of monitor in paragraph 6.4).
4.4.2 Noise Reduction. If an analysis of the industrial noise indicates that
hazardous environment exists, a number of steps should immediately be
ken to protect the workers in that area. The primary approach is to reduce
esound levels to be less than those allowed.
Modifying the Source ( See Chapter 9). Protection commonly begins with
tanalysis of the environment to locate the source of noise.ence located,
eoffending device should be modified or replaced with aquieter model> At
le very least, specifications for future equipment to be purchased should be
ritten to attenuate the noise production.
Modifying the Path ( See Chapter 9). If little can be done to alleviate the
rise at its source, then perhaps some modification of the noise path can be
fected.<0ften the distance between the source and the worker can be
creased, or the source or the worker can be enclosed in anoise-attenuating
wrier.'
Protecting the Ear. A third means of implementing_noise reduction in-
)Ives protection of the receptor itself—the human ea&This can be accom-
ished as indirectly as rotating the worker in and out of the noisy environ-
ent, thereby increasing the intermittency of the noise exposure and decreas-
g the possibility of noise damagè) ore directly, such protection can—and
.:ten does—involve anoise-attenuating device worn over or in the ears of the
orker)The particular type of ear protector worn by the worker depends
pon anumber of factors, among them the individual's ear anatomy and his
rimediate working environment. In general, there are three types of ear
rotectors: ear plugs, ear muffs, and helmet.
Ear plugs, made of anonporous pliable material, are designed to fit into
le ear canal. Properly designed, plugs can provide as much as 15-dB attenua-
on at 125 Hz and 40 dB attenuation at 4 kHz. Dry cotton or similar
taterial is often stuffed into the ear canal as asubstitute for well designed
lugs. Contrary to popular belief, such material offers virtually no protection
om noise, and should not be used as a substitute for plugs, unless the
taterial is well impregnated with wax, amodification which greatly improves
seffectiveness.
Ear muffs surround and cover the external ear completely. When designed
ld fitted properly, the ear muff can meet and, at low frequencies, exceed
le attenuation affprded by plugs. However, since these devices surround the
ir and adjacent portions of the face, their attenuation is greatly affected by
tovements of the jaw in chewing, swallowing, and talking.

62
 The helmet, can be as simple as a heavy-duty cloth covering the upper
head and ears, or some variation of the hard hat. Generally, the helmet is not
an effective protector, having attenuation characteristics less efficient than
those of ear plugs.
In summary, we can say that accepted devices, such as ear plugs and muffs,
potentially provide virtually the same amount of attenuation. The actual
amount of attenuation afforded by these protective devices is primarily
dependent upon how well they are fitted, and how consistently they are
worn. A recent study ( Heftier, 1967) on the effectiveness of ear plugs
revealed that, when loosely inserted, plastic plugs provided only 10-dB
attenuation at low frequencies, and 35-dB attenuation or better at other test
frequencies. It is a fact that the attenuation efficiency of other types of
protectors—not merely plugs—falls off as their positioning and fit become
poorer. But, even a poorly fitted protective device is better than none at all,
and workers jn hazardous environments, who use such devices inconsistently
or not at all, risk permanent hearing damage.
The percentage of employees misusing protective devices is alarmingly
high. Maas ( 1969), Audiologist and Hearing Consultant for Employers, Mu-
tual of Wausau, indicated at one point that little more than 20% of personal-
protection programs could be termed in any way successful. Evidence in
support of this statement lies in the fact that, in spite of the institution of
protective programs, individual hearing records taken across several years
continue to show increasing hearing loss. And the reason protective devices
are not working is not that the devices are inadequate, but that they are being
used improperly, if at all.
There are a number of steps that can be taken to correct this situation.
Most important, the worker must be convinced of the desirability of using
these devices. This conviction can be most effectively instilled by a good
indoctrination program, which clearly spells out the hazards of unrestricted
noise and the utility of personal protective devices ‘k/orn properly.
Second, the worker should be allowed some choice in the selection of the
device he will be wearing. It is essential that for him the device be: comfort-
able, capable of being worn up to 8 hours per day; non-toxic, causing no
irritation or skin reaction when worn; easy to use, store, fit, clean and carry;
and durable, capable of surviving for several years of daily use. If these
conditions are met, personal-protection programs will have a much higher
likelihood of success.

4.5 HEARING MONITORING.

çl'he third and perhaps the most important phase of the hearing-conserva-
tion program involves the ongoing monitoring of employees exposed to noisy
environmenHosey and Powell, 1967). If noise is indeed having an adverse
effect upon the ear,<hen ameaningful record of each worker's hearing across
time should reflect this change.) The measurement of the human hearing
function is called "audiometry," and basically requires that the subject make
judgments about specified auditory stimuli. These judgments can relate to the
least-intense sound that can be detected (absolute threshold), the minimal
detectable difference between two sounds (differential threshold), or the
matching of one sound characteristic to another. Similarly, the stimuli can be
pure tones, white noise, narrow-band noise, or avariety of speech modes. The
choice of which judgment is requested and which stimulus is presented is a

63
function of many variables, among them the purpose of the test and the
environment in which it is administered.
In industrial situations, the requirements placed upon the audiometric test
are, first, that it should be simple to administer and easily understood by the
individual taking the test. Second, the test should be asensitive index of the
subject's hearing changes. And finally, the test should be sufficiently well
specified and standardized that the comparison of results is possible. All of
these requirements are met by the pure-tone absolute-threshold test, which is
used extensively in industrial audiometry.

4.5.1 Audiogranni" .The end result of the administration of a pure-tone

threshold test is an altdiogram, apermanent record of the subject's hearing
sensitivity for each of several pure-tone frequencies presented> Sensitivity is
defined in terms of deviation, in dB, from "normal hearing." Kormal hearing,
in turn is defined both by the International Standards Organization ( ISO) and
by the American National Standards Institute (ANSI) and is shown by the
absolute threshold values in dB SPL shown in Table 4-1. These values, based
on 15 different studies of absolute threshold sensitivity in young adults, are
representative of an international agreement on the definition of normal
hearing. Collectively, these SPL values are referred to as the ISO/ANSI zero
reference level, and are used as abasis for uniform calibration of audiometers.

Table 4 1

PURE- TONE REFERENCE THRESHOLD LEVELS*

Frequency dB
(Hz) (re 20 µN/m 2 )
125 45
250 25.5
500 11.5
1000 7
1500 6.5
2000 9
3000 10
4000 9.5
6000 15.5
8000 13

•1964 ISO/1969 ANSI, based on measurements

made on National Bureau of Standards 9-A Coupler
and Telephonics TDH-39 Earphone fitted with
MX-41/AR Cushion. Similar values have been
developed for other sources.

On the audiogram form shown below, the 0-dB horizontal line represents this
zero reference level, which, in turn, is equivalent to the standardized normal
threshold values shown in Table 4-1. Deviations from this line are noted as
dB Hearing Threshold Level ( HTL). A subject, requiring amore intense signal
to reach threshold than normal, will show apositive HTL; one requiring aless
intense signal than normal will show anegative HTL. ( See paragraph 4.7 for a
discussion of factors responsible for deviations around the standardized
normal value.)
It should be noted that, before the current ISO/ANSI Standards defining
normal hearing were accepted, other standards were in effect ( Davis and
Silverman, 1970, Chapter 7). In 1951, the American Standards Association
(ASA) published reference levels that were generally used until 1964 for
audiometric calibration. These two sets of scales- 1951 ASA and 1964

64
 5.2.5 Components. The measured value in a band is sometimes called the
value of a " component." This term is more commonly used for an analysis
that divides the range of interest into a very large number of bands. The
center frequency of a band is used to designate the particular component;
thus, "the component at 120 Hz," " the 120-Hz component", or " the compo-
nent whose frequency is 120 Hz." The term " component" is considered
particularly appropriate if it is expected that the energy in aparticular band is
concentrated in a very narrow frequency region, as often occurs at frequen-
cies that are multiples of the power-line frequency or of arotational frequen-
cy of a motor. Then the resultant analysis may be described as showing
"lines" at certain frequencies.

5.2.6 Conversion of Octave-Band to A-Weighted Levels. Because A-weight-

ed sound levels are so widely used for noise ratings, some may wish to convert
measured octave-band levels to the equivalent A-level when that sound level
did not happen to be measured. This conversion is readily accomplished by
means of Table 5-2, which is used as follows:
1. Add the correction numbers given in the table to each of the corres-
ponding measured octave-band levels.
2. By means of the table in Appendix Iconvert these corrected numbers
to relative power.
3. Add the relative powers of all the bands.
4. Convert back from power to level in dB.
Note that instead of steps 2, 3, and 4 the summing of the corrected levels
can be done in pairs by the chart of Appendix II.
Table 5 2
CORRECTIONS FOR A-WEIGHTED OCTAVE- BAND ANALYSIS

Preferred Series of Octave Bands Older Series of Octave Bands

Band Center
Frequency Original Weighting Octave Band Original Weighting
(Hz) Flat C (Hz) Flat C
31.5 -39.4 -36.4 18.75 - 37.5 -43.4 -39.3
63 -26.2 -25.4 37.5 - 75 -29.2 -28.0
125 -16.1 -15.9 75 - 150 -18.3 -18.0
250 - 8.6 - 8.6 150 - 300 -10.3 -10.3
500 - 3.2 - 3.2 300 - 600 - 4.4 - 4.4
1,000 0 0 600 - 1,200 - 0.5 - 0.5
2,000 + 1.2 + 1.4 1,200 - 2.400 + 1.0 + 1.1
4,000 + 1.0 + 1.8 2,400 - 4,800 + 1.1 + 1.7
8,000 - 1.1 + 1.9 4,800 - 9,600 - 0.4 + 1.9

Table 53
EXAMPLE OF CALCULATlONs•
Octave
Band Band Correction Corrected
Center Level for Level Relative
(Fia) (dB) A-veng. (dB) Power/10 6
31.5 it, - .30 39 .0 I
63 76 -26 50 1.1
125 78 -16 62 1.6
250 82 -9 73 20.0
500 81 -3 78 63.1
1,000 80 0 80 100
2,000 80 + 1 81 125.9
4,000 73 + 1 74 25.1
8,000 65 -1 64 2.5

338.3 1
For the factory noise used previously.
t338 x 10. corresponds to 85.3 dB(A)
77
 employed varies, many automatic audiometers incorporate the Bekesy ( 1947)
technique; which records the subject's threshold estimates continuously,
either as afunction of frequency (sweep-frequency Bekesy), or as afunction
of time (fixed-frequency Bekesy). In atypical industrial testing situation, a
recording pen is moved along the horizontal axis of the audiogram form, on
which several frequencies are located. In the absence of asubject response, an
automatic attenuator associated with the subject switch increases the sound
level and simultaneously moves the recording pen down along the HTL
(vertical) axis of the form. In the presence of a subject response, the
attenuator decreases the sound level and moves the recording pen up along
the chart's vertical axis. The end result of this procedure is that the subject,
instructed to press the switch when he hears the sound and to release when
the sound is not heard, adjusts and simultaneously records his threshold
continuously across time, for each of several frequencies. In addition to the
obvious benefit of operational ease over manually controlled units, automatic
audiometers have the advantage of providing increased test-retest reliability
by presenting astandard test sequence, free from operator intervention and
the consequent likelihood of error. The legal ramifications of such reliability
are quite extensive.
Several automatic audiometers ot the sort mentioned above are currently
available for industrial audiometry. The Grason-Stadler Model 1703 Audiom-
eter ( Figure 4-2), for example, is a fully automatic unit that presents a
fixed-frequency Bekesy test sequence while providing a record of the sub-
ject's responses on an integre X- Y recorder. In addition to standard frequen-
cies, the 1703 tests both ears at 8 kHz, then retests at 1kHz, the latter used
as an indication of test reliability. The 1703 features aunique variable-speed
intensity drive that allows the subject to "home in" rapidly on the threshold
region, then spend more time precisely defining threshold.
If an audiometer is to provide atrue record of an individual's hearing, it
must be calibrated in accordance with specified standards. When the unit is
first manufactured, its many signal parameters—level, purity, rise/fall time,
etc—are all precisely calibrated, often with the aid of wave analyzers and
other electronic instrumentation> Once in the field the audiometer should be
sent to acalibration laboratory once or twice ayear for thorough inspection
and testing, to insure this continued calibration.
Typically, unless audible cues indicate amalfunction in the unit's compo-
nents, adaily field calibration, involving only acheck of signal level, provides
the needed assurance of accurate operation between such annual or semiannu-
al certifications. To facilitate this field calibration, General Radio offers the
Type 1562-Z Audiometer Calibration Set which consists of the Type 1565-B
Sound Level Meter and a choice of couplers that connect the meter to the
audiometer. To ensure accurate readings from the sound-level meter, a
sound-level calibrator is also included. This calibration technique is easy to
implement, can be accomplished in ashort time, and is extremely reliable. Its
frequent use is recommended over alternate methods.
< The validity of threshold measurements is affected not only by the
calibration of the audiometric instrument, but also the test environment.? If
there is ahigh ambient noise level, the test stimuli can be masked, and, as a
consequence, the threshold values of the subject can be erroneously elevated
(Cox, 1955).
(if the audiometric tests are questionable because of high noise levels in the

66
test area, ahearing-conservation program cannot be effectively carried out>
he acceptable levels of background noise for audiometric tests are given
in an American Standard (ANSI S3.1-1960).* For the important frequencies
used in industrial audiometry these levels are given in Table 4-2> where the
center frequency is that of the audiometric test frequency and of the analyzer
bands.
Table 4 2
PERMISSIBLE BACKGROUND NOISE LEVELS
Octave Band Test Room Level (dB) GR 1939 Level Outside GR 1939
Center Freq. Attenuation dB)
(Hz) Max 10 dB Margin (dB) Max 10-dB Margin
500 40 30 46 86 76
1000 40 30 53 93 83
2000 47 37 58 105 95
4000 57 47 61 118 108
8000 67 57 63 130 120

s The AA00 ( 1969) recommends levels that are 10-dB lower than those
---

shown for the octave-band levels, except for the 8000-Hz band, where the
recommended level is 52 dB>
In the selection of a location for audiometric testing, the noise environ-
ment must, therefore, be considered. It is rarely possible to make the
selection on the basis of the noise environment alone, since it is usually
desired to make the tests near the medical or personnel departments. Al-
though these areas are usually relatively quiet, they do not often have alow
enough low-frequency background-noise level, particularly in aplant that has
noise levels high enough to constitute ahazard. In addition, these areas are
often variable in level because of the way that personnel movements fluctu-
ate, the occasional doors that slam, the rearrangements of office and factory
operations, and the changes in production rate or techniques. One might
allow asafety factor for these expected changes, 10 dB is often suggested, in
assessing aparticular space for suitability and long-term adequacy.
<fhe space set aside for audiometric tests might be in an existing room, if
sufficient sound isolation is provided. A more satisfactory procedure is to use
a professionally designed and built prefabricated audiometric room(Hir-
schorn, 1967). Such a room will provide adequate attenuation of noise to
permit location in many areas.
The attenuation characteristics of the GR Type 1939 Audiometric Exami-
nation Room (manufactured by the Industrial Acoustics Company) are
shown in the Table 4-2, along with the permissible background noise levels. If
this room is to be used, an octave-band analysis of the noise at the proposed
site should show levels that are lower than those shown in the right-hand
column.

4.6 PERSONNEL.
There are no firm standards regarding what type of individual is most
appropriate to operate the audiometer used in industrial situations. Naturally,
since automatic audiometers simplify the testing situation, their use means
that less training will be required in order for personnel to administer the test.
Regardless of what type instrument is employed, of prime importance is the

The standard uses the older series of octave bands, but the corresponding levels with
the new standard series of bands should be essentially the same.

67
with sampled points (the crosses) uniformly spaced on the time axis. In the
middle example shown, the period of the sampling is one-fifth that of the
period of the wave; sampling frequency of 5000 Hz = 5X frequency of the
wave ( 1000). One can see that it is not possible to pass asinusoid through the
points shown that has a lower frequency (longer period) than that shown in
the middle.
The lower waveform in the figure is shown sampled at arate five-fourths
the frequency of the waveform, and the upper waveform sampling rate is
five-ninths the frequency of the wave. As shown in the figure we can draw
another lower-frequency wave through the sample points. The waves have
been set up so that all 3 go through equivalent points. If we call the
frequency of the middle wave 1000 Hz, and the sampling frequency 5000 Hz,
the other frequencies are 4000 Hz and 9000 Hz. We could have shown waves
whose frequencies are 6000, 11000, 14000, 16000, 19000, 21000 Hz, etc.,
all of the same peak amplitude and all going through the same points. Since
these cannot be distinguished from one another by the selected set of points,
they are called aliases ( Blackman and Tukey, 1958, p. 167; Bendat and
Piersol, 1966, 278ff). The frequencies of components that are aliases are
related by the equation il l = f2 ±kfs,where f 1 and f
2 are the alias frequen-

cies, k is an integer, and f

s is the sampling frequency.

When these sampled points are treated by digital processing, they are
usually assumed to be from awave of the lowest frequency. If the sampling
rate for an incoming signal is not greater than twice the highest frequency of
any component in the signal, then some of the high-frequency components of
the signal will be effectively translated down to be less than one-half the
sampling rate. This translation may cause serious problems with interference
of high and low-frequency components.
How are these interference effects avoided? Either we sample at a suffi-
ciently high rate to avoid them, or we put in a low-pass filter to reduce the
amplitudes of the higher-frequency components, so that they are no longer
large enough to be troublesome.
In order to see what is required here, consider the filter-response character-
istic, shown in Figure 5-6, which is for alow-pass filter with anominal cutoff
frequency at 5 kHz. This statement merely means that, in asignal applied to
the input of the filter, components with frequencies above 5 kHz are
attenuated compared to those components having frequencies below 5 kHz.

Figure 5-6. Low-pass filter char-

acteristics required to reduce ef-
fects of aliasing.

FREQUENCY -- kHz

80
 <A second and equally important use for the audiometry program is the
implementation of the follow-up tests of hearing ability)lt is imperative that
persons stationed in possibly hazardous noise environments have their hearing
checked,`Isually at 9-12 month intervals after job placementor earlier<if
they have been regularly exposed to noise levels in excess of 80 dB ( A).
Retest should not be postponed until speech discrimination problems are
noted, since noise can affect hearing in the region of 3, 4, and 6 kHz, even
when it does not markedly interfere with hearing for speechAn interval of at
least 16 hours after exposure to high noise levels should elapse before
measuring hearing>This period provides the ear mechanism sufficient time to
recover from the effects of TTS—obviously a prerequisite to making an
accurate estimate of threshold.
Follow-up tests should employ standard threshold-testing procedures used
in the original examinations. By comparing the current record with the
preemployment-threshold audiogram, aqualified tester can estimate what, if
any, hearing losses have occurred. Of course, more sophisticated tests may
have to be made, to differentiate a real permanent hearing loss from a
temporary or false loss—due to inattentiveness or malingering by the sub-
ject—and, in the case of the permanent loss, to pinpoint its locus.
4.7.2 Evaluation of Changes. The final—and in many ways the most
important—aspect of the hearing program comes from the analysis of em-
ployee audiograms. Where hearing losses are indicated, some qualified indivi-
dual—in some cases the technician or physician in the plant, in others, an
outside audiologisr-
-
--must determine the cause of the hearing change.›There
are a host of relevant variables that can be investigated as contributors:
physical variables, such as improper earphone placement or high ambient
noise level; physiological variables, such as the sex of the subject, his age and
general health; psychological variables, such as the subject's motivation and
attention fluctuation; and methodical variables, such as the testing technique
and instructional set used. Finally, there is the very real possibility that none
of these variables can account for the current audiogram and that aperma-
nent hearing loss has occurred. From this point on, it is up to the qualified
professional, the audiologist and his associates, to ascertain the type of
difficulty, its cause and, where possible, its cure. Toward these ends, the data
from the industrial monitoring program will be used—the audiogram, informa-
tion about the character of the noise, and medical history of the individual.
For these reasons, it is essential that this information be kept as current and
complete as possible.

4.8 APPROPRIATE LEGISLATION.

Just as the awareness and specification of the hazardous properties of
noise has increased in recent years, so also has federal and state legislation to
protect employees whose occupation requires that they work in a noisy
environment. A number of states have made legislative provisions to compen-
sate workers suffering from industry-related hearing losses. Wisconsin, one of
the leaders in this area, has had loss of hearing listed as acompensable claim
since 1919 ( Frederickson, 1967). The first case in the state, however, was
filed only in 1951, culminating in aState Supreme Court decision ( 1953) that
allowed the claimant compensation for industrial-noise-induced hearing loss.
This case, now a classic, was one of the first instances in which it was
agreed that a disability was sustained, even though the employee had not

69
missed work and a wage loss had not been sustained. As a result of this
decision, some 500 more claims were filed in ashort period of time until, in
1955, the present section of state legislation pertinent to occupational hear-
ing loss was passed ( Frederickson, 1967). Briefly, this legislation acknowl-
edges the reality of occupational deafness, defined by permanent partial or
total loss of hearing. It further permits the awarding of financial compensa-
tion to employees sustaining such a loss, whether partial or total, provided
that the loss is verified by hearing tests or medical evaluation. This last phrase
emphasizes, once again, the importance of validated audiograms.
The Wisconsin legislation is not unique by any means. AA00 indicates
that over 35 of the 50 states acknowledge the existence of occupational
hearing loss due to continuous noise exposure, and provide compensation for
such loss. In the remaining states, responsibility for hearing loss claims must
be determined by acourt decision.
In addition, Federal legislation has recently been enacted to protect
workers in firms engaged in interstate commerce (see paragraph 4.4.1), and a
number of states have similar regulations.*
These regulations, as well as others proposed, underscore the prevalence of
high-level noise and its damaging effects on human hearing. It also makes it
essential, for the protection of employer and employee alike, that industry-
wide hearing conservation programs be implemented immediately to assess,
reduce, and monitor the effects of noise.
•Cf. American Conference Governmental Industrial Hygienists, Threshold Limit Values
of Physical Agents Adopted by ACGIH, ( 1014 Broadway, Cincinnati, Ohio 45202).

REFERENCES
AA00 ( 1969), " Guide for Conservation of Hearing in Noise", Supplement to Trans
Amer Acad Ophthalmology and Otolaryngology, Revised 1969. (Committee on the
Conservation of Hearing, do Callier Hearing and Speech Center, 1966 Inwood Road,
Dallas, Texas 75325).
AIHA ( 1966), Industrial Noise Manual, Second Edition, American Industrial Hygiene
Association, 14125 Prevost, Detroit, Michigan 48227.
G.R.C. Atherley, W. G. Noble, and D. B. Sugden ( 1967), " Foundry Noise and Hearing in
Foundrymen," Ann Occup Hyg, Vol 10, # 3, July 1967, pp 255-261.
W. L. Baughn ( 1966), " Noise Control-Percent of Population Protected", International
Audiology, Vol V, # 3, Sept 1966, pp 331-338.
G. von Bekesy ( 1947), "A New Audiometer", Acta Oto-Laiyngol, Vol 35 1947, pp
411-422.
R. W. Benson, G. Cook, S. H. Pearsall, and J. Woods ( 1964), " Experimental Investigation
of a Prototype Noise Exposure Meter," Aerospace Medical Research Laboratories,
WPAFB Tech Doc Rept #AMRL-TDR-64-4.
J. H. Botsford ( 1969), " Prevalence of Impaired Hearing and Sound Levels at Work", J
Acoust Soc Am, Vol 45, # I, Jan 1969, pp 79-82.
J. H. Botsford and B. R. Laks ( 1970), " Noise Hazard Meter", Acoust Soc Am, Vol 47,
#1, Jan 1970, p90.
W. Burns, R. Hinchcliffe, and J. S. Little ( 1964), "An Exploratory Study of Hearing and
Noise Exposure in TextileWorkers", An of Occup Hyg, Vol 7, #4, Dec 1964, pp
323-333.
D. L. Chadwick ( 1963), " Noise-Induced Deafness", The Practitioner, Vol 191, # 1146,
Dec 1963, pp 733-741.
A. Cohen, J. R.Anticaglia, and H. H. Jones ( 1970), " Noise Induced Hearing Loss, "Arch
Environ Health," Vol 20, # 5, May 1970, pp 614-623.
A. Cohen and K. C. Baumann ( 1964), "Temporary Hearing Losses Following Exposure
to Pronounced Single Frequency Components in Broad-Band Noise", J Acoust Soc
Am, Vol 36, #6, June 1964, pp 1167-1175.
J. R. Cox, Jr., R. H. Mansur, and C. R. Williams ( 1953), " Noise and Audiometric
Histories Resulting from Cotton Textile Operations", AMA Arch Indus Hyg and
Occup Med, Vol 8, July 1953, pp 36-47.

70
H. Davis, G. T. Morgan, J. E. Hawkins, R. Galambos, and F. W. Smith ( 1950),
"Temporary Deafness Following Exposure to Loud Tones and Noise", Acta Oto-
Laryngol, Suppl 88, 1950, pp 28ff.
H. Davis and S. R. Silverman ( 1970), Hearing and Deafness, third edition, Holt, Rinehart
and Winston, New York.
John Fosbroke, ( 1831), " Practical Observations on the Pathology and Treatment of
Deafness" No. II, The Lancet, 1830-31 Vol I, pp 645-648.
A. Glorig ( 1967), " Some Relations between TTS and PTS," Fourth Annual Symposium
on Noise Effects in Industry, The East Range Clinic, Dept of Indus Med, Sept, 1967.
A. Glorig, W. D. Ward, and J. Nixon ( 1961), "Damage Risk Criteria and Noise-Induced
Hearing Loss", Arch Oto-laryngol, Vol 74, 1961, pp 413-423.
J. D. Harris ( 1965), " Hearing-Loss Trend Curves and the Damage- Risk Criterion in
Diesel-Engine Room Personnel", J Acoust Soc Am, Vol 37, # 3, March 1965, pp
444-452.
A. J. Heffler ( 1967), " Personal Protection from Noise", Fourth Annual Symposium on
Noise Effects in Industry, The East Range Clinic, Dept of Indus Med, Sept 1967, pp
VII 1-16.
M. Hirschom ( 1967), " Acoustical Environments for Industrial Audiometric Programs"
Sound and Vibration, Vol 1, #7, July 1967, pp 8-15.
I. J. Hirsh and R. C. Bilger ( 1955), " Auditory Threshold Recovery After Exposure to
Pure Tones," JAcoust Soc Am, Vol 27, #6, Nov 1955, pp 1186-1194.
A. D. Hosey and C. H. Powell, eds ( 1967), "Industrial Noise, A Guide to its Evaluation
and Control" U.S. Dept. of Health, Education, and Welfare, Public Health Service
Publication # 1572, US Govt Printing Office, Washington, D.C. 20402.
J. S. Keeler ( 1968), "Compatible Exposure and Recovery Functions for Temporary
Threshold Shift- Mechanical and Electrical Models," J Sound and Vib, Vol 7, #2,
March 1968, pp 220-235.
K. J. Kronoveter and G. W. Somerville ( 1970), "Airplane Cockpit Noise Levels and Pilot
Hearing Sensitivity", Arch of Env Health, Vol 20, #4, April 1970, pp 495-499.
K. D. Kryter, W. D. Ward, J. D. Miller, and D. H. Eldredge ( 1966), " Hazardous Exposure
to Intermittent and Steady-State Noise", JAcoust Soc Am, Vol 39, # 3, March 1966,
pp 451-464.
B. Kylin ( 1960), "Temporary Threshold Shift and Auditory Trauma Following Exposure
to Steady-State Noise," Acta Oto-Laryngol, Supplement 152. ( Stockholm).
P. LaBenz, A. Cohen, and B. Pearson ( 1967), "A Noise and Hearing Survey of Earth
Moving Equipment Operators," An: Ind Hyg Assoc J, Vol 28, #2, March/April 1967,
pp 117-128.
R. M. Maas ( 1965), " Hearing Conservation- Legislation, Insurance and Compensation
Claims", Second Annual Symposium on Noise Effects in Industry, The East Range
Clinic, Dept of Ind Med, June 1965, pp 11-12.
R. B. Maas ( 1969), " Personal Hearing Protection: The Occupational Health Nurses
Challenge and Opportunity," Occup Health Nurs, May 1969, pp 25-27.
L. M. Moser ( 1970), " Noise Hazard MeterSased on Temporary Threshold Shift", J
Acoust Soc Am, Vol 48, # 1 ( Pt 1), July 1970, pp 105f.
J. C. Nixon and A. Glorig ( 1961), " Noise-Induced Permanent Threshold Shift at 2000
cps and 4000 cps", J Acoust Soc Am, Vol 33, #7, July 1961, pp 904-908.
M. H. Noweir, A-A El-Dakhakhuy and F. Vale ( 1968), " Exposure to Noise in the
Textile Industry of the UAR", Amer Ind Hyg Assoc J, Vol 29, #6, Nov-Dec 1968, pp
541-546.
W. A. Rosenblith ( 1942), " Industrial Noises and Industrial Deafness", JAcoust Soc Am,
Vol 13, # 3, Jan 1942, pp 220-225.
W. A. Rosenblith, Chmn ( 1954), "The Relations Of Hearing Loss to Noise Exposure",
Jan 1954, ANSI Subcommittee Z24-X-2, New York.
W. Rudmose ( 1957), " Hearing Loss Resulting from Noise Exposure", in C. M. Harris, ed,
Handbook of Noise Control, McGraw-Hill, New York.
L. Ruedi and W. Furrer ( 1946), " Physics and Physiology of Acoustic Trauma,"J Acoust
Soc Am, Vol 18, #2, Oct 1946, pp 409412.
J. T. Sanderson and J. Steel ( 1967), "Noise Induced Hearing Loss in Bench Glass
Blowers," Ann of Occup Hyg, Vol 10, #2, April 1967, pp 135-141.
E. J. Schneider, J. E. Peterson, H. R. Hoyle, E. H. Ode, and B. B. Holder ( 1961),
"Correlation of Industrial Noise Exposures with Audiometric Findings," Amer Ind
Hyg Assoc J, Vol 22, Aug 1961, pp 245-251.

71
W. Taylor, J. C. G. Pearson, A. Moir, and W. Burns ( 1965), "Study of Noise and Hearing
in Jute Weaving,"J Acoust Soc Am, Vol 38 # 1, July 1965, pp 113-120.
W. Taylor, J. C. G. Pearson, R. Kell and A. Moir ( 1967), "A Pilot Study of Hearing Loss
and Social Handicap in Female Jute Weavers," Proc of the Royal Soc of Med, Vol 60,
November 1967, pp 1117-1121.
W. D. Ward ( 1962), " Damage Risk Criteria for Line Spectra," JAcoust Soc Am, Vol 34,
#10, Oct 1962, pp 1610-1619.
W. D. Ward, A. Glorig, and D. L. Sklar ( 1958), " Dependence of rrs at 4 kHz on
Intensity and Time", JAcoust Soc Am, Vol 30, # 1, Oct 1958, pp 944-954.
W. D. Ward, A. Glorig and D. L. Sklar ( 1959), "TTS from Octave- Band Noise-Applica-
tions to Damage Risk Criteria", J Acoust Soc Am, Vol 31, #4 April 1959, pp
522-5?8.
C. D. Yaffe and H. H. Jones ( 1961), " Noise and Hearing: Relationship of Industrial
Noise to Hearing Acuity in aControlled Population," U. S. Dept. of HEW, Pub Health
Ser Publication #850 ( 1961)•

72
 Chapter 5

Analysis
5.1 INTRODUCTION.

Electronic techniques can provide more information about sound or vibra-

tion signals than merely the over-all levels. We can find out how the energy of
asignal is distributed over the range of frequencies of interest, aprocess that
we can describe as analysis in the frequency domain. We can find relations
among signals as a function of time by correlation techniques, and we can
enhance the appearance of coherent elements in asignal, if asynchronizing
trigger is available, by waveform averaging. These two we can class as analysis
in the time domain. We can find the amplitude distribution of asignal, which
shows how often the signal is at any of apossible range of values, and this
process we class as analysis in the amplitude domain.
We shall first discuss analysis in the frequency domain in general terms.
This type of analysis, which has been called "frequency analysis", "wave
analysis", " spectrum analysis", " time-series analysis," and "harmonic analy-
sis", has been widely used for noise measurements. It is invaluable in guiding
one to reduce noise and vibration efficiently (see Figure 5-1). It is also helpful
a.

NOT011-11NAFT
SALANCE
:new

PREGLIEN •
173 IN)

• NOTOM-MAIT
IMBALANCE INKOMNCY
• LOW-SPEED
• IMEMIENC,
1142115)

100 •CO

i•

. , • , • ,..
/1.111.0011CV 1•040111,00 01 0 10 0 100 • MOO

Figure 5-1. Chart records of the

FAN BLADES
vibration acceleration spectrum of a 46 TEETH ( 12)

motor and gear- train assembly (see 9 TEETH

LOW-SPEED RUSHES (2)
sketch). (a) A 1%- band width analysis, SHAFT
COMMUTATOR
taken with the Type 1568- Wave SEGMENTS (( 2)

Analyzerand (b) A 1/3- octave analysis.

ROTOR SPEED
For diagnostic measurements of this 9 TEETH 2gC3/014,RPM

kind, the detail of the spectrum 40 TEETH

ROTOR
SLOTS 02) lose• -
el
provided by the 1% bandwidth is by
far the better.

73
for preventive maintenance. As we have seen in an earlier chapter, it is used in
a number of procedures for estimating the probable effects of noise and
vibration on man.
The development of electronic digital techniques has made instruments for
analysis in the time domain and amplitude domain practical. One aspect of
digital techniques, called "sampling", will be described, because it is helpful
in understanding the concepts of autocorrelation, crosscorrelation, waveform
averaging, and amplitude distribution. After those and some additional pro-
cesses have been described, the implementation of various forms of analysis
and their characteristics will be discussed.
The application of some of these techniques is not so obvious from the
earlier discussion as is spectrum analysis, and some references will be given to
their use as they are discussed.

5.2 ANALYSIS IN FREQUENCY BANDS.

To make an analysis in the frequency domain, the signal energy is electron-
ically separated into various frequency bands, for example, octave bands,
each of which covers a2-to- 1range of frequencies. The analysis yields aseries
of levels, one for each band, called "band levels", or for octave bands,
"octave-band levels" or "octave-band sound-pressure levels". Here it is ap-
parent that the band in which areading of level is obtained must be specified
if the information is to be of value.

5.2.1 Octave Bands. The preferred series of octave bands for acoustic
measurements covers the audible range in ten bands. The center frequencies
of these bands are 31.5, 63, 125, 250, 500, 1000, 2000, 4000, 8000, and
16,000 Hz. The actual nominal frequency range of any one of these bands is
2-to- 1; for example, the effective band for the 1000-Hz octave band extends
from 707 to 1414 Hz.
Another series of octave bands has been widely used in the past. The
older bands were a 75-Hz low-pass unit, and the octave bands of 75 to 150,
150 to 300, 300 to 600, 600 to 1200, 1200 to 2400, 2400 to 4800, and
4800 to 9600 Hz, but these are no longer preferred, according to American
National Standards. This older series is still specified in a number of test
codes, however, and the published data obtained with this series is extensive.*

...

-- TOTAL
TOTAL NOISE

-ro Figure 5-2. A plot of the

octave- band analysis of
noise from a calculating
io
machine,(Graphic paper for
BACK./
plotting octave band levels
GROOM
is available from Codex
50
Book Co. Inc., Norwood,
Mass., as Forms 31464 and
90
BACKGROUND NOISE
31460 for the preferred
octaves and the older ser-
ies, respectively.)
30.
OVER-AU. 63 125 250 500 1000 2000 4000 8000
OCTAVE BAND CENTER FREQUENCY

*A method for converting octave- band levels measured with this older series to levels for
the new series is given in Appendix A of ANSI S1.11-1966, American National Standard
Specification for Octave, Half- Octave, and Third-Octave Band Filter Sets.

74
 When agraph is made of the results of octave-band pressure level measure-
ments, the frequency scale is commonly divided into equal intervals, between
the position designated for each band and the position for the band adjacent
to it in frequency. The pressure level in each band is plotted as a point on
each of these positions along the other axis. Adjacent points are then
connected by straight lines. An example of a plot of this type is given in
Figure 5-2. An alternative presentation uses horizontal lines centered on the
band at the measured level.

5.2.2 One-Third-Octave Bands. For more detailed analysis of the distribu-

tion of sound energy as a function of frequency, still narrower bands are
used. The next popular division is asplit of the octave into three parts. This
choice is based partly on the fact that ten such filters can be arranged
effectively to cover a 10-to- 1frequency range. The preferred center frequen-
cies for such a series would be, for example, 100, 125, 160, 200, 250, 315,
400, 500, 630, and 800 Hz. The next 10-to- 1set would start with 1000 Hz as
the center frequency and continue by multiplying each number by 10, 100
and so on ( 1000, 1250, 1600, 2000 . ..). Similarly, lower preferred frequen-
cies are obtained by adivision of 10, 100, etc. For practical reasons the usual
span of third octaves for acoustic noise analysis runs from 25 to 10,000 Hz.
The actual effective band for aone-third-octave filter at 1000 Hz extends
from about 891 to 1122 Hz. That is, the bandwidth is about 23% of the
center freauencv.
Third-octaves vs. octaves. When we wish to compare a one-third-octave
analysis and an octave analysis, it is best to combine the one third octaves, in
groups of three to get equivalent octaves. Thus, for example, to find the
equivalent 1000-Hz octave-band level, combine the third-octave levels at 800
Hz, 1000 Hz, and 1250 Hz. Suppose the levels are 74.5, 73.0 and 71.0 dB.
These levels can be converted to relative power, summed, and then translated
back to level. Or we can use the chart of Figure 2-4 to combine them. The
combination of 74.5 and 73.0 is 74.5 + 2.3 = 76.8. This result combined with
71.0 is 76.8 + 1.0 = 77.8 dB.

5.2.3 Narrower Bands. Analyzers that use third-octave and octave band-
widths are widely used in acoustics, but still narrower bands are essential for
some purposes. One-tenth-octave ( 6.9%), one-twelfth-octave ( 5.8%), one fif-
Yee.

743

50

40

Figure 5-3. Plot of one

30
third-octave analysis of
noise from calculating ma-
chine. (Codex Book Co.,
Form 31462.)

10:0
FREQUENCY IN HERTZ

75
 0.5/e

04/o

0.3/s

0.2/o

o
-4o - 3o
3o 4o

Figure 5-14. The amplitude density distribution p ( y) of Gaussian random noise.

100

075

0.50

0.25

o
-4o - 3a -o 2o 3o 4u

Figure 5-15. The amplitude distribution function P (y) of Gaussian random noise.

If the signal being sampled is a periodic signal, even fewer observations of

he signal are required for agood estimate of the rms value. But some unusual
ampling situations must be avoided. Thus, the sampling rate must not have
he same period as the period of the signal or some integer multiple of that
)eriod. When sampling procedures are used in the process of determining the
ms value of awave that may be periodic, these conditions are avoided by one
ef the following:

1. Sampling at arandom rate.

2. Sampling at arate that changes according to some rule.
3. Sampling with aperiod less than half the period of the signal.
The measurement of amplitude distribution, in a practical case, is not
'early so useful as a spectrum measurement and it is rarely done (Piersol,
967). It is helpful, however, to explain it to bring in various ideas and
echniques that are used in other applications.

88
 5.2.5 Components. The measured value in a band is sometimes called the
value of a " component." This term is more commonly used for an analysis
that divides the range of interest into a very large number of bands. The
center frequency of a band is used to designate the particular component;
thus, "the component at 120 Hz," " the 120-Hz component", or " the compo-
nent whose frequency is 120 Hz." The term " component" is considered
particularly appropriate if it is expected that the energy in aparticular band is
concentrated in a very narrow frequency region, as often occurs at frequen-
cies that are multiples of the power-line frequency or of arotational frequen-
cy of a motor. Then the resultant analysis may be described as showing
"lines" at certain frequencies.

5.2.6 Conversion of Octave-Band to A-Weighted Levels. Because A-weight-

ed sound levels are so widely used for noise ratings, some may wish to convert
measured octave-band levels to the equivalent A-level when that sound level
did not happen to be measured. This conversion is readily accomplished by
means of Table 5-2, which is used as follows:
1. Add the correction numbers given in the table to each of the corres-
ponding measured octave-band levels.
2. By means of the table in Appendix Iconvert these corrected numbers
to relative power.
3. Add the relative powers of all the bands.
4. Convert back from power to level in dB.
Note that instead of steps 2, 3, and 4 the summing of the corrected levels
can be done in pairs by the chart of Appendix II.
Table 5 2
CORRECTIONS FOR A-WEIGHTED OCTAVE- BAND ANALYSIS

Preferred Series of Octave Bands Older Series of Octave Bands

Band Center
Frequency Original Weighting Octave Band Original Weighting
(Hz) Flat C (Hz) Flat C
31.5 -39.4 -36.4 18.75 - 37.5 -43.4 -39.3
63 -26.2 -25.4 37.5 - 75 -29.2 -28.0
125 -16.1 -15.9 75 - 150 -18.3 -18.0
250 - 8.6 - 8.6 150 - 300 -10.3 -10.3
500 - 3.2 - 3.2 300 - 600 - 4.4 - 4.4
1,000 0 0 600 - 1,200 - 0.5 - 0.5
2,000 + 1.2 + 1.4 1,200 - 2.400 + 1.0 + 1.1
4,000 + 1.0 + 1.8 2,400 - 4,800 + 1.1 + 1.7
8,000 - 1.1 + 1.9 4,800 - 9,600 - 0.4 + 1.9

Table 53
EXAMPLE OF CALCULATlONs•
Octave
Band Band Correction Corrected
Center Level for Level Relative
(Fia) (dB) A-veng. (dB) Power/10 6
31.5 it, - .30 39 .0 I
63 76 -26 50 1.1
125 78 -16 62 1.6
250 82 -9 73 20.0
500 81 -3 78 63.1
1,000 80 0 80 100
2,000 80 + 1 81 125.9
4,000 73 + 1 74 25.1
8,000 65 -1 64 2.5

338.3 1
For the factory noise used previously.
t338 x 10. corresponds to 85.3 dB(A)
77
 5.3 TIME SERIES AND SAMPLING.
5.3.1 Time Series. An acoustic signal that is monitored at apoint in space
can be considered as a variation in pressure as a function of time. This
variation is acontinuous function that we transform into asimilar electrical
function by the use of a microphone. We can operate on this electrical signal
as acontinuous function by electronic circuits, and we call this " using analog
techniques."
It is also useful to convert the function into aseries of discrete numerical
values, which is called a " time series." We can then operate on these discrete
values by digital techniques with acomputer program. This computer may be
one that is easily recognized as such, or it may be built into the system in
such away that its operation is not apparent externally.

• 5.3.2 Sampling. The process of obtaining a series of discrete numerical

values from a continuous function is known as sampling. We use electronic
circuits to observe the instantaneous voltage of the signal at regular intervals.
This instantaneous voltage is then converted into an electrical signal that
represents anumerical value proportional to that voltage.
What requirements do we have to put on this time series so that it is an
adequate representation of the original continuous function? The answer
depends on what information we want from the time series. If we wish to be
assured that we can perform adirect spectrum analysis on the time series, we
can begin with aspecification in the frequency domain. We can see that one
factor is how frequently we sample, compared to how rapidly the signal
changes. The actual rule is that the sampling rate must be at least twice the
highest frequency component in the signal.
The frequency of the signal that corresponds to that minimum sampling
rate has been called the " Nyquist frequency," which is then one-half the
minimum sampling frequency. The corresponding maximum time between
samples, which is the reciprocal of the minimum sampling frequency, is called
the " Nyquist interval". As we shall see later we can sample less frequently for
operations in the amplitude domain. Sampling less frequently has also been
used to transfer information from a limited band at high frequencies to a
low-frequency region, but then, care must be taken to ensure that overlapping
of information does not confuse the process. The basic point here is that
more frequent sampling than at the Nyquist interval is sufficient sampling,
but it is not always necessary when we need only a limited amount of
information from the signal.
• 5.3.3 Quantization. When the sampled value is converted into an equiva-
lent numerical value by an analog-to-digital converter, each value is repre-
sented by a finite number of on-or-off states of electronic elements. The
converter has in effect an input-output relation that is aseries of steps. The
size of the steps depends on the total range to be covered and the number of
steps available. Since the equivalent electrical signal is usually coded in binary
form, the number of steps is then, say 256, 512, 1024, etc., which is 2raised
to some integer exponent.
The exponent of 2 used is called the number of "bits," and a 10-bit
converter would have 1024 discrete values. Since the signal has both positive
and negative values, this number may be cited as 512 values plus asign bit.
Each of the sampled values is now rounded off or quantized to the nearest
number of units in the range available. This quantization leads to an error

78
that may be as much as one-half the quantum step or interval in the
conversion. On the average, however, it will be less than that, and the
equivalent noise contributed by this error has an rms value of about 0.3 times
the interval. Furthermore, this noise is distributed in frequency, and when an
analysis is made, the equivalent noise level will be correspondingly lower in
each band.
An example of small-signal operation is sometimes cited as a limiting
feature of the analog-to-digital converter. Assume that the signal applied to
the converter has atotal voltage excursion that is less than one interval in its
range. Then, the output will remain at azero value, and the signal will not be
recognized. This mode sets one limit on a type of dynamic range. But in
practice such asituation should not ordinarily occur. The converter should be
preceded by electronic amplifiers of sufficient gain so that any signal to be
studied can be brought up to the level required for proper utilization of the
converter. The dynamic range that is then of interest is the level of the
smallest component that can be observed compared to the level of the total
signal, which we shall discuss briefly.
In analysis one is often interested in each of anumber of small compo-
nents of a signal, which may contain one or more dominating components.
What happens to the small components when the signal is sampled? The
answer is that they are preserved but, as described above, quantization noise
appears ( Sloane, 1968). Since this noise is distributed in frequency, when an
analysis into many bands is made, the noise on the average in each band
becomes very small. With ahigh-resolution processor, one can then observe
components that need be only somewhat greater than the noise level in the
band, which may be appreciably less than the total noise level.
The significant reductions in effective quantization noise that occur in
some processing procedures mean that the computing system, used to process
the digital data, can advantageously utilize a significantly higher resolution
than is used in the analog-to-digital converter. It is for this reason that one
can justify the use of acomputer with 16-bit resolution for processing data
from a10-bit converter, for example ( Korn, 1966, Chapter 6; Widrow, 1961).

• 5.3.4 Aliasing and Filtering. In order to appreciate other effects of this

sampling process, it is useful to look at the sampling of some sinusoidal
pressure waveforms. "Waveform" is used here to describe the instantaneous
amplitude as a function of time. In Figure 5-5 three waveforms are shown

Figure 5-5. Waveforms of three

frequencies sampled at the same
rates for digital processing.

79
with sampled points (the crosses) uniformly spaced on the time axis. In the
middle example shown, the period of the sampling is one-fifth that of the
period of the wave; sampling frequency of 5000 Hz = 5X frequency of the
wave ( 1000). One can see that it is not possible to pass asinusoid through the
points shown that has a lower frequency (longer period) than that shown in
the middle.
The lower waveform in the figure is shown sampled at arate five-fourths
the frequency of the waveform, and the upper waveform sampling rate is
five-ninths the frequency of the wave. As shown in the figure we can draw
another lower-frequency wave through the sample points. The waves have
been set up so that all 3 go through equivalent points. If we call the
frequency of the middle wave 1000 Hz, and the sampling frequency 5000 Hz,
the other frequencies are 4000 Hz and 9000 Hz. We could have shown waves
whose frequencies are 6000, 11000, 14000, 16000, 19000, 21000 Hz, etc.,
all of the same peak amplitude and all going through the same points. Since
these cannot be distinguished from one another by the selected set of points,
they are called aliases ( Blackman and Tukey, 1958, p. 167; Bendat and
Piersol, 1966, 278ff). The frequencies of components that are aliases are
related by the equation il l = f2 ±kfs,where f 1 and f
2 are the alias frequen-

cies, k is an integer, and f

s is the sampling frequency.

When these sampled points are treated by digital processing, they are
usually assumed to be from awave of the lowest frequency. If the sampling
rate for an incoming signal is not greater than twice the highest frequency of
any component in the signal, then some of the high-frequency components of
the signal will be effectively translated down to be less than one-half the
sampling rate. This translation may cause serious problems with interference
of high and low-frequency components.
How are these interference effects avoided? Either we sample at a suffi-
ciently high rate to avoid them, or we put in a low-pass filter to reduce the
amplitudes of the higher-frequency components, so that they are no longer
large enough to be troublesome.
In order to see what is required here, consider the filter-response character-
istic, shown in Figure 5-6, which is for alow-pass filter with anominal cutoff
frequency at 5 kHz. This statement merely means that, in asignal applied to
the input of the filter, components with frequencies above 5 kHz are
attenuated compared to those components having frequencies below 5 kHz.

Figure 5-6. Low-pass filter char-

acteristics required to reduce ef-
fects of aliasing.

FREQUENCY -- kHz

80
Assume we are interested only in the range below 5 kHz. If we sampled at a
10-kHz rate, and acomponent at 5.1 kHz was present, it would be equivalent
to one at 4.9 kHz and the filter would not adequately reject the component
at 5.1 kHz. Our processing would lead us to believe there was acomponent at
4.9 kHz and we would be misled. Suppose we sampled at a 15-kHz rate as
shown on the figure. The 5.1 kHz component would be there, but it is
beyond the range of interest, and we would ignore it.
If however, there is a component at 10 kHz, it would appear to be at 5
kHz in the processing, and we might be misled. But now the filter would
attenuate this component by some 66 dB, which would probably make it so
small that we would not be concerned about it. Any components at a
frequency higher than 10 kHz would be attenuated even more and be
correspondingly less important. The components originally at frequencies
between 5 and 7.5 kHz would appear, but with some attenuation, and they
would be ignored, because we assumed interest only up to 5 kHz. Compo-
nents between 7.5 and 10 kHz would be translated into the range between 5
and 7.5 kHz and also be ignored.
It is sometimes easier to think of the component frequencies as being
folded about afrequency equal to one-half the sampling rate. This folding is
illustrated on the figure for the attenuation characteristic.
If we wanted to suppress the extraneous components at 10 kHz and
beyond even more than 66 dB, we could either use afilter that had more
attenuation at those frequencies or go to ahigher sampling rate, or both.
In general, it is advantageous to use filters to limit the frequency range of
the input data, since we can then use alower sampling rate and a minimum
number of points for agiven sample time. These filters when used in this way
are often called, "anti-aliasing filters" or, simply, "aliasing filters." If we have
to process a signal that lasts for asecond and that could have components up
to 12,000 Hz, we would have to sample at more than a 24,000-Hz rate. We
would then have more than 24,000 values to process. If we are interested
only in the range below 1000 Hz, we could use alow-pass filter that starts
attenuating at 1000 Hz. We could then sample at a 3000- to 4000-Hz rate,
and we would save much time in processing.
• 5.3.5 Frame Size. Some digital operations can be applied in a running
fashion to the sampled values as they are produced. But other operations are
done in batches, and it is convenient then to think of aset of points or values
that are processed as a group, and we shall call such aset a "frame." The
word "sample" is also used for such a set, but this usage may lead to
confusion, since " sample" is thought by some to be an individual value.
"Block" is also used for this set of points, but it is convenient to reserve this
word as amore general term for sets of points.
The frame size, which is the number of sampled values in the frame, is
most conveniently set up in powers of 2, that is, 64, 128, 256 etc., for many
of the operations, but particularly for the calculation of the spectrum values
by the fast Fourier transform.
5.4 ANALYSIS IN THE TIME DOMAIN.
5.4.1 Correlation. Correlation is a measure of the similarity of two time
series or waveforms, and it is afunction of the time displacement between the
two ( Lathi, 1965, Chapter 12; Anstey, 1966). If a waveform is compared
with itself by the correlation process, it is called an autocorrelation. If the
waveforms to be compared are distinct, the process is called crosscorrelation.

81
 Original signals

100 13ms
TIME- MILLISECONDS T LAG - 0.1 ms/POINT

Figure 5-7. Auto correlation functions.

• 5.4.2 Autocorrelation. To obtain the autocorrelation of awaveform or

time series we calculate a whole set of averaged products. We multiply
point-by-point that waveform with an identical one, and we take the average
of these multiplied values, over the full range of the wave. If there is no time
displacement, the result is the mean square value, and it is usually termed,
power, although it may not be related to physical power. If the two wave-
forms are shifted in time with respect to each other, another averaged
product can be obtained, and so on for many shifts in time. These averaged
products as afunction of the time displacement, or delay (usually designated
by rand also called "lag"), are the autocorrelation function.
The autocorrelation function is seen to be an extension of the concept of
the mean-square value of awave ( more often used, after the square root has
been taken, as the root-mean-square value). The autocorrelation function is
symetrical about the point of zero delay, and at zero delay it has the
maximum value equal to the mean-square value ( Figure 5-7).
If the waveform has aperiodic component, the autocorrelation function
will show a periodic character. This behavior is illustrated by the example
shown in Figure 5-8, where a noisy signal contains a periodic component,
which is not obvious in the original waveform, but the autocorrelation
function shows it clearly ( Lee, 1960; Heizman, 1970). Some of the important
applications of autocorrelation are based on this property.
• 5.4.3 Cross-correlation. When the cross-correlation function is calculated,
we have a measure of the similarity of the two waveforms used in the
calculation. Now the maximum value does not always occur with zero delay,
but the time at which the maximum value occurs may be significant. As

Auto.correlation function of original si.nal

I00
TIME - MILLISECONDS r LAG - 0.1 ms/POINT

Figure 5-8. Auto correlation analysis of sinewave buried in noise.

82
shown in Figure 5-9 for the two waveforms there correlated, they are most
similar when a delay of 0.7 millisecond is used. This time may be an
important clue in tracking down the source of adisturbance.
The waves shown are sampled waves. The original waves can be thought
of as those shown, with successive points connected by asmooth continuous
line.
Applications of Correlation Techniques. Correlation techniques have been
applied to determine which among a number of sources is contributing most
significantly to the noise at agiven point ( Goff, 1955); to locate noise sources
by direction finding ( Faran and Hills, 1952b, Gilbreck and Binder, 1958), to
separate aerodynamic noise of turbulent boundary layers from noise radiated
by jet engines ( Bhat, 1971), to measure panel-transmission loss without
requiring that flanking signals be eliminated ( Goff, 1955; Burd, 1964; Imai;
1968; Burd, 1968); to reduce the contribution of wind noise to a measure-
ment of noise from a device ( Goff, 1955; Burd, 1964); to measure the
transient response of rooms ( Goff, 1955); to measure the diffuseness of sound
in aroom ( Cook et al, 1955; Balachandran and Robinson, 1967/68), and to
measure sound absorption ( Goff, 1955; Burd, 1964). The measurements of
transmission loss and sound absorption are limited in frequency range by the
size of the object and the bandwidth of the applied noise signal. They are also
measurements by a particular path and not the averaged random-incidence
measurements of the standard methods. They are, therefore, more suited for
research studies, for tests of materials in place or for tracking down difficul-
ties rather than for rating procedures.
The application of correlation to signal detection in underwater echo
ranging and signal transmission has been extensively investigated ( Faran and
Hills, 1952a; D. Middleton, 1960; Horton, 1969; Tolstoy and Clay, 1966; and
many papers in the,/ Acoust Soc Am).
Correlation has been used to study the vibrations on the two sides of a
panel in panel sound-transmission research ( Nakamuru and Koyasu, 1968), to
study the flow of energy in structures ( White, 1969), and to analyze seismic
vibrations in geophysical exploration ( Anstey, 1966).

• 5.4.4 Convolution and Superposition. A digital operation called " convolu-

tion" is sometimes used in the time domain ( Lathi, 1965, Chapter 10; Lee,
1960). If one waveform is convolved with another, one of them is reversed in
time or folded back. Corresponding ordinates are multiplied and the products
added. A whole series of these sums are obtained for different positions of
the two waves, and one has aresult or output that is a function of the time
displament of the two waves. Except for the reversal in time, the operation
is the same as cross-correlation.

001 13.ns
r LAG — OI ,,, t/POINT

Figure 5-9. Cross-correlation function of the original signals of Figure 5-7.

83
 If one of the waveforms is the input signal, the convolution can be used to
produce arunning average or smoothing of the wave or it can do adifferenc-
ing operation, depending on the second function used in the convolution.
The operation can also be looked upon as afiltering of the input wave-
form. In this process the second function is known as the impulse response of
the filter. It corresponds to the waveform that results when the filter is
stimulated by an idealized impulse of finite energy but zero duration. The
filtered output waveform is thus obtained by a convolution of the input
waveform with the impulse response of the filter. The relations among the
corresponding filter operations in the time domain and the frequency domain
are shown in Figure 5-10.
Convolution in this type of operation is also superposition, which is more
readily understood. If an input wave is assumed to be acollection of impulses
occurring at successive instants of time, we can obtain the output wave by
superposition of the individual responses to the impulses.
Each impulse response normally decays as time goes on. Thus, at any one
instant the contribution to the output of preceding elements in the input
wave is reduced, the farther back in time that we go. The output at any
instant is a weighted sum of the past input. The impulse response is the
weighting function.

• 5.4.5. Waveform Averaging. Summation Analysis . In the study of noisy

signals that include periodic components or that are responses to stimuli, the
periodic component of the wave, or the evoked response, can be emphasized
with respect to random noise or stray signals by aprocess known as waveform
averaging (Geisler and Rosenblith, 1962; Clark et al, 1961; Nelson and
Lassmann, 1968; Rothman, 1970). It is essential in this process that a
reference or triggering signal be available.
This process is a simple summing of corresponding ordinates of selected
samples of the wave. Because of this summing, waveform averaging is also
called summation analysis. Often it is left as a sum, but a division by the
number of selected samples is necessary to convert to an average. If we state
the operation in sampled terms, it might go like this: We select and store a
frame of data as determined by the reference or trigger signal. When the
trigger initiates the sampling of another frame, the first of the sampled values
of the new frame is added to the first sampled value of the original frame.

FREQUENCY
DOMAIN

TIME -.-
DOMAtN

Figure 5-10. Graphical representation of interrelationships

between time and frequency functions.

84
• Example of Random-Noise Measurement. To illustrate by an actual numer-
ical example the type of behavior that occurs, some measurements were made
of an arbitrary level of arandom-noise generator in the octave band from 150
to 300 Hz. With the fast meter speed, the average of the fluctuating levels
indicated on the meter was estimated to be about + 5dB, where in aperiod of
30 seconds the level fluctuated from aminimum of + 3.3 dB to amaximum of
+6.5 dB, arange of 3.2 dB. In the slow position the estimated level was +4.7
dB, and the level fluctuated over athree-minute period from a minimum of
+3.8 to a maximum of +5.7, arange of 1.9 dB. Some sample readings were as
follows: fast position: 4.8, 4.1, 5.3, 3.7, 5.8, 4.9, 5.3, 5.2, 6.2, 4.6; slow
position: 4.4, 5.1, 3.9, 4.9, 4.2, 5.0, 4.7, 4.1, 4.3, 4.9. (These sample readings
were taken with the help of astroboscope, to avoid observer bias in selecting
readings and to make it possible to take definite readings on the rapidly
moving pointer in the fast position.) One hundred samples were taken for
each position. The average value on an energy basis for slow was +4.72, with
the lowest reading + 3.8 and the highest + 5.8. The average for fast was +4.74,
with alow reading of + 3.1 and 2, high reading of +6.2.
Taking such aset of readings is not the usual way to obtain the indicated
level; rather, one estimates avalue by observing the pointer fluctuations. But
these discrete samples permit one to describe statistically the behavior that
can be expected.
For the fast position one would expect only 1 in 1000 readings to differ
from the average by more than about - 3dB or +2.4 dB, arange of 5.4 dB.
The corresponding extremes for one chance in 100 is about - 2.3 dB or + 1.9
dB, a range of 4.2 dB; for 1 in 10, about - 1.4 to + 1.2, arange of 2.6 dB.
Note that the range is not symmetrical.
These statements about variability can be expressed in another way, which
is the comerse of that above. If any reading is taken in the fast position, the
chances are only 1 in 100 that the long-time average value of the noise is
below the observed value by more than 1.9 dB or above the observed value by
more than 2.3 dB. These limits are called the 99% confidence limits.
Confidence Limits for Octave Bands. A chart of the 99% confidence limits
for octave bands for random noise measurement is given in Table 5-4.
These ranges of uncertainty can be reduced by the use of the average of a
number of independent readings. The reduction in the range is approximately
inversely proportional to the square root of the number of independent
observations. Thus, the average of four observations would reduce the uncer-
tainty to about one-half that shown.
Table 5-4
CONFIDENCE LIMITS FOR RANDOM
NOISE IN OCTAVE BANDS
99% Confidence Limits (dB)
Meter Speed
Center Freq
(Hz) Fast Slow
31.5 -4.2, +7.0 -2.5, +3.3
63 -3.2, +4.7 -1.8, +2.2
125 -2.4, +3.1 -1.3, + 1.5
250 -1.7, +2.1 -1.0, + 1.1
500 -1.2, + 1.4 -0.7, +0.7
1,000 -0.9, + 1.0 -0.5, +0.5
2,000 -0.6, +0.7 -0.3, +0.3
4,000 -0.5, +0.5 -0.2, +0.2
8,000 -0,3, +0.3 -0.2, +0.2
16,000 -0.2, +0.2 -0.1, +0.1

101
1.0 and > 1.0. A similar series is used for the negative values of sound
pressure (note that sound pressure is the change in pressure with respect to
the ambient average atmospheric pressure).
After these timers have been running for awhile, we record the running
total for each. We divide each value by the total time for all intervals and plot
the results. Suppose we get the plot shown in Figure 5-11. This type of plot is
often called a " histogram" or a "frequency distribution" ( here "frequency" is
used in the sense of "frequency of occurrence" and not "frequency in Hz").
Someone familiar with these plots would guess from Figure 5-11 that the
sound-pressure wave that was observed was probably essentially sinusoidal in
form. ( Actually one cannot be definite about such an observation, since an
infinite variety of waveforms could produce the same distribution.) If we
refined the interval resolution to the limit, we could obtain the characteristic
curve of Figure 5-12 for a sinewave. This plot is an amplitude-density
distribution. The area under the wave between any two values is the fraction
of the total time that the instantaneous pressure has avalue that falls within
that interval.
If we measured the amplitude-density distribution for arandom noise we
might get one of the form shown in Figure 5-13. If we did, we would assume
we were dealing with a "Gaussian" noise, which is characterized by the
bell-shaped curve. It is also said to have a "normal" distribution. An ampli-
tude-density distribution of this form is commonly observed for acoustic
noises.
If astrong sinusoidal component is mixed with random noise, the ampli-
tude-density distribution will be a mixture of the two basic density distribu-
tions shown.

LA

1.•

II

E, OA

0.1

Figure 5-12. Amplitude den-

or sity distribution of a finely
sampled sinewave.
?LA - OA - es - 0.4 - OA 0 OA OA OA OA IA

AMPLITUDE INTERVALS

Figure 5-13. Amplitude histograms of

random noise.
-2 0 +2
AMPLITUDE - VOLTS

86
meter is used, the fluctuations would be asmall fraction of adecibel. As a
matter of fact, one can find many examples of an over-all sound level that
fluctuates over many decibels.
One example is the background noise of private offices. Here, for C
weighting in the slow meter position, one can commonly find fluctuations of
three or more decibels. The fluctuation corresponds to aband that is only
tens of hertz wide rather than 8000 to 10,000 hertz wide, such as that of the
response of the sound-level meter. This is because the energy in the sound is
concentrated in the low frequencies over a relatively narrow band. The
fluctuations reflect only the relation between the equivalent frequency band
of the signal applied to the metering circuit and the averaging time of the
circuit. Whether the energy is concentrated in anarrow band by means of an
electrical analyzer or by the source and the path to the microphone is im-
material.
Interpretation of Fluctuations. One can conclude, then, that if the ob-
served fluctuations are significantly greater than would be expected, an
important part of the random-noise energy is concentrated in aband or bands
that are narrower than the pass band of the measuring system. ( Another
possibility is that the type of noise is sufficiently different from normal that
the fluctuations for agiven bandwidth are inherently excessive. This behavior
is possible for atone whose frequency varies in aregion where the response of
the measuring system varies markedly with frequency.) It is also clear that if
the fluctuations are significantly smaller than would be expected, the noise
very likely includes some discrete tones that have significant amounts of
energy.
5.6.6 Speed of Processing — " Real Time." If an analyzer can operate to
process the input signal continuously, it is often called a "real-time" analyzer.
This type of operation usually requires a parallel type of analyzer or some
storage system. The accuracy and frequency range over which it may operate
in real time is usually significant, particularly with digital equipment, and the
cost usually increases with the speed and accuracy.
In an FFT system, which operates on discrete frames of data, real-time
operation requires what is called "buffered-mode" operation. Here one frame
of data is being stored while another is being processed. Then, if the process-
ing can be done within the time taken to acquire aframe of input data, real-
time operation is possible. Since manynoise analyzers require that anumber
of spectra be summed, the real-time operation with the buffered mode can
make it possible to utilize the full data available in agiven time,
The real-time feature is particularly important for signals that vary in
character with time, for example, the sounds from aircraft, missiles, speech,
music, and many machinery operations, and when it is unproductive to tape
record the sounds for later analysis.
Other definitions of real time have been used, and the basic requirement
seems to be that the operation must be completed quickly enough to suit the
application at hand. The user must recognize this time factor as another
element in his choice of analysis equipment. To illustrate the range of times
involved, consider the problems of analysis with a10-Hz band over the range
from 20 Hz to 20000 Hz. A serial analyzer will take about 1000 to 2000
times the time required for avery fast FFT system to do the basic processing.
5.6.7 Dynamic Range. The dynamic range of an analyzer is set by an upper
limit of distortion and a lower limit of internal noise, selectivity, or arith-

103
 0.5/e

04/o

0.3/s

0.2/o

o
-4o - 3o
3o 4o

Figure 5-14. The amplitude density distribution p ( y) of Gaussian random noise.

100

075

0.50

0.25

o
-4o - 3a -o 2o 3o 4u

Figure 5-15. The amplitude distribution function P (y) of Gaussian random noise.

If the signal being sampled is a periodic signal, even fewer observations of

he signal are required for agood estimate of the rms value. But some unusual
ampling situations must be avoided. Thus, the sampling rate must not have
he same period as the period of the signal or some integer multiple of that
)eriod. When sampling procedures are used in the process of determining the
ms value of awave that may be periodic, these conditions are avoided by one
ef the following:

1. Sampling at arandom rate.

2. Sampling at arate that changes according to some rule.
3. Sampling with aperiod less than half the period of the signal.
The measurement of amplitude distribution, in a practical case, is not
'early so useful as a spectrum measurement and it is rarely done (Piersol,
967). It is helpful, however, to explain it to bring in various ideas and
echniques that are used in other applications.

88
 5.5.2 Level Distribution. If the spectrum and the amplitude distribution of
a noise remain essentially the same regardless of the time when a set of
observations is made, it is stationary. But, actually, we are ordinarily con-
cerned only with alimited time span. If we ignore the startup and shutdown
phase of the cycle, the noise of most refrigerators can be regarded as
stationary. We can make many measurements while it is running, we can run
it again and again, and we can make more measurements that will give
consistent results if we duplicate conditions.
Many other noises are not so readily made effectively stationary. Thunder,
asonic boom, adoor slam, and an explosion are examples of transient sounds
that have to be treated differently. But there are also noises of intermediate
types or combinations of noises.
When amachine operator in ashop is setting up his work, the noise from
his machine may be at alow level, but there will be some background noise
from other machines that are in operation. When he runs his machine, the
noise in his vicinity may increase significantly. If we are trying to evaluate the
noise he is exposed to, we need to take this variability into account. We can
proceed in a fashion similar to that for amplitude distribution to obtain a
"level distribution."
We use a sound-level meter to measure the level. (Now it is not the
instantaneous value of the wave, but it is the rms value averaged over some
time.) We use intervals of level, say 90-92 dB, 92-95 dB, 95 -97 dB, etc., with
times arranged to indicate the total time that the noise stays within each of
the intervals. At the end of the day we could plot ahistogram of the result,
and we would have alevel distribution of his exposure to noise.

5.6 TECHNIQUES AND CHARACTERISTICS OF ANALYSIS.

The basic processes used for spectrum analysis are afiltering, or resonance,
technique and a Fourier transform, or correlation, technique. Frequency
shifting, or heterodyning, is sometimes used to extend the basic range of
these systems, and time compression or scaling is used in some instruments to
speed up the filtering process.
As indicated in the earlier sections, the signal can be processed by analog
techniques, digitally, or by acombination of the two.
• 5.6.1 Direct Filtering. Many analyzers now use sets of electronic band-pass
filters to separate the signal components into the required number of stand-
ard bands. When those filters are operating directly at the frequencies of the
desired bands, we call the analysis "direct filtering."
Most octave and one-third-octave analyzers use such electronic systems,
either combinations of inductors and capacitors, or networks involving resis-
tors, capacitors and amplifiers in feedback circuits, to produce the required
resonant effects.
Serial or Parallel Operation. We can classify analyzers by another feature
into two types, that is, serial or parallel. In strictly parallel operation, the
input signal is passed simultaneously through aset of filters with detectors at
the output of each filter (see Figure 5-16). The output level is then available
continuously for each filter band.
In serial operation, however, only one detector is used, and the filters are
sequentially switched into the circuit or are tuned sequentially to the re-
quired frequencies. The level in each band is determined before the next band
is measured.

89
 FILT ER DE TEC TOR --I CAPAC ITOR
#I #I — STORAGE
*1

AND OUTPUT CIRC

FI LTER DE TECTOR CAPA CI
TO R
#2 STORAGE
#2
*2

FI LTER DETECTOR CAPA CITOR

STOR AGE
#3 #3

ANNING
#3

r i r 1 I -
L. i FI LTER IDETECTO R I I
-CAPACI
TOR I 1
-— I r---1i N
I
- - - -I STOR AGE
I
I-
L I
«J L J
I
L
N I
.
1

Figure 5-16. Typical parallel-operation analyzer.

The tuning of the filter in serial operation may be continuous through the
frequency range or it may proceed in steps. In acoustics, the stepped opera-
tion that goes from one standardized preferred center frequency to the next
is more widely used than the continuous one, mainly because the spectrum is
then characterized by afinite set of numbers.
Serial-type analyzers are usually smaller and less expensive than corres-
ponding parallel types, because they require fewer elements to do the job.
The parallel type, on the other hand, can be much faster. Consider the
following example. If we dwell for one second to determine the level in each
of 30 third-octave bands, the serial analyzer would require 30 seconds, while
the parallel analyzer would have the output available after only one second.
Some time must be allowed for presenting the data in usable form, but it is
obvious that if many spectra with many bands must be measured, much time
can be saved by the use of aparallel analyzer.
One can take advantage of the sampling techniques in the amplitude
domain, where sampling less often than at the Nyquist interval is generally
satisfactory, to make a hybrid type of parallel operation. By the use of
sampling, multiplexing, and storage techniques one detector system can be
made to act as many.
Digital Filtering. Band-pass filtering can be done digitally on a sampled
time series (Gold and Rader, 1969; Oppenheim, 1969; Kuo and Kaiser, 1966;
Enochson and Otnes, 1968). An output time series is generated from the input
time series by the use of adifference equation, which relates an output value
at a particular time to the input values at that time and for some previous
times, as well as to one or more previous output values.
By the use of agroup of these difference equations, we can simulate aset
of one-third-octave filters, for example, that each produce a time series
equivalent to a filtered signal. From each of the output time series we can
calculate in the computer an rms value to give the one-third-octave levels.
Some of the parameters that affect the dynamic range, the discrimination
against interfering signals, and the stability of the output time series are:
1. Accuracy and resolution of conversion to digital form.
2. Length of time series.
3. Actual difference equations used.
4. Accuracy of calculations.

90
 5.6.2 Fast Fourier Transform. Another important spectrum-analysis tech-
nique is the basic one of direct application of the Fourier transform. It has
only recently become significant in the analysis of actual acoustic and
vibration signals as aresult of the development of the calculation procedure
called the Fast Fourier Transform ( FFT) ( Bingham et al, 1967; Singleton,
1969; Gold and Rader, 1969). FFT is an efficient way of calculating correla-
tion of awaveform with sinewaves whose frequencies are integer multiples of
the frequency corresponding to the time duration of the wave.
In order to apply the FFT to a signal, the signal is transformed into a
digital time series as explained previously. A fixed number of consecutive
points in the series is selected. Usually, this number is a power of 2, for
example, 1024, or 2048. This time series is then transformed into correspond-
ing components in the frequency domain by the use of the FFT calculation in
acomputer. If the signal is noise, or has the character of noise, anumber of
transforms must be combined to get statistical stability in the answer.
Some of the parameters that affect the characteristics of the process are
the following:
1. Number of samples in selected time period.
2. Accuracy and resolution of the conversion to digital form.
3. Weighting functions used on the sampled values.
4. Number of transforms combined.
5. Accuracy of the calculation procedure.
These parameters determine the frequency resolution, the dynamic range,
the discrimination against interfering signals, and the statistical stability of
the output spectrum.
The number of output values developed in the analysis is equal to the
number of data points in the original frame. But they are in pairs with areal
(cosine) value and an imaginary (sine) value for each integer multiple of the
fundamental frequency. The two together are then usually described as a
frequency component with areal and imaginary part or as avector with an
absolute magnitude and a phase angle. For many acoustical problems, the
phase angle is ignored and the magnitude at each integer multiple of the
fundamental frequency is the value used as the result of the analysis.
The square of this magnitude is sometimes called an "autospectral value,"
and the set of squares is the "autospectrum." This set of squared values is also
sometimes called the "power spectrum." Since the actual values are hardly
ever "power," and since the " cross spectrum" is also used (see paragraph
5.7.1), it is convenient to use the terms "autospectrum" and " cross spec-
trum" with the similar and related "autocorrelation" and " cross-correlation."
Some simple relations for the frequency transform are as follows:
Number of component lines = 1/2 number of data points in frame.
Frequency range = 1/2 sample rate.
Freq. range 2X Freq range
Resolution/line = —
Number of lines Number of data points

Frame period = Number of data points/sample rate

= Number of data points xsampling interval
Resolution/line = 1/Frame period
• Windows/Truncation. When aframe of points is selected and used in an
FFT analysis, the time for the initial point becomes the reference for the start
of the sine and cosine waves used for the analysis. In addition, the standard

91
transform uses the duration of the frame as the basic period for the analysis.
The fundamental component in the analysis has a frequency that is the
reciprocal of that basic period, and the frequencies of the other components
are integer multiples of that fundamental.
If the actual signal from which the frame is taken is periodic, and if the
duration of the frame is an integer multiple of the signal period, then the
analysis can give excellent results without modification of the frame. Another
way of stating this favorable situation is that the component frequencies of
the actual signal are all integer multiples of the fundamental frequency
corresponding to the frame period. This situation is unusual, however, and we
need to look at the problems of amore general case.
There are anumber of ways of looking at what happens in an analysis with
a finite duration input. One approach uses the concept of a data or time
window. The input signal is regarded as extending indefinitely in time, and
the sampled frame can be regarded as the input signal looked at through a
finite window in time, or multiplied by adata window that is 0 everywhere
except during the sampling period and then it is unity.
When atransform is made, the developed analysis fits the input frame of
data correctly, but in doing so, it in effect is analyzing asignal that is the
original frame continuously repeated. A simple example of what can happen
is shown in Figure 5-17, where asinewave has been sampled. The discontinu-
ity at the ends leads to an extensive set of components in the analysis that
may obscure or interfere with the components of interest. This effect is
sometimes called "leakage." If the frame were changed to include more of the
original signal, the discontinuity would be different, and the results of the
analysis would be different. In this example, if the frame were set to coincide
with the period of the sinusoid, there would be no discontinuity, and no
difficulty would occur. We cannot, however, adjust the frame to fit any
possible signal, because many are not periodic, and even for those that are we
would not necessarily know what period to use. With afixed frame, let us
observe the response of the system as we vary the incoming signal frequency.
With a rectangular window we will obtain aresponse of the form shown in
Figure 5-18. This response shows the desired peak at f o when the applied

signal and the frame fit correctly. When the frequency deviates from this
optimum value, the response decreases.
Another way of looking at the problem is that the Fourier transtorm is
designed to produce a set of functions that will combine to reproduce the
original data. The discrete transform uses only aminimum number of these

WI NDOW
DURATION

Figure 5-17. A repeated frame of data that shows the

discontinuity that can occur when the window dura-
tion is not an integer multiple of the basic period of
the sampled wave.

92
 ,
I I

, wo,k
RECTANGULAR WINDOW

VIIFOOte
OURAT,OPCI
A.

o
1n nd

t
o

4 t
- - -2 _ 0 2 3
4
FREQUENCY OF INPUT SINUSOIDAL WAVE RELATIVE TO THAT OF
MAXIMUM RESPONSE IN UNITS EQUAL TO FREQUENCY RESOLUTION

Figure 5-18. Effect of truncation with abrupt rectangular window

on the response of the transform.

functions to reproduce these data and the behavior of the combined func-
tions beyond the data window may not be what we would always desire. If
we wish to control the behavior beyond the original window, we have to
specify more data points outside the original set and thereby enlarge the
window.
Several techniques have been developed to reduce the effects of "leakage"
or of the finite window. The most commonly used one is amodification of
the window to have asmooth transition from zero to the full value, instead of
the abruptness of the rectangular window. The simple tapering of the " raised
cosine" or the hanning window, shown in Figure 5-19, is the one most widely
used, and it is very effective. The response now becomes as shown.
When hanning is used, the data at the ends of the frame are ignored, since
they are multiplied by avalue near zero. It is important then that the data
window be positioned and made wide enough to ensure that the important
behavior is centered within the window.

(
The tapered weighting used in banning and other windows also broadens
the main response of the transform. This broadening can be compensated for
by increasing the number of points in the data frame.

O
I 1
HANNING WINDOW
10

20

ri
S
w 30

›'

g 40

nn
:
3

50

60 5
-• -2 - 2 3 4
FREQUENCY OF INPUT SINUSOIDAL WAVE RELATIVE TO THAT OF
MAXIMUM RESPONSE IN UNITS EQUAL TO FREQUENCY RESOLUTION

Figure 5-19. Effect of tapered window (harming) on the response

of the transform.

93
 Reducing the effects of interfering components is so important that
hanning or some similar tapered weighting should almost always be used.
• 5.6.3 Time Compression or Scaling. When aserial type of analyzer is to be
used to analyze a signal that changes with time, the signal is often first
recorded on a tape recorder. Sections of the record are then selected, made
into tape loops, and played back to the analyzer until the analyzer has had
time to go through the desired range of analysis.
If the speed at which the tape loop is reproduced is different from the
original recording speed, all the signal components are translated in frequency
by the speed ratio, and the repetition period of the loop is changed by the
inverse of this ratio. Thus, if we have a one-second loop recorded at 1 7/8
in./s and played back at 15 in./s, the loop will repeat every 1/8th of asecond,
and a1000-Hz component will become an 8000-Hz component.
The loop when played back at its original speed repeats every second. The
output signal will then have components that are spaced 1 Hz apart. The
speeded-up loop will have components with 8-Hz spacing.
If we now analyze this speeded-up loop in third-octave bands, we use the
band at 8000 Hz to find the value of the components in the original signal in
the 1000-Hz band. The band at 8000 Hz is 8times as wide as the one at 1000
Hz, and the response of its filter is correspondingly 8 times as fast. Now we
can in effect process the signal 8times as rapidly as at the original speed.
If the signal is converted into digital form, it can be stored in acirculating
digital memory rather than in a tape loop. The speed-up that is then possible
is many times greater, being 1000-to-1or even more.
• 5.6.4 Frequency Translation or Heterodyning. Both the serial and parallel
types of analyzer can be operated over a wide range of input-signal frequen-
cies by translating the input signal frequencies to be within the range of the
analyzer. This technique has been widely used with serial analysis to translate
the effective center frequency of asingle highly selective filter.
The technique is illustrated in Figure 5-20. Assume we are concerned with
the signal components in the vicinity of 1000 Hz. A local oscillator in the
device is set to generate asinewave at 101 kHz. This wave is mixed with the
incoming signal, and the resulting components, with frequencies in the
immediate vicinity of 100 kHz, pass through the 100-kHz filter and are
indicated by the detector system. The component in the original signal at
1000 Hz would be the principal component measured. But components at
100 kHz and at 201 kHz also produce an output at the detector, but they are

INPUT 100- kHz 100-1H,

PHASE BALANCED CRYSTAL CRYSTAL
SPLIT TER MOGUL ATOR F5.TER FILTER

BANDWIDTH

CALIBRATING 100 - IS4 MHT AFC DISCRIMINATOR

SIGNAL OSCILLATOR AND
REACTANCE DIODES

100 11.
RECORDER
OUTPUTS
42)— RECTIFIER IIRA DC

Leer
INDIC AT D
MIXER FREQUENCY
FILTERED
INPUT
COMPONENT
„del TRACKING
ANALYZER

r—o
NORMAL
100- laM
OSCILLATOR

Figure 5-20. Block diagram, aheterodyne analyzer.

94
 •4
I 1 1
RANDOM- INCIDENCE
PERPENDICULAR- INCIDENCE

I; 0 .ir
- P --- ---

•
50 70 100 200 500 700 2k 5 r r 20 50k
FREQUENCY IN HERTZ

Figure 6-3. Frequency characteristics of two different half in. condenser microphones.
They optimize response for specified sound incidence.

search, Chesepeake Instrument Corporation, Clevite Ordnance, Gulton Indus-

tries, Scientific-Atlanta, and Wilcoxon Research.

6.2.5 Vibration Pickups. The vibration pickups supplied by General Radio,

as shown in Figure 6-5, are all inertia-operated, lead-zirconate, lead-citanate,
piezoelectric devices that generate a voltage proportional to the acceleration
of the pickup ( Dranetz and Orlacchio, 1961; Mason, 1961, Carlson, 1952).
Voltages proportional to velocity and displacement of the vibrating body are

Figure 6-4. A typical hydrophone.

Figure 6-5. Typical appli-

cation of a vibration pick-
up. It is mounted on the
end bell of the left motor
in the foreground and the
vibration signal is being ana-
lyzed by a General Radio
automatic-recording 1/3-
octave-band analyzer.

112
 • • •

I
!!!
• • 1.1•11•0<1.111 OMNI IFOLIM•1111,140••

1/10 OCTAVE
00d8

ue 1% is OAVE

Figure 5-21. Blurring effect and

ó2
excess level error for various band-
nor, • 011 . i• 2
widths (slope = 60 dB/octave).
FREQUENCY MULTIPLIER 1000

HUAI SI

wider bandwidths. There is an obvious rounding at the lower corner, and the
wider the band used the more the corner is rounded.
The wider bands also show ashift in level, which will occur whenever the
spectrum slopes steeply, because the energy passed by the filter, in the region
where the level is high, will more than counterbalance the decrease in energy
in the opposite region. This shift has sometimes been regarded as an apparent
shift in the center frequency of the filter, but when levels are to be specified
at preferred frequencies, it may be more convenient to think of it as ashift in
level. The extent of the effect can be expressed as an excess in apparent
spectrum level, and Figure 5-22 shows the results of calculations for various
bandwidths and spectrum slopes. It illustrates the limited resolving power of
wide filters even when ideal. This effect is important mainly when comparing
results with analyzers of different bandwidths. For most practical noises, the
effect is not great except for octave bands. For the one-third-octave band and
the extremely steep slope in the example of Figure 5-21, the shift in level in
the middle of the slope should be somewhat over 3dB, which is essentially
the shift observed.
A similar effect results in filling in anarrow dip and rounding off the top
of a peak in a spectrum. It is merely what is to be expected from the
limitations in resolution of afinite bandwidth.
• Degrees of Freedom/Spectrum Averaging. Assume we analyze successive
frames of sampled values of arandom-noise signal. The results of a Fourier

EXCESS SPECTRUM LEVEL OF IDEAL FILTERS

OCTAVE 22- OCTAVE

Ils- OCTAVE
2

1.0.0CTAVE Figure 5-22. Excess level error

n
for various bandwidths.
-10 -00 -40 - 20 0 20

SPECTRUM SLOPE - de/OCTAVE

96
transform would show avery large variability in the indicated level in any one
band, from one frame to the next. This variability is characteristic of a
random signal ( Bendat and Piersol, 1966; Sloane, 1969).
In order to produce values that have some significance for the band levels
of noise, we must combine many samples. We can sum the corresponding
squares of the measured spectrum values (autospectra) for a number of
frames, which we call averaging. We can sum the squares of the measured
spectrum values for anumber of adjacent bands, which reduces the resolution
but improves the statistical stability. Or we can do both, that is, average and
combine bands.
In order to show what needs to be done, we shall describe what happens
with awhite-noise signal of agiven bandwidth, B, in Hz. If this is sampled at
the Nyquist interval, we will get all the information available in the signal. If
the frame of points is taken over atotal time span T, we will have aframe of
2 BT points. This value for bandwidth-limited Gaussian white noise is the
number of statistically independent sampled values, which is sometimes called
the number of degrees of freedom. It can be used to describe the expected
behavior of the variability of the random signal.
If acomplete Fourier transform of this frame is now made, each autospec-
tral component ( the sum of the squares of the amplitudes of the sine and
cosine terms) will have two degrees of freedom. For each independent frame
of data that is summed in, two degrees of freedom are added. Thus, if we had
32 autospectral values at 1000 Hz that we had summed (we could normalize
by dividing by 32 to get the average), we would have 64 degrees of freedom.
From the chart of Figure 5-23 we find that we would have a95% confidence
that the resultant level is not greater than the long-term true value by more
than 1.4 dB or less than the true value by more than 1.6 dB.
The results of some measurements on white noise, with a Time/Data
1923-C FFT Analyzer, will illustrate what these statements mean. One

.4

.2

95%
/.........."C0.2110E
L NITS 99%
90%

2 0 16 32 64 126 256 SIS .024

DEGREES OF FREEDOM
.....,
Figure 5-23. Reliability of amplitude analyses as a function of degrees of freedom from
the data sample for various confidence limits.

97
tndred successive levels of the same single band of stationary noise were
easured. Each of these 100 levels was a result of 32 averages. (The high-
eed processing made it possible to get the 3200 measurements rapidly.) The
easured levels were then arranged for plotting on probability paper as
own in Figure 5-24. The expected distributions for 32 and 64 degrees of
!edom are also shown and they have been set at the level to give abest fit to
e observed data. The results show that any particular averaged level that
es 32 independent frames is very likely to be within a 3-dB span, as
edicted. Similar results are shown for a measurement with a hanning
indow.
The total number of degrees of freedom for aframe of data is reduced by
le use of atapered window, such as hanning. Thus, if anumber of adjacent
Inds are combined by summing the squares of the component values, the
amber of degrees of freedom will not be twice the number of bands that are
)mbined. If hanning is used, it will approach only about one-half that value.
•the noise is not white over the range of the combined bands, or if the noise
not essentially stationary over the time for the frame, the total number of
egrees of freedom will be reduced even further.
Thus the chart of Figure 5-23 should be used mainly as aguide.
The summing and averaging that is used here is different from that used in
•aveform averaging (paragraph 5.4.5). In waveform averaging, the summing is
near and includes the sign, and therefore, random noise, which is random in
alue, will add up more slowly than asignal that is always of the same wave
lape with respect to the starting point of the frame. This type of averaging
cduces random noise.
The averaged waveform will have the coherent signal emphasized with
espect to the noise by a factor equal to the square root of the number of

e'

! lI
.0

'
_

,,

II
«ED
,
*/
UNWAUMS:32— AV
./
MANNED- 32 AVG

I 4111111111 II

IVE LEVEL - ES «LAE VI LEVU IA

Figure 5-24. Comparison of the Figure 5-25. Comparison of the

statistical stability of experimen-
statistical stability of experimen-
tal measurements with theoreti- tal measurements with theoreti-
cal curves for 32 and 64 degrees cal curves for 32 and 64 degrees
of freedom.
of freedom — hanning was used.

98
 (9D

Single average

8 averages

600 averages

FREQUENCY— HERTZ 5k

Figure 5-26. Averaged auto spectrum of band-limited noise.

sums. This technique " pulls" asignal out of noise. If one were interested in
the spectrum of the coherent signal only, waveform averaging before trans-
forming to obtain aspectrum would be agood approach when it is possible.
The spectrum-averaging procedure used in this section sums the squares of
the magnitudes of the components, and the sign does not enter, since the
squared value is always positive. This spectrum averaging gives amore stable,
that is, less variable, value for the spectrum levels of a random-noise signal,
provided the noisy signal is stationary over the averaging period. When
measurement of the noise level is important, as is often the case in acoustics,
spectrum averaging makes possible a more reliable measurement of the level,
as shown in Figure 5-26.
It is also interesting to note that spectrum averaging sometimes makes it
possible to find a periodic signal that may otherwise be obscured by noise,
provided the noise is stationary. The point involved here concerns the fact
that a transform of a single frame yields spectrum levels for the noise that
have only two degrees of freedom. The variability of level is then very large.
A pure-tone component that is of the same order as the band level of the
noise will appear as just another noise component among many in its vicinity
that are similar in amplitude.
If many frames are averaged as shown in Figure 5-27, the noise-spectral
values are more uniform in level. Then the effect of an added periodic
component will be apparent by its projection beyond the relatively uniform

99
 AUTOSPECTRuM
(SINGLE FRAME)

a
a io 14 16 le 20
FREQUENCY - kHz

SPECTRUM AVERAGING
(256 FRAMES AVERAGED)
-. ( PRESSURE)

o 2 4 6 8 10 12 18 20 e
FREQUENCY - Mg:

Figure 5-27. Autospectra of a signal that contains a sinewave buried in random noise.
The autospectrum for a single frame is shown at (a), and the normalized average of tl..i
autospectra for 256 frames is shown at (b).

amplitude of the average noise-spectral values in its vicinity. This averaging

procedure is particularly appropriate when both periodic and random signals
are important.
• Statistical Stability in Analog Systems. The measurement of random noise
on an analog system involves essentially the same concepts as in a digital
system. Now, however, the metering circuit provides some averaging, and two
meter speeds are often provided in acoustic measuring instruments. These two
speeds are designated fast and slow, with the slow condition yielding the
longer averaging time.
The fact than an averaging time is used leads to a dependence of the
statistical stability on bandwidth. That is, if arandom noise is analyzed with
an analog system, the extent of the meter fluctuations depends on the
bandwidth. The narrower the band, the greater are the fluctuations and the
longer is the averaging time required for asatisfactory estimate of the level.
A relatively simple principle is involved here. A narrow band is used to get
fineness of detail. The finer the detail that is desired, the more time is needed
to obtain the result to acertain degree of confidence.

100
• Example of Random-Noise Measurement. To illustrate by an actual numer-
ical example the type of behavior that occurs, some measurements were made
of an arbitrary level of arandom-noise generator in the octave band from 150
to 300 Hz. With the fast meter speed, the average of the fluctuating levels
indicated on the meter was estimated to be about + 5dB, where in aperiod of
30 seconds the level fluctuated from aminimum of + 3.3 dB to amaximum of
+6.5 dB, arange of 3.2 dB. In the slow position the estimated level was +4.7
dB, and the level fluctuated over athree-minute period from a minimum of
+3.8 to a maximum of +5.7, arange of 1.9 dB. Some sample readings were as
follows: fast position: 4.8, 4.1, 5.3, 3.7, 5.8, 4.9, 5.3, 5.2, 6.2, 4.6; slow
position: 4.4, 5.1, 3.9, 4.9, 4.2, 5.0, 4.7, 4.1, 4.3, 4.9. (These sample readings
were taken with the help of astroboscope, to avoid observer bias in selecting
readings and to make it possible to take definite readings on the rapidly
moving pointer in the fast position.) One hundred samples were taken for
each position. The average value on an energy basis for slow was +4.72, with
the lowest reading + 3.8 and the highest + 5.8. The average for fast was +4.74,
with alow reading of + 3.1 and 2, high reading of +6.2.
Taking such aset of readings is not the usual way to obtain the indicated
level; rather, one estimates avalue by observing the pointer fluctuations. But
these discrete samples permit one to describe statistically the behavior that
can be expected.
For the fast position one would expect only 1 in 1000 readings to differ
from the average by more than about - 3dB or +2.4 dB, arange of 5.4 dB.
The corresponding extremes for one chance in 100 is about - 2.3 dB or + 1.9
dB, a range of 4.2 dB; for 1 in 10, about - 1.4 to + 1.2, arange of 2.6 dB.
Note that the range is not symmetrical.
These statements about variability can be expressed in another way, which
is the comerse of that above. If any reading is taken in the fast position, the
chances are only 1 in 100 that the long-time average value of the noise is
below the observed value by more than 1.9 dB or above the observed value by
more than 2.3 dB. These limits are called the 99% confidence limits.
Confidence Limits for Octave Bands. A chart of the 99% confidence limits
for octave bands for random noise measurement is given in Table 5-4.
These ranges of uncertainty can be reduced by the use of the average of a
number of independent readings. The reduction in the range is approximately
inversely proportional to the square root of the number of independent
observations. Thus, the average of four observations would reduce the uncer-
tainty to about one-half that shown.
Table 5-4
CONFIDENCE LIMITS FOR RANDOM
NOISE IN OCTAVE BANDS
99% Confidence Limits (dB)
Meter Speed
Center Freq
(Hz) Fast Slow
31.5 -4.2, +7.0 -2.5, +3.3
63 -3.2, +4.7 -1.8, +2.2
125 -2.4, +3.1 -1.3, + 1.5
250 -1.7, +2.1 -1.0, + 1.1
500 -1.2, + 1.4 -0.7, +0.7
1,000 -0.9, + 1.0 -0.5, +0.5
2,000 -0.6, +0.7 -0.3, +0.3
4,000 -0.5, +0.5 -0.2, +0.2
8,000 -0,3, +0.3 -0.2, +0.2
16,000 -0.2, +0.2 -0.1, +0.1

101
 The range of uncertainty is sometimes called the statistical error.
Averaging By Observation. When one observes the fluctuations of ameter
for a time and estimates an average, the extent of the reduction of the
uncertainty is limited by the fact that all the observations are not indepen-
dent, and one can remember and use only a small portion of the total
observed behavior. The observations are not independent because of the finite
time required for the pointer to assume anew value. In the fast position of
the meter, one should allow about one-half second between observations; in
the slow position, an interval of one to two seconds is desirable.
• Averaging By Circuit Time Constants. The smoothing or reduction in
fluctuations achieved by the electrical circuit is often characterized by an
equivalent time constant. The averaging time constant required to reduce
fluctuations of arectified noise signal to adesired amount for agiven band
width of noise is approximately:
1250
T --k: seconds
Afa 2

where T is the averaging time

M is the bandwidth of the noise in Hz.
a is the standard deviation of the fluctuations at the output of alinear
detector in percent.
Duration of a Sample. The uncertainty that results from the limited
observation time, in comparison with the detail desired in the frequency
domain, occurs for other time limitations as well. Moreover, some of these
may not be under the control of the operator. Thus, the sound source may
not perform uniformly over an extended period of time; for example, a
rocket may run for only a fraction of a minute. During launch, the time
available for observing arocket may be only afew seconds or less.
When anoise signal, recorded on amagnetic- tupe recorder, is to be studied,
it is customary to take short samples for analysis. These samples are cut from
the full recording and formed into loops that can be run continuously in the
recorder. This procedure directly limits the fineness of detail possible in the
analysis and also limits the accuracy with which one can determine the actual
level in aband.
This limitation of accuracy results from the fac, that the maximum time
during which independent information can be obtained is the sample dura-
tion. If the noise is sufficiently uniform with time, alonger sample can be
used to obtain increased accuracy, or measurements on anumber of samples
can be averaged.
Because of the inherent variability of random noise, analyses of distinct
samples of the same noise will not yield identical results. The expected spread
in values predicted by statistical theory can be used as a guide in judging
whether the results of such analyses agree well enough to be useful. Unless
this inherent variability is appreciated, one can be led into rejecting useful
data, rejecting a useful analysis system, or placing too much reliance on a
particular measurement.
Fluctuations Produced in Practice. The table of values shown for the
octave bands is based on the analysis of noise that is uniform in energy per
hertz throughout the band. In the wider bands, the values shown are mislead-
ing for acoustical signals, because the energy is not uniformly distributed.
One should expect from such values that, when the full range of asound-level

102
meter is used, the fluctuations would be asmall fraction of adecibel. As a
matter of fact, one can find many examples of an over-all sound level that
fluctuates over many decibels.
One example is the background noise of private offices. Here, for C
weighting in the slow meter position, one can commonly find fluctuations of
three or more decibels. The fluctuation corresponds to aband that is only
tens of hertz wide rather than 8000 to 10,000 hertz wide, such as that of the
response of the sound-level meter. This is because the energy in the sound is
concentrated in the low frequencies over a relatively narrow band. The
fluctuations reflect only the relation between the equivalent frequency band
of the signal applied to the metering circuit and the averaging time of the
circuit. Whether the energy is concentrated in anarrow band by means of an
electrical analyzer or by the source and the path to the microphone is im-
material.
Interpretation of Fluctuations. One can conclude, then, that if the ob-
served fluctuations are significantly greater than would be expected, an
important part of the random-noise energy is concentrated in aband or bands
that are narrower than the pass band of the measuring system. ( Another
possibility is that the type of noise is sufficiently different from normal that
the fluctuations for agiven bandwidth are inherently excessive. This behavior
is possible for atone whose frequency varies in aregion where the response of
the measuring system varies markedly with frequency.) It is also clear that if
the fluctuations are significantly smaller than would be expected, the noise
very likely includes some discrete tones that have significant amounts of
energy.
5.6.6 Speed of Processing — " Real Time." If an analyzer can operate to
process the input signal continuously, it is often called a "real-time" analyzer.
This type of operation usually requires a parallel type of analyzer or some
storage system. The accuracy and frequency range over which it may operate
in real time is usually significant, particularly with digital equipment, and the
cost usually increases with the speed and accuracy.
In an FFT system, which operates on discrete frames of data, real-time
operation requires what is called "buffered-mode" operation. Here one frame
of data is being stored while another is being processed. Then, if the process-
ing can be done within the time taken to acquire aframe of input data, real-
time operation is possible. Since manynoise analyzers require that anumber
of spectra be summed, the real-time operation with the buffered mode can
make it possible to utilize the full data available in agiven time,
The real-time feature is particularly important for signals that vary in
character with time, for example, the sounds from aircraft, missiles, speech,
music, and many machinery operations, and when it is unproductive to tape
record the sounds for later analysis.
Other definitions of real time have been used, and the basic requirement
seems to be that the operation must be completed quickly enough to suit the
application at hand. The user must recognize this time factor as another
element in his choice of analysis equipment. To illustrate the range of times
involved, consider the problems of analysis with a10-Hz band over the range
from 20 Hz to 20000 Hz. A serial analyzer will take about 1000 to 2000
times the time required for avery fast FFT system to do the basic processing.
5.6.7 Dynamic Range. The dynamic range of an analyzer is set by an upper
limit of distortion and a lower limit of internal noise, selectivity, or arith-

103
metic processing errors. If a signal that is too high in level is applied to an
analyzer, the analyzer will be overloaded. As aresult of the overload, the in-
dicated spectrum will be different from the actual spectrum. How seriously
the spectrum is distorted depends on the way in which the overload occurs.
But, ordinarily one can avoid overloading by the use of reasonable care in fol-
lowing the procedures given in the instructions for the analyzer.
Internal noise of an analyzer is often alower limiting factor in the analysis
range possible. Unless this internal noise is significantly less than the applied
signal, it can affect the indicated spectrum. The selectivity characteristic of
the analyzer also limits the range, particularly close in frequency to strong
components. In addition, some analyzers, particularly older designs, have
ultimate rejection of components outside the passbands of only 30 to 40 dB.
Other factors that enter into the dynamic range of digital systems are
aliasing and quantization, which are discussed in paragraph 5.3.4. A factor
that is related to the quantization is the detail of arithmetic processing. For
example, in additions, subtractions or multiplications, it is frequently neces-
sary to maintain aconstant number of bits in the results. If this is done by
simple truncation or dissymmetrical rounding, the noise introduced is usually
greater than for symmetrical rounding. An even more important effect can
occur in squaring the amplitudes of components, as is done for autospectral
values. Sometimes the squared values are limited to the same number of bits
as the basic values, in order to save storage space. This procedure can result in
the loss of all information for low-level components and effectively reduce
the dynamic range by afactor of two.

F.7 TWO-SIGNAL FUNCTIONS.

• 5.7.1 Cross Spectrum and Cross-Correlation. A number of functions show
relations between two signals. We have already briefly described cross-correla-
tion which expresses the similarity of two signals as afunction of time. A
related function is cross spectrum, which is the Fourier transform of the
cross-correlation and expresses the similarity as afunction of frequency.
The cross-spectral function can also be calculated from the Fourier trans-
forms of the two time series by a conjugate multiplication. This type of
multiplication gives the products of the magnitude and the differences of the

TIME DOMAIN FREQUENCY DOMAIN

COMPLEX
TIME SERIES SPECTRUM
4 FOURIER TRANSFORM
r) , X( oe)

COMPLEX
TIME SERIES SPECTRUM
FOURIER TRANSFORM
YU)

CROSS CORRELATION CONJUGATE MULTIPLICATION

CROSS— CROSS
CORRELATION SPECTRAL
FUNCTION 4— FOURIER TRANSFOR
R„ r) S„(1w)

Figure 5-28. Time/frequency-domain diagram

104
phases of the two signals. The alternative routes to the cross-spectral function
are shown in Figure 5-28 ( Heizman, 1970). One of the important applications
of the cross spectrum is in the calculation of the transfer function.
Just as for autocorrelation and cross-correlation, the autospectrum and the
cross spectrum are related in that the autospectrum is the cross spectrum with
both signals being the same.
However, it would be wasteful to calculate the auto spectrum by either
process shown in Figure 5-28, since the only output is the square of the
magnitude of the components and only one transform is necessary for the
autospectrum.
5.7.2 Transfer Functions and Coherence. The transfer function of adevice
or a system is the ratio of the output to the input. This function, which is
ordinarily defined in the frequency domain, can be useful for studying noise
and vibration problems. If we have the frequency analysis of both the input
and the output signals, we can then have the transfer function by taking the
ratio of the output to the input at each corresponding frequency component.
It is easy to see that we can select many different points in asystem to be
the output. Assume we have an electrically driven hydraulic pump that is in
an enclosure. We are concerned about the noise and vibration that it pro-
duces. What do we regard as the output? It could be the acoustic noise at
various places outside the enclosure, or inside the enclosure. It could be the
vibration at the pump housing or the vibration of the pipes.
We can look at the input in asimilar way. If we are interested in the noise
outside the enclosure, it is hardly useful to regard the electrical-power-line
terminals as the input. The vibration at various points on the pump or on the
pipes could be useful as an input, for the noise outside the enclosure as the
output.
In most practical acoustical and vibration problems, enough extraneous
noises are present to make the simple measurement of the transfer function
unreliable. By the use of an alternative form of the transfer function, these
extraneous effects can be reduced. The transfer function is also the ratio of
the cross spectrum of the input and output to the autospectrum of the input
(Sloane, 1970; Roth, 1971). The use of the cross spectrum, which includes
those components that are common to both input and output, eliminates any
extraneous components that do not appear in both.
Another function that is related to the transfer function is the coherence
function. It is normalized by including the input and output in both numera-
tor and denominator, as opposed to the transfer function, which has only the
input in the denominator. The coherence function is the ratio of the square
of the magnitude of the cross spectrum to the product of the input and
output auto spectrums. The coherence function covers arange of magnitudes
from 0 to 1as afunction of frequency, and the value depends on how well
the input and output values at each frequency are related. It provides auseful
further parameter to help in interpreting the transfer function.
The combination of transfer and coherence functions provides apowerful
technique in the study of noise sources and transmission paths.

5.8 ANALOG VS DIGITAL.

In an analog filter system, each filter band is essentially aseparate element
or is achieved by individual tuning. This technique is well suited for process-
ing of, say, 8 octave bands or 30 third-octave bands or for serial analysis.

105
Analog equipment for these tasks is available with excellent characteristics
and at lower cost than for digital equipment.
In adigital system that uses the Fast Fourier Transform, awhole series of
filter bands is achieved by the one transform. The difference then between 32
or 2048 filter bands is mainly amatter of memory size. Accordingly, digital
techniques become preferred as the desired number of filter bands increases.
In addition, many types of operations on the data are often easy to
include once the data is in digital form. Examples of these are averaging,
addition of corrections, combining bands in different ways to give both
narrow bands and third-octave bands, use of weightings of various kinds
either in the time domain or the frequency domain, and calculation of noise
ratings. In other words digital equipment is more flexible than analog equip-
ment.
Digital processing can have very high resolution, has unlimited repeatabil-
ity, can handle vast amounts of data, and is almost insensitive to environ-
mental factors such as ambient temperature and pressure.
Since the original signal is in analog form, some processing is often
conveniently done by analog techniques before sampling and conversion to a
digital form. The system is then ahybrid one and may combine some of the
desirable features of both techniques.

REFERENCES

Standards
ANSI S1.11-1966, Specification for Octave, Half-Octave, and Third-Octave Band Filter
Sets.
ANSI S2.10-1971, Methods for Analysis and Presentation of Shock and Vibration Data.
IEC 225-1966, Octave, half-octave and third-octave band filters intended for the analysis
of sounds and vibrations.

Other
W. B. Allen and E. C. Westerfield ( 1964), " Digital Compressed-Time Correlators and
Matched Filters for Active Sonar", JAcoust Soc Am, Vol 36, #1, jan, pp 121-139.
N. A. Anstey ( 1966), " Correlation Techniques - A Review", J. Can Soc Explor
Geophys, Vol 2, #1, Dec 1966, pp 1-28 and Geophys Prospecting, Vol 12, #4, 1964.
P. K. Baade ( 1966), " Instrumentation Trends tor Noise Reduction, as Seen by a
Mechanical Engineer", J Audio Eng Soc, Vol 14, #4, Oct, pp 302-306.
C. G. Balachandran and D. W. Robinson ( 1967/68), " Diffusion of the Decaying Sound
Field", Acustica, Vol 19, #5, pp 245-257.
J. S. Bendat and A. G. Piersol ( 1966), Measurement and Analysis of Random Data, John
Wiley and Sons, New York.
W. V. Bhat ( 1971), " Use of Correlation Technique for Estimating In-flight Noise
Radiated by Wing-mounted Jet Engines on a Fuselage", J. Sound and Vib, Vol 17,
#3, pp 349-355.
C. Bingham, M. D. Godfrey and J. W. Tukey ( 1967), " Modern Techniques of Power
Spectrum Estimation", IEEE Trans on Audio Electroacoust, Vol AU- 15, #2, June, pp
56-66.
R. B. Blackman and J. W. Tukey ( 1958), The Measurement of Power Spectra, Dover
Publications, Inc., New York.
R. Bracewell ( 1965), The Fourier Transform and Its Applications, McGraw-Hill, New
York.
A. N. Burd ( 1964), "Correlation Techniques in Studio Testing", Radio and Elec
Engineer, Vol 27, #5, May, pp 387-395.
A. N. Burd ( 1968), " The Measurement of Sound Insulation in the Presence of Flanking
Paths,"J Sound and Vib, Vol 7, # 1, January 1968, pp 13-26.
L. J. Chamberlain ( 1971), " A Simple Discussion of Time-Series Analysis", Sound and
Vibration, Vol 5, #4, April 1971, pp 18-25.

106
 W. A. Clark, R. M. Brown, M. H. Goldstein, Jr., C. E. Molnar, D. F. O'Brien, and H. E.
Zieman ( 1961), "The Average Response Computer ( ARC): A Digital Device for
Computing Averages and Amplitude and Time Histograms of Electrophysiological
Response", IRE Trans on Rio-Medical Electronics, Vol BME-8, # 1, Jan, pp 46-51.
R. K. Cook, R. V. Waterhouse, R. D. Berendt, S. Edelman, and M. C. Thompson, Jr.
(1955), " Measurement of Correlation Co- efficients in Reverberant Sound Fields", J
Acoust Soc Am, Vol 27, #6, Nov, pp 1072-1077.
G. R. Cooper and D. C. McGillem ( 1967), Methods of Signal and System Analysis, Holt,
Rinehart and Winston, Inc., New York.
P. W. Cooper ( 1964), "Correlation Detection in aDuovelocity Medium with Application
to Seismic Disturbances", JAcoust Soc Am, Vol 36, #7, July, pp 1378-1382.
L. D. Enochson and R. K. Otnes ( 1968), Programming and Analysis for Digital Time
Series Data, Navy Publication and Printing Service Office, Washington, D.C. 20390.
J. J. Faran, Jr. ( 1968), " Random-Noise Generators", General Radio Experimenter, Vol
42, # 1, Jan, pp 3-13.
J. J. Faran, Jr., and R. Hills ( 1952a), "Correlators for Signal Reception", Tech Memo 27,
Sept, Acoustics Res Lab Div. App Sci, Harvard Un, Cambridge Mass.
J. J. Faran, Jr. and R. Hills ( 1952b), " The Application of Correlation Techniques to
Acoustic Receiving Systems", Tech. Memo # 28, Nov., Acoustics Res Lab, Div. Appl
Sci, Harvard Un, Cambridge, Mass.
C. D. Geisler and W. A. Rosenblith ( 1962), "Average Responses to Clicks Recorded from
the Human Scalp", JAcoust Soc Am, Vol 34, # 1, Jan, pp 125-127.
D. A. Gilbrech and R. C. Binder ( 1958), " Portable Instrument for Locating Noise
Sources in Mechanical Equipment", JAcoust Soc Am, Vol 30, #9, Sept, pp 842-846.
K. W. Goff ( 1955), " Application of Correlation Techniques to Some Acoustic Measure-
ments", J Acoust Soc Am, Vol 27, #2, Mar, pp 236-246.
B. Gold and C. M. Rader ( 1969), Digital Processing of Signals, McGraw-Hill, New York.
A. Hannavy ( 1967), " New Detection Devices Help Predict Potential Failure", Prod
Engineering, Mar 27, pp 37-44.
C. L. Heizman ( 1970), "Signal Analysis with Digital Time Series Analyzers", General
Radio Experimenter, Vol 44, #7-9, July-Sept, pp 3-7.
C. W. Horton, Sr. ( 1969), Signal Processing of Underwater Acoustic Waves, U.S. Gov't
Printing Office, Washington, D.C. 20402.
S. Imai ( 1968), "On a Simplified Hybrid Analog-Digital Correlator for Sound Measure-
ments," Reports of the 6th Inter Gong on Acoust, Tokyo, Aug, pp E-13-16.
T. C. Kincaid ( 1968), "Optimum Waveforms for Correlation Detection in the Sonar
Environment: Noise-Limited Conditions," J Acoust Soc Am, Vol 43, #2, Feb, pp
258-268.
G. A. Korn ( 1966), Random-Process Simulation and Measurements, McGraw-Hill, New
York.
W. R. Kundert and A. P. G. Peterson ( 1969), "Spectrum Analysis of Stationary Noise
Signals", Sound and Vibration, Vol 3, #6 June, pp 16-24.
F. F. Kuo and J. F. Kaiser ( 1966), System Analysis by Digital Computer, John Wiley &
Sons, Inc., New York.
H. F. H. Lange ( 1967), Correlation Techniques, Iliffe Books Ltd, London, and D. Van
Nostrand Company, New York.
B. P. Lathi ( 1965), Signals, Systems and Communication, John Wiley 8c Sons, Inc., New
York.
Y. W. Lee ( 1960) Statistical Theory of Communication, John Wiley & Sons, Inc., New
York, 1960.
G. C. Maling, Jr., W. T. Morrey and W. W. Lang ( 1967), " Digital Determination of
Third-Octave and Full-Octave Spectra of Acoustical Noise", IEEE Trans Audio
Electroacous, Vol AU- 15, #2, June, pp 98-104.
D. Middleton ( 1960), An Introduction to Statistical Communication Theory, McGraw-
Hill, New York.
S. Nakamuru and M. Koyasu ( 1968), "Measurements of the Vibration Correlation
between Both Surfaces of a Multi-Layer Panel," Reports of the 6th International
Confess on Acoustics, Tokyo, Aug, pp E-125-128.
D. A. Nelson and F. M. Lassman ( 1968), " Effects of Intersignal Interval on the Human
Auditory Evoked Response," JAcoust Soc Am, Vol 44, #6, Dec, pp 1529-1532.
A. V. Oppenheim, cd ( 1969), Papers on Digital Signal Processing, The M.I.T. Press,
Cambridge, Mass.

107
A. G. Piersol ( 1967), " Practical Interpretations of Unusual Characteristics in Reduced
Vibration Data", J Env Sci, Vol 10, #1, Feb, pp 17-21.
P. R. Roth ( 1971), " Effective Measurements Using Digital Signal Analysis", IEEE
Spectrum, Vol 8, #4, April, pp 62-70.
H. H. Rothman ( 1970), " Effects of High Frequencies and Intersubject Variability on the
Auditory-Evoked Critical Response", J Acoustical Soc Am, Vol 47, #2 ( Part 2),
February 1970, pp 569-573.
R. C. Singleton ( 1969), "A Short Bibliography on the Fast Fourier Transform", IEEE
Trans Audio Electroacous, Vol AU- 17, #2, June, pp 166-169.
E. A. Sloane ( 1966), " An Introduction to Time-Series Analysis", Monographs I, II, and
Ill, Time/Data Corporation, Palo Alto, Calif.
E. A. Sloane ( 1968), "The Effects of Quantization on Small Signals", Technical Bulletin
No. 1, Time/Data Corporation, Palo Alto, Calif.
E. A. Sloane ( 1969), "Comparison of Linearly and Quadratically Modified Spectral
Estimates of Gaussian Signals", IEEE Trans Audio Electroacous, Vol AU- 17, #2,
June, pp 133-137.
E. A. Sloane ( 1970), " Measurement of Transfer Function and Impedance", General
Radio Experimenter, Vol 44, #7-9, July-Sept, pp 8-13.
M. W. Smith and R. F. Lambert ( 1960), "Acoustical Signal Detection in Turbulent
Airflow", J Acoust Soc Am, Vol 32, #7, July pp 858-866.
R. A. Thompson and B. Weichbrodt ( 1969), "Gear Diagnostics and Wear Detection",
ASME Vibration Conference Paper #69-VIBR-10, and Machine Design, Vol 41, May
15, pp 181-183.
I. Tolstoy and C. S. Clay ( 1966), Ocean Acoustics, McGraw-Hill, New York, 1966.
P. H. White ( 1969), "Cross Correlation in Structural Systems: Dispersion and Nondisper-
sion Waves", JAcoust Soc Am Vol 45, #5, May, pp 1118-1128.
B. Widrow ( 1961), " Statistical Analysis of Amplitude-Quantized Sampled Data Sys-
tems", AIEE Trans, Vol 79, Part II, Applications and Industry, 1960, Jan, pp
555-568.
J. F. Wilby and F. L. Gloyna ( 1970), "Correlation Measurements of Airplane Fuselage
Vibrations", J Acoust Soc Am, Vol 48, # 1 ( Part 1), July, p 80.

108
 Chapter 6

Instrumentation for
Noise and Vibration
Measurement
6.1 GENERAL.
Sound measuring systems use amicrophone (or, as amore general term, a
transducer) to transform the sound-pressure variations into acorresponding
electrical signal. This signal is amplified, measured and analyzed by electronic
instruments.
Although some nonelectronic instruments have been used in the past for
sound measurements, hardly any are used at present. The dominance of
electronic techniques is aresult of their versatility and extensive development
and the need in acoustics for operation over awide range of frequencies for
which those techniques are well suited.
Many vibration measurements, however, are still made with nonelectronic
techniques. When the vibratory motion is slow and large, the measurement
can sometimes be made visually, with ascale. If the motion is slow but small,
ameasuring microscope may be used. For rapid motion, astroboscope can be
used to produce an apparent slow-motion replica of the rapid motion for
optical measurement. This technique is discussed in more detail later in this
handbook.
Table 6-1

GENERAL RADIO SOUND- AND VIBRATION- MEASURING INSTRUMENTS

Sound &
Vibration Meters, Grace..
Transducers & Tape Recorder, Analyzers/ Recorders &
Calibration Accessories Preamplif ier Impact Met« Filt(111 Display Units

Sound- Level Ceramic Precision Sound- Level Meter

Calibrator Microphones and Analyzer ( Octave)
II - .se")

Electret Preamplifier Sound- Level Meter Level Recorder and

Microphones lOptional I 1/3- Octave Band Analyzer
Noise Exposure
Preamplifier Meter Sound & Vibration Graphic Level
Condenser with Bias Analyzer Recorder
Microphones Supply Impact Noise (1/? & 1 / 10 octave)
11/2, 1/4" Analyzer
1/8") Multichannel Wave Analyzer
Amplifier Cassette Data 11% Bandwidth)
Coupler, Recorder
Real Time Analyzer DC Recorder
Cables 11/1.1/3.or 1/100ct)
Storage Display
Windscreens EFT Analyzer Unit

Tripod

Vibration Accelerometers Preamplifier Vibration Meter

.5 Calibrator (Pickups) (Optional)

Z
E Cables Multichannel Cassette Data Universal
5. Magnet Clamp Amplifier Recorder Filter

Control Boxes
to use with
SLM or Analyzers

109
Figure 6-18. The
Type 1921 Real-
Time Analysis
System operat-
ing under com-
puter control.
The computer
and peripherals
are shown at left.

ID: FLU/VEX 1EST 3

ULIAATION IN SECS: 12

0JUUT: EPNL6, PM.» TPEL

STAUT

c1.11L . 10960

Pi61. ? NET
.5 91.1 9960
1.0 98.4 03.0 ...... .
1.5 0066 01.9 ....... .......»,..
e.e
. 0262 44.2 . ..
265 04.8 0661
'!.0 07.1 98.3
3.5 1,
,,„k ii.3
660 13.3 13.3
9.9 1464 19.6
5.0 11.2 1162
505 13.0 13.6 .................
6.0 12.7 1267 .. • ....... 66.....
665 11.0 1166
7.0 02.2 09.9 i•

745 05.1 0561

1.0 00.9 02.0 S SS .... te........•.*

2.5 92.1 99.2

9.0 9669 9769
965 9562 96.5
10.0 9509 9762
10.5 9269 94.3
11.0 9262 93.5
11.5 90.9 92.8
12.0 8404 86616

Figure 6-19. Typical teletype printout of an EPNL analysis of airport noise generated by
acomputer-controlled Type 1921 system.

124
 The 1-in. size (actually 0.936 in. or 23.77 mm in diameter) is the most
commonly used microphone, because it has an acceptable combination of
characteristics with regard to sensitivity, frequency response, and omnidirec-
tionality. The 1/2-in. size ( 12.7 mm in diameter) has abetter high-frequency
response and is less directional but these improvements are obtained at the
expense of sensitivity.
Typical responses for the two sizes of ceramic microphones are shown as a
function of frequency in Figures 6-1 and 6-2. The responses are usually given
for particular values of incidence of the sound on the microphone. For a
cylindrically symmetrical microphone with the diaphragm mounted perpen-
dicular to the axis of symmetry, sound propagation along the axis toward
the microphone is called "0° incidence" and sometimes " perpendicular inci-
dence." Propagation perpendicular to the axis is " 90 ° incidence" or "grazing
incidence."
When the angle of incidence of sound is equally likely to be any value, as is
essentially the situation in a highly reverberant room, an averaged value of
response is calculated from the total directivity characteristic. This averaged
response is called the "random-incidence response." It is the one used for
rating response in the American National Standard for sound-level meters.
For low frequencies, as would be expected, the microphone is essentially
omnidirectional. As the frequency of the sound increases, its wavelength
becomes more nearly comparable with the dimensions of the microphone and
directional effects are noticeable. At agiven frequency the directional effect
for the smaller microphone is correspondingly less than for the larger one.

6.2.2 Condenser Microphones. Another type of microphone, known as the

condenser, electrostatic or capacitor microphone, is used for measurement
purposes ( Hawley, 1955; Rasmussen, 1960/1963). Again adiaphragm is used
and it is set in motion by the sound pressure. Here, the variation of an
electrical capacitance, formed between the thin, stretched diaphragm and a
backplate, is used to develop an electrical signal when a high polarizing
voltage is applied to the capacitor.
Responses as afunction of frequency for such microphones are shown in
Figure 6-3. These small microphones have excellent response to high frequen-
cies and are used for wide-frequency-range acoustical investigations.

6.2.3 Electret Microphones. Another microphone of the condenser type

uses a thin, plastic sheet, which has a conductive coating on one side. The
other side of the sheet rests on a perforated, metallic backplate with many
small, supporting raised points. The sound pressure is applied to the plastic
surface, which moves with respect to the backplate, thus varying the capaci-
tance between the conductive coating and the backplate. By the use of a
suitable plastic and proper treatment, the microphone maintains its own
polarization, and the capacitance change generates a corresponding electrical
voltage. When the microphone is self polarized in this way, it is called an
"electret microphone" ( Sessler and West, 1966; Djuric, 1972). This micro-
phone can also be built to have an excellent combination of characteristics.

6.2.4 Hydrophones. Microphones used for underwater sound measure-

ments are called hydrophones ( Figure 64). They also generally use piezoelec-
tric ceramics as the sensitive element (Tucker and Gazey, 1966). Various
types of hydrophones are available from such companies as Atlantic Re-

111
 •4
I 1 1
RANDOM- INCIDENCE
PERPENDICULAR- INCIDENCE

I; 0 .ir
- P --- ---

•
50 70 100 200 500 700 2k 5 r r 20 50k
FREQUENCY IN HERTZ

Figure 6-3. Frequency characteristics of two different half in. condenser microphones.
They optimize response for specified sound incidence.

search, Chesepeake Instrument Corporation, Clevite Ordnance, Gulton Indus-

tries, Scientific-Atlanta, and Wilcoxon Research.

6.2.5 Vibration Pickups. The vibration pickups supplied by General Radio,

as shown in Figure 6-5, are all inertia-operated, lead-zirconate, lead-citanate,
piezoelectric devices that generate a voltage proportional to the acceleration
of the pickup ( Dranetz and Orlacchio, 1961; Mason, 1961, Carlson, 1952).
Voltages proportional to velocity and displacement of the vibrating body are

Figure 6-4. A typical hydrophone.

Figure 6-5. Typical appli-

cation of a vibration pick-
up. It is mounted on the
end bell of the left motor
in the foreground and the
vibration signal is being ana-
lyzed by a General Radio
automatic-recording 1/3-
octave-band analyzer.

112
also obtainable by the use of electronic integrating networks to convert the
voltage generated by the pickup. This type of pickup has the advantage of
small size, light weight, and wide frequency range, and it does not require a
fixed frame of reference for the measurement.

6.2.6 Transducer Sensitivity. Microphone sensitivity is now rated in terms

of the reference sensitivity of 1-volt output for a pressure of 1 newton/
(meter) or i n terms of the older reference, 1 Wear. The apparent sensitiv-
2

ity levels for these two ratings differ by 20 dB; a pressure of 1 N/m 2
corresponds to 94 dB re 20 µN/m 2 ,whereas 1µbar corresponds to 74 dB re
20 µN/m 2 .A typical microphone open-circuit (unloaded) sensitivity level is
—40 dB re 1V/(N/m 2 ). If the sound-pressure level were 134 dB re 20 µN/m 2 ,
the open-circuit output from this microphone would be 1V ( see Figure 6-6).
The sensitivity of an accelerometer is commonly rated in terms of open-
circuit (unloaded) output in millivolts (.001 V) for an acceleration that
corresponds to that of gravity (g). Since the acceleration of gravity (9.80665

10,000 mV
7,000
5,000

2,000

1000 mV
700
500

200

100 mV
70
50

20

10 mV
7
5

1— 2

;5%0°
o

,?•. 200

100 µV
70
50

20

10µV
7

1µV
07
0.5

02

0 1b. V0
10 20 30 40 50 60 70 80 90 100 110 120 130 140
SOUND PRESSURE LEVEL ( SPL) dB

Figure 6-6. Comparative transducer sensitivity levels. 156053

113
m/s 2 or 386.09 in./s 2 )is always directed toward the center of the earth, an
alternating acceleration rated in terms of "g" units is to be interpreted as a
use of the numerical value only. As such, it is sometimes used for the peak
value of asinusoidal acceleration or the rms value (0.707 times the peak for a
sinusoidal vibration). As long as it is recognized that the electrical output and
the acceleration are to be measured in the same way, rms and rms, or peak
and peak, it is not necessary to specify in the sensitivity statement which is
meant; the numerical sensitivity should be the same for both.
The relations among velocity, displacement, and acceleration depend on
the frequency as well as the amplitude ( Appendix IV). To find the transducer
output for a given frequency and velocity or displacement, convert the
vibration to the equivalent acceleration (Appendix IV, Figure IV- 1and IV-2),
and then the chart of Figure 6-7 can be used.
6.2.7 Preamplifier. The preamplifier shown in Figure 6-8 is a high-input-
impedance, low-noise amplifier ( Marteney, 1970). It is particularly well suited
for amplification of the output of capacitive sources, such as ceramic and
condenser microphones and piezoelectric vibration pickups, and for use with
sound-level meters and analyzers when along cable must be used between the
microphone and the instrument. A gain of 20 dB ( 10:1) is available for
increasing the ultimate sensitivity of analyzers for low-level measurements.

PARAMETER ON LINES IS ACCELERATION LEVEL

dB re 10- ' 0 / s2

'

,k9

o
_
1.9

2
0
0
- 01
.
A%
‘A
I.. o
..,
o. 4

,
2
cc
a %. •
001
cn omae eb
0°-
- 4 0

ci•
0

.0001

C>

00001 J 1
110,..v) 4 8 2 4 8 2 4 8
10 100 1000
TRANSDUCER SENSITIVITY - rnvig

Figure 6-7. Chart to determine transducer output for various acceleration parameters.

114
 Figure 6-8. A typical field installation for micro-
phone use. The tripod mount is shown supporting
a Type 1560-P42 Preamplifier, into which a 1/2-
inch microphone has been plugged. The foam
sphere at the top is a windscreen that fits over the
microphone to permit accurate measurements out-
doors.

The preamplifier shown can also provide a 200-V polarizing voltage for a
condenser microphone. It obtains its own power from another instrument or
aseparate supply.
6.2.8 Wind Screen. In order to reduce the wind noise produced when a
microphone is placed in a wind, awind screen, shown in Figure 6-8, should
almost always be used outdoors.

6.2.9 Tripod. A microphone and preamplifier is often best suspended in

position by its connecting cable. When this arrangement is not possible, a
tripod with suitable adaptor devices can be used to hold the microphone in
the desired position ( see Figure 6-8).

Figure 6-9. The Type 1566 Multi-

channel Amplifier shown mounted
in a field transportable automatic
sound analysis system in which
it scans the inputs of up to 16
remote microphones.

115
 6.2.10 Multichannel Amplifier. Many sound and vibration measurements
can be simplified by use of a scanner that connects, in sequence or in any
arbitrary order, the outputs from a number of transducers to a single
metering or analyzer system. The unit shown in Figure 6-9 scans up to 16
channels and amplifies each.

6.3 SOUND- LEVEL METERS.

The basic instrument of a sound-measuring system is the sound-level
meter. It is a portable meter for reading in terms of a standard reference
pressure ( 20 µN/m 2 )the sound level at its microphone. Fundamentally, the
instrument consists of an omnidirectional microphone, acalibrated attenua-
tor, astabilized amplifier, an indicating meter, and weighting networks. The
networks provide the three common sound-level meter responses, A, B, and C
(see Figure 2-3).
Four types of sound-level meters are specified in the latest American
National Standard Specification for Sound-Level Meters, S1.4-1971. These
are called Types 1, 2, 3, and S or " Precision," "General Purpose," " Survey"
and " Special Purpose," respectively. The first three types differ in their

Figure 6-10. Sound level meters a. Type 1933

Precision (ANSI Type 1) b. General Purpose
(Type 2) c. Survey ( Type 3).

116
performance requirements, with the requirements being most strict for the
Type 1 or Precision Sound Level Meter and progressively less strict for the
types 2 and 3. The special-purpose sound-level meter is one that meets the
requirements of one of the other types, but does not contain all three
weighting networks.
The three basic types are available and are shown in Figures 6-10. They are
all small, light in weight, and easy to use. They can be mounted on atripod,
held in the hand or placed on table or bench with equal facility. Readings and
settings are easily made with the microphone in a vertical or horizontal
position.
The units shown also include an output for driving analyzers, recorders,
oscilloscopes, and earphones.
The Type 1933 Precision Sound-Level Meter and Analyzer ( Kundert and
Marteney, 1971) also includes aspecial detector circuit for measuring impulse
sounds ( paragraph 6.9) and it has a set of octave-band filters ( paragraph
6.6.1). This instrument is the key one in thf! 1933 Sound Analysis System.
This system permits one to make awide variety of measurements in the field,
and its versatility is aresult of the features included in the 1933.
In the precision sound-level meter shown, the microphones fit atop a
telescoping 18-in, extension to reduce the effects of the instrument and
operator on the source field. When the operator and instrument must be
removed even farther from the microphone, a cable can be used, and a10-ft
cable is included. The measurements are not affected by the use of the cable

Figure 6-11. The Type 1934 Noise Exposure Monitor, shown in an industrial application.

117
because the preamplifier in the 1933 is detachable and connects to the cable
at the microphone end to prevent loss in signal level.

6.4 NOISE- EXPOSURE METER.

An instrument for evaluation of noise exposure according to the regula-
tions of the U.S. Department of Labor (OSHA, 1970) is shown in Figure 6-11
(Partridge, 1971). It is aspecial-purpose sound-level meter, which includes a
calculator to do the measuring, accumulation, exposure calculation, and even
impact detection automatically. The result is read out on adigital display
and is also available as digital and analog electrical signals for use with
recorders or printers. The recorded time of occurrence of high exposure levels
is helpful in determining how to reduce the noise-exposure level.
Pushbuttons permit the user to select the appropriate length of working
day for the exposure calculation. At the end of this measurement period, the
instrument holds the noise-exposure percentage until it is reset. At any
intermediate time one can flip aswitch to read the percentage of allowable
noise exposure accumulated to that instant.
A security case is also available to reduce the possibility of tampering with
the instrument or measurement.

6.5 VIBRATION METER.

The instrument for vibration measurements that corresponds to the
sound-level meter is the vibration meter, shown in Figure 6-12. It takes
advantage of the wide frequency range of the piezoelectric type of pickup
with a response for the measurement of acceleration extending smoothly
from 2 to 20,000 Hz. The meter is calibrated directly in terms of peak,

Figure 6-12. Type 1553 Vibration Meter shown with pickup connecte°.

118
 POWER ACCELEROMETER
AMPLIFIER AMPLIFIERS MULTIPLEXER

OSCILLOSCOPE

SHAKER

D/A

X-Y CONVERTER A/D

PLOTTER CONVERTER

INVERSE PS D
FOURIER ANALYSIS
TRANSFORM I T/D - 90
FOURIER -
TRANSFORM
PROCESSOR

SYSTEM JUNCTION UNIT

L
-
OPERATORS
CONSOLE
COMPLEX INPUT PSI) REFERENCE
IDEC
FOURIER SPECTRUM COMPARATOR PSD
POP- Ii
SPECTRUM 4- MODIFIER 4^ STORAGE
I COMPUTER
GENERATOR
ASR 33
TELETYPEWRITER

Figure 6-
30. System configuration for signal analysis and control of shaker system.

scopes have provision for the addition of a camera, which makes possible a
permanent record of the wave shape being studied.

6.15 VIBRATION SHAKERS.

Several types of vibration shakers are widely used ( Booth, 1963). One of
the most versatile is the electrodynamic shaker. These shakers, produced in a
wide range of sizes, are used by vibration engineers in many ways to help
evaluate performance of instruments, components, and structures. Typical
uses are: endurance or fatigue testing of electrical and mechanical structures,
testing of resilient or shock mounts, shake testing of electrical components
such as switches, relays, or amplifiers, determination of damping characteris-
tics of materials, and calibration of vibration pickups.
Some tests use sinewave motion, with the frequency either set to a
resonance of the device under test or swept over aspecified band. Random
motion is becoming widely accepted in vibration testing, with arandom-noise
generator used as the signal source (see paragraph 6.10), and an adjustable
band-pass filter used to shape the noise spectrum.
Since the motion of the shaker is affected by the structure fastened to it,
the drive system cannot always be readily preset to produce a required
motion of the device being tested. The motion at the fastening points of the
device may be monitored with vibration pickups and analyzed to check the
spectrum. This information can then be used to set the shaker drive signal to
the proper spectrum shape and level. The ultimate approach is to arrange the
system to monitor and control the level automatically. This approach be-
comes very involved for random vibration and requires acontrol system of
the complexity shown in Figure 6-30.

135
 1I 11I
TYPE 1560-PIIB VIBRATION PICKUP SYSTEM
FREQUENCY RESPONSE CHARACTERISTICS

Ha°

CI DJ

20 c 50 100 200 500 lk

FREQUENC ( IN HERTZ ragian

I I I I I

. . . . .„
TYPE 1560-PI4 V BRATION PICKUP SYSTEM
FREQUENCY RESPONSE CHARACTERISTICS a
&
9

0
- 10
20 cis 50 100 200 500 lk 2 5 10
FREQUENCY IN HERTZ

TYPE 1560-P13 VIBRATION PICKUP SYSTEM

FREQUENCY RESPONSE CHARACTERISTICS Lt.
d o o
>

50 100 ZOO 500 ik 10 20

FREQUENCY IN HERTZ

Figure 6-14. Over-all frequency response of Type 1560-

Pl 18, 1560-P13, and 1560-P14
Vibration Pickup Systems and sound-level meter for constant applied acceleration,
displacement, and velocity.

with each vibration pickup system gives the proper conversion factors for that
system when it is used with asound-level meter. By means of these data, plus
the decibel table in the Appendix, the readings can be readily converted to
in. ( displacement),inls(velocity), or in./s/s(acceleration), or to the proper
metric units.
Three such pickup systems are available. The Type 1560-P11B Vibration
Pickup System is a low-cost unit for general use, the Type 1560-P13 is for
high-frequency measurements, and the Type 1560-P14 is for very low-vibra-
tion-level measurements. Typical over-all frequency response characteristics
for these systems used with asound-level meter are shown in Figure 6-14.

6.6 ANALYZERS.
Even if asound-level meter were perfect ( i.e. fit with no tolerance all the
design objectives of the ANSI or IEC Standards), the reading obtained by it in
any given noise field is inadequate for a complete understanding of the
problem. The number of decibels indicated by a sound-level meter tells
nothing about the frequency distribution of the noise. It is true that by
judicious use of the weighting networks in asound-level meter one can learn
something about the frequencies present, but this knowledge is only qualita-
tive. For most important problems it is necessary to use some type of
frequency analyzer to determine the noise spectrum, as described in Chapter

120
5. It is also often helpful to measure the correlation of two noise signals, the
transfer and coherence functions, and the other measures described in Chap-
ter 5.
The vibration meter measures the displacement, velocity, acceleration, or
jerk of avibration. Unless the waveform is substantially sinusoidal, however,
the vibration meter by itself gives little information about the frequencies of
the individual vibration components. An analyzer, therefore, is desirable and
often is anecessity. As with noise, the analysis of vibration provides clues to
the sources of the vibration components and information necessary in the
suppression of the vibration.
A number of analyzers are available for use with the sound-level meter or
the vibration meter or for use with microphones and vibration pickups
directly or with preamplifiers. These analyzers vary in cost, complexity, and
ease of operation. Choice among them is generally determined by the amount
of detailed information needed, the speed of processing required, the nature
of the output format, and the auxiliary processing that may be required.
The simple, serial analyzers will be described first. These analyzers can be
hand operated, and the band levels can be read from an attenuator setting and
a meter reading. Some of them can also be coupled to recorders to yield a
descriptive plot of the band levels as afunction of frequency.
The analyzing systems that can provide detailed data rapidly will then be
described.

6.6.1 Octave-Band Analyzers. The Type 1933 Precision Sound Level Meter
and Analyzer, shown in Figure 6-10, includes an octave-band filter set that
makes possible the simple and rapid analysis of noises having complex spectra
(Kundert and Marteney, 1971). As described in Chapter 8, it is widely used
for frequency analysis of noise, particularly for product rating, production-
line testing, preventive maintenance, checking for compliance with some
ordinances, and for estimating some subjective effects.
This portable, battery-operated instrument is acomplete sound-level meter
with a microphone, a preamplifier, an attenuator, weighting networks, an
amplifier (which drives the filters), an indicating meter, and monitoring
outputs. The set of octave-band-pass filters, selected by means of arotary
switch, range in center frequency from 31.5 to 16,000 Hz. The direct-reading
level range with the microphones supplied is 10 to 150 dB re 20 ¡.LN/m 2 .The
filters meet the requirements of ANSI S1.11-1966, Class II, Type E.

o
,o
i OCTAVE
u - 20

3°

1 40
1
if, OCTAVE
2 50

60

70
fo 2o

Figure 6-15. The Type 1564 Sound

and Vibration Analyzer shown with
selectivity characteristics.

121
Figure 6-16. Type 1523 Graphic Level Recorder being operated with the Type 1523-P3
plug-in to make automatic stepped 1/3-octave-band analyses of compressor noise.

6.6.2 Third-Octave-Bandwidth Analyzer. For more detailed analysis of

noise, a third-octave-band analyzer, such as that shown in Figure 6-15, is
often used ( Kundert, 1963). In addition to the third-octave band, this
instrument also has aone-tenth-octave band. A typical selectivity characteris-
tic for these two bands is also shown. This instrument can be tuned to any
center frequency between 2.5 and 25,000 Hz, and the shape of the selectivity
curve is constant in terms of percentage of the resonant frequency over the
entire range.
The meter scale is calibrated in decibels for use in sound measurements
and in linear units for vibration measurements. The direct-reading range for
sound-pressure levels with a microphone attached is 44 to 150 dB. With the
Type 1560-P40 or P42 Preamplifier, the minimum is extended to 24 dB and
long cables can be used without loss in sensitivity.
This analyzer can be operated from the power line or from arechargeable
battery.
The filter characteristics meet the requirements of ANSI S1.11-1966, Class
II, Type E.
Third-octave analysis is now widely used, particularly for checking compli-
ance with noise and vibration specifications. It is most often used with a
graphic-level recorder to give agraph of the energy distribution of the noise
and vibration as afunction of frequency (see paragraph 6.7).
The analysis of a noise is made more automatic with the stepped 1/3-oc-
tave-band analyzer system shown in Figure 6-16. Here, the analyzer is part of
alevel-recorder system, and successive, preferred, 1/3-octave bands are selec-
ted as the recorder chart moves and the pen records the band level. The range
of 1/3-octave-band center frequencies is from 1 Hz to 500 kHz, and the
beginning and final frequencies are selectable within that range.

122
 o
/11111111111111111111111111111
11111111111111111111111111111
11111111111111111111111111111
mornimmuu mum
1111 111111111111in 1111
Ill 11111111111111111 III
11 1111111111111111111111111
1 1111111111111111111111111
1U1111111111111111111111111
W1111111111111111111111111
11111111111111111111111111n
,rrftrumrfrrrfrcrigrurrIf

1„
-

THIRMOCIAMIMAMO CrI•11101 IIIIMOUIEMCII WI IM

Figure 6-17. Type 1921 Real Time Analyzer (left) with Type 1921-P1 Display Scope
and Type 1522 Recorder shown recording a measurement. Enlarged displays at right.

A wide choice of averaging time is selectable. In addition, the averaging

can also be automatically programmed to decrease with increasing frequency
in such afashion that essentially aconstant confidence level is achieved.
In order to achieve the most rapid 1/3-octave analysis, acomplete set of
filters must be driven in parallel. This technique is used in the Type 1925
Multifilter, ( Kundert, 1968a). It can include up to 30 such filters anywhere in
the frequency range from 3.15 Hz to 80,000 Hz. Full octave or 1/10-octave
filters in that range are also available. These filters meet the requirements of
IEC 225-1966 and the highest applicable requirements of ANSI S1.11-1966.
Attenuators for each filter channel are also available. These can be set to
compensate for irregularities in frequency response of other parts of asystem,
or to weight the band level in any desired way.
The detector system for this multifilter is the 1926 Multichannel RMS
Detector ( Lapointe, 1969). It combines with the multifilter to form the Type
1921 Real-Time Analyzer, shown in Figure 6-17. This system has a 70-dB
dynamic range with integration times selectable from 1/8 sto 32 s. The band
numbers and measured band levels can be displayed on high-intensity neon-
readout tubes, or the complete results of aspectrum analysis can be displayed
on the 1921-P1 Storage Display Unit.

123
measured by acoustic instruments but are derived from them by various
calculations.
Here are some of the measures and effects we may wish to calculate or
predict from acoustical measurements.
1. Human reactions
Loudness and loudness level
Perceived-noise level
Articulation index
Speech-interference level
Hearing-damage risk
Noise criteria
2. Speech
Spectrum vs time
Formant tracking
3. Source characteristics
Spectrum vs time
Sound power
Directivity
Loudness efficiency rating of loudspeakers
4. Room acoustics and materials testing
Reverberation
Transmission loss
Absorption
5. Statistical measures — searching and sorting
Signature analysis and comparisons
Averaging or integrating multiple runs
Maxima of aseries of spectra
Amplitude distribution
Correlations
6. Miscellaneous
Microphone corrections
Weighted levels
In the past, most of the above have been calculated from analog measure-
ments without the use of an electronic computer, or in some instances special
analog devices have been designed to permit the direct measurement of the
desired result. Computer systems will be increasingly used for these acoustic-
measurement problems because of the benefits that such systems can provide.
It takes 3or 4 minutes to calculate loudness level or perceived noise level
with adesk calculator from 25 third-octave band levels, and it may take even
longer to calculate the discrete-tone corrections for perceived noise level.
Furthermore, there aren't many who can do anumber of these calculations,
even with adesk calculator, without making mistakes.
If an electronic computer is programmed to do this calculation, the
calculation time is insignificant. It may take asecond or so for the teletype-
writer to print out the answer, however.
Accurate and rapid calculation of the desired end result is only one of the
benefits of a computer system. We shall describe others in the role of the
computer as asystem controller.
Computer as System Controller. As acontroller, the computer makes the
various elements operate together as acomplete, versatile system. The com-
puter stores the parameters required for one or more tests, accepts particular

143
 The results of the analysis can also be recorded on the 1522 Dc Recorder,
also shown in Figure 6-17 and described in paragraph 6.7. A sample of such a
recording is shown as well.
Since the 1921 Real-Time Analyzer is programmable by digital signals and
the spectrum is available in digital form, it can be combined with acomputer
to provide on-line calculations of effective perceived noise level ( EPNL)
according to FAA specifications, or ASHRAE, ARI, AMCA, and SAE ratings
of anoise, make vibration signature comparisons for preventive maintenance,
etc. Such a system is shown in Figure 6-18, and a teletype printout of an
EPNL calculation from the system is shown in Figure 6-19.
6.6.3 One-Percent-Bandwidth Wave Analyzer. When still finer detail of
analysis is desired, the Type 1568 One-Percent-Bandwidth Wave Analyzer,
shown in Figure 6-20, may be used ( Kundert, 1966). It has avery selective
tunable filter covering the frequency range of 20 to 20,000 Hz. The shape of
the selectivity curve is constant in terms of the percentage of the resonant
frequency over the entire range. The meter scale is calibrated in linear units
for vibration measurements and in dB for sound measurements.
When a Type 1560-9531 1-in, ceramic microphone set, (for which the
analyzer supplies the power), is connected to the analyzer, component levels
from 24 to 128 dB sound-pressure level can be measured. The output of the
analyzer will drive agraphic-level recorder so that an automatic recording of a
noise or vibration spectrum is readily obtained.

6.6.4 Wave Analyzer. The Type 1900 Wave Analyzer uses afixed-frequen-
cy filter in atunable heterodyne system similar in principle to the common
superheterodyne radio receiver (Peterson, 1964). It is continuously tunable
from below 20 Hz throughout the audio band in asingle sweep of the main
tuning dial. The resulting filter characteristic is constant in response, with
respect to the number of hertz deviation from the center frequency, over the
entire tuning range.
This characteristic is convenient for analyzing random noise, because the
spectrum level is obtained by a constant correction of the indicated level.
Most such analyzers are narrow in bandwidth, however, and an analysis of
noise must then proceed slowly because of the long averaging time required.
A significantly wider band, such as 50 Hz, which is available on the Type
1900 is very much faster and relatively easy to use for noise analysis. Narrow
bands of 3and 10 Hz are also provided.

05 10 2O 30 10.0
NORMALIZED FREQUENCY .•••

Figure 6-20. The Type 1568 Wave

Analyzer shown with attenuation
characteristics of the filter.

125
 This analyzer has an output for recording and a linear frequency scale.
When an analysis that is linear in frequency scale is made, one can readily
track down harmonic relations, since successive integral harmonics are spaced
uniformly. Thus, the analysis of rotating or reciprocating machinery, includ-
ing gear trains, electric motors, and turbines, by awave analyzer is often to be
preferred to other types of analysis.
The analyzer has an electrical output arranged so that the system is a
continuously tunable filter. Thus, one can listen by means of earphones to
the component or band selected by the analyzer. Furthermore, if one applies
a broad-band noise signal to the input, one can obtain at the output anarrow
band of noise, preferably 50-Hz wide for most acoustic measurements, whose
center frequency is continuously tunable over the full range of the analyzer.
This signal is desirable for some acoustic tests of rooms, walls, and hearing.
Another mode of operation of the analyzer yields asine-wave signal at the
output that is always at the frequency to which the analyzer is tuned. This is
then a convenient source (to drive an amplifier and speaker) and detector for
over-all electrical or acoustical response measurements.

6.6.5 Fast Fourier Transform Analyzers. As explained in Chapter 5, digital

techniques of time-series analysis are providing another versatile approach to
the study of acoustic and vibration signals, including not only spectrum
analysis, but also transfer functions, coherence functions, cross spectra, and
auto and cross correlation. The T/D 1923 Analyzer, shown in Figure 6-21, is
a pushbutton-controlled, programmed system of this type (Chamberlain,
1971). It has been built to be exceptionally versatile and yet easy to use. The
basic functions are selectable by pushbuttons, and the various stages in the
process can be viewed on the storage display oscilloscope. The signal can be
viewed initially as afunction of time. When aframe of input points is selected
for analysis, they can be displayed to ensure that the desired section was
selected. The frame can be transformed, and the result displayed to show the
analysis.
Spectrum averaging for noise signals, and rectangular or hanning weighting
of the input data, are also provided for improving processing. The buffered-
mode of operation permits real-time continuous processing of noise signals,
up to a certain frequency range, for processing all the data for maximum
statistical accuracy. The upper frequency limit for this real-time mode de-
pends on the function performed. For auto-spectral calculations, the 1923-C
can operate in real time to about 500 Hz and the 1923-A to about 25,000 Hz.
Although the basic analysis provided by the FFT is one with uniform
resolution in Hz, the results can be transformed into an equivalent 1/3-octave
analysis for comparison with such data. This operation is not as efficient for
the third octaves as the parallel processing of the 1921 system, but it does
permit use of the same processor for detailed analysis with a frequency
resolution of as fine as .025% of the desired maximum frequency to 50 kHz,
as well as in 1/3-octaves or octaves, depending on the programming specified.
The FFT can also provide phase information, which may be useful in
transfer-function studies or in tracking down troubles in repetitive systems,
where the time of occurrence of related events is significant.
An alphanumeric display of the selected function and control settings is
furnished. The output can be preserved on aprinter or arecorder.
Since the analysis results are in the proper form for computer use,
additional processing of the data can readily be done in the computer by the

126
 1,1›, 1-1.1 111111111- 1 HU

AC I

Figure 6-21. T/D 1923 Real- Time Time-Series Analyzer (left). Diagram suggests possible
expansions towards complete on-line systems. Scope detail, lower right.

use of additional programming. For example, the spectrum signature can be

compared with stored criteria, or with previous signatures, to determine if
action or further study is necessary. The spectrum can be modified by a
calculation routine to yield arating of the noise.
The scope of applications for these versatile processors can be seen from
comments about some specific applications. In noise and vibration control,
there is noise and vibration source, path, and direction identification. In
addition, for preventive maintenance, the spectrum is sometimes used to
identify incipient troubles. This technique is sometimes called signature
analysis. The processing techniques used for these purposes are not only the
straightforward spectrum analysis, but also the transfer and coherence func-
tion to identify probable cause-and-effect relations, the cross-correlation
function to relate two or more signal pick-up points in time, waveform
averaging and spectrum analysis to dig a repetitive signal out of noise to find
its waveform or spectrum, and spectrum averaging to obtain a statistically
stable spectrum of astationary noise signal.
The processing capabilities of such devices are so broad in frequency range,
time span, and function that these techniques are being applied in many other
areas beyond the scope of this book. Examples can be found in medicine,
biophysics, economics and business analysis and forecasting, physics, com-
munications, and geophysics.

127
 -111111 el Di
•
Olt
•

rerrrvirivrvEus
31111111 Inaxl1211.111ILISII.11,111.1.311,11y1

"•i t"

I 1111111111111111111111111111111111
11111181811110111111111111111181111181111111111111110111110811

Figure 6- 22. Type 1523 Graphic Level Recorder with -P3 1/3 Octave Band Analyzer
plug-in accessory at right. Typical chart record shown below.

6.7 GRAPHIC RECORDER.

Graphic recorders of various types are used to produce a permanent,
reproducible record of the results of a measurement. As an accessory to
sound and vibration instruments, they can be used to record over periods of
time the sound level near highways, airports, residences, or the vibration
levels of building floors or walls, bridges, or airframes and to measure
reverberation time. The resulting information is much more extensive than
that obtainable from afew readings of ameter, and when observations over a
long period are desired, the recorder can be unattended.
Some recorders can be used with an analyzer to plot the frequency
spectrum of a noise source or of a vibrating object. The combination of a
recorder with atunable oscillator and other accessories can produce records
of the acoustic-transfer characteristics of rooms, walls, microphones and
loudspeakers, the electrical response of analyzers, networks, and amplifiers,
and the vibration transfer characteristics of shakers, vibration pickups, and
stru ctures.
The 1523 Level Recorder shown in Figure 6-22, is the basic element in a
system to do these many different operations. It is aservo-type strip-chart,
pen recorder that uses adisposable cartridge. It has awide range of speeds for
the chart drive, of averaging times for the detector system, and of level ranges
for recording. The maximum level range is 100 dB. Limit comparators are
included for use in production test systems.
When this recorder is used with the appropriate plug-in, it forms a
level-recorder system, and it can be used for the applications of recording as a

128
function of time mentioned above. When the recorder is used with the
1523-P3 Stepped 1/3-Octave-Band Analyzer plug-in, it becomes arecording
analyzer as described in paragraph 6.6.2. When it is used with the 1523-P2
Sweep Oscillator plug-in and other accessories, it can record responses and
transfer characteristics as afunction of frequency.
This sweep oscillator covers the frequency range from 1 to 500,000 Hz
and the upper and lower limits of the sweep can be selected in the range of 5
to 200,000 sec/10:1 frequency change. The time/decade can be maintained
constant, or it can be set to decrease ( increase speed) with increasing fre-
quency. This decreasing sweep time is usually satisfactory for measurements
on most physical systems, and it saves measurement time compared with
maintaining the slow rate required for the low-frequency end of the range.
The 1522 Dc Recorder shown in Figure 6-17, is aversatile, programmable
strip-chart recorder that is particularly well suited for recording the results of
an analysis on the 1921 Real-Time Analyzer, which it will do much faster
than the conventional X- Y plotters ( Basch, 1969).
A number of the manually operated analyzers can be driven mechanically
from the 1521-B Graphic Level Recorder and the results of the analysis can
be automatically plotted ( Basch, 1964). Figure 6-5 shows the 1911 Recording
Sound and Vibration Analyzer for producing continuous frequency plots of
the 1/3 or 1/10-octave spectrum of sound and vibration signals. Figure 6-23
shows the chart output of the 1913 Recording Wave Analyzer for 1%-band-
width plots of the spectrum of acoustic and vibration signals.
A d-coutput is available from the 1933 Precision Sound-Level Meter and
Analyzer that is directly proportional to the indicated level in decibels over a
40-dB range. It can be used to drive a simple recorder, when the sound level
or vibration level is to be recorded as afunction of time.

6.8 MAGNETIC- TAPE RECORDER.

The magnetic-tape recorder has become a useful tool for the acoustical
engineer both in research and in development, ( Kamperman, 1958; Tall,
1958; Mee, 1964; Peterson, 1967). It stores a signal as variations in the
magnetic state of the particles on the tape. The time scale then becomes a
length scale on the tape.
The signal to be stored must be supplied to the recorder as an electrical
signal; and, for recording noise as afunction of time, this electrical signal is
usually obtained from a high-quality microphone. When measurements are to
be made on the stored signal, the recorded tape is played back on the
recorder and measurements are made on the electrical output signal.

Figute 6-23. Frequency spectrum analysis of a 1.0 ms pulse at a 70 Hz repetition rate.

The 1% bandwidth yields high resolution at low frequencies, shows the envelope at high
frequencies.

129
 The magnetic tape recorder is being used to perform the following func-
tions in the field of noise measurements:
1. To keep reproducible records of progressive changes in asound. These
changes may be a result of the application of successive noise-control
procedures, for example.
2. To record a noise for analysis by a number of techniques, when the
particular approach to be used is not at first obvious and it is not
convenient to use the original source repeatedly.
3. To record a noise in the field for detailed study in the laboratory,
where more complex instrumentation systems can be used.
4. To record asound that varies with time. Samples can then be selected
from the recording for analysis to obtain the change in spectrum as a
function of time.
5. To record a short-duration sound, which can then be played back
repetitively to simplify analysis.
6. To monitor over long periods to catch intermittent sounds, which can
then be separated out for analysis.
7. To record noises that are erratic or intermittent, possibly by binaural
techniques, to aid in tracking down sources.
8. To record a noise to permit afrequency translation for convenience in
analysis.
9. To record a transient noise in order to change the time scale or to
invert the time scale for ease of graphic recording.
10. To permit subjective or objective comparison among sounds recorded
at different times. The subjective judgment can then be made by groups
listening under similar conditions.
11. To permit observation of the subjective effects of altering asignal, for
example, by filtering, clipping, or adding noise.
12. As a measurement system with arecorded signal as the source and a
recording channel as the detector, for example, in the measurement of
reverberation characteristics.

Figure 6-24. Type 1935 Cassette

Data Recorder storing inputs from
a GR 1933 SLM.

130
 1562 1953

Figure 7-2. Octave-band analysis.

Acoustic Power Output. The use of acoustic-power output for rating

noisy devices is widely recognized as the best approach for certain measure-
ments. Acoustic power is calculated from the results of a number of sound-
pressure level measurements, usually octave or 1/3-octave-band levels. The
procedure requires a controlled environment, usually an anechoic room, a
reflecting floor in an anechoic room, or areverberation room. Under certain
conditions the requirements on the environment may be relaxed. The instru-
mentation used here covers awide range.
For example: It can be a group of microphones scanned with a 1566
Multichannel Amplifier and analyzed with a 1921 Real-Time Analyzer. The
1921 may be adjusted to read the sound power level in each 1/3-octave band
directly. If the calculation of directivity factor or other data reduction is
required, asmall computer may be connected to the 1921. Calculated results
can be printed out on a Teletype. In place of agroup of microphones, it can
be a moving microphone that scans agiven area as its output is analyzed on a
1921 Real Time Analyzer, with the averaged output plotted on a 1522 Dc
Recorder. It can be a1933 Precision Sound Level Meter and Analyzer with its
microphone on a tripod. An analysis at each of a number of microphone
positions is made. Here, the results are recorded and calculated by hand.
Product Noise and Test Codes. In order to be more specific, some
examples of instrumentation for certain test codes will be given.
The American Society of Heating, Refrigerating, and Air-Conditioning
Engineers ( ASHRAE) has prepared a set of standards* for measuring the
noise radiated by heating, refrigerating, and air-conditioning equipment.
These standards are used by the Air-Conditioning and Refrigeration Institute
(ARI) to obtain the basic data in their standards for rating the equipment.
The preferred basic instrument for this use is the 1921 Real-Time
Analyzer. It provides the highly selective third-octave filters and the long
averaging times required by the standards.
A group of microphones scanned with a 1566 Multichannel Amplifier, or
a microphone that moves over a long path as the sound-pressure-squared is
integrated in the 1921, may be appropriate.
The Compressed Air and Gas Institute (CAGI) and the European Com-
mittee of Manufacturers of Compressed Air Equipment ( PNEUROP) have
prepared a " CAGI-PNEUROP Test Code for the Measurement of Sound from

*See Appendix VII for list of standards and codes.

153
Pneumatic Equipment," ANSI S5.1-1971. It specifies procedures and operat-
ing conditions for the equipment, and it requires octave-band analysis of the
noise at anumber of points near the equipment.
A precision sound-level meter and a calibrator are required, and the
observer and measuring instrument are to be at least one meter away from the
microphone. These requirements lead to the use of the 1933 Precision
Sound-Level Meter and Analyzer with 10-feet of cable, atripod, and a1562
Sound-Level Calibrator, which are all apart of the Sound Analysis System.
If the measurements are to be part of a production test, various degrees
of sophistication can be used in the instrumentation, procedures, and setup to
speed up the measurement. Since 5 or 10 measurement locations are speci-
fied, fixed supports and microphones at each location could be used. These
could connect into the 1566 Multichannel Amplifier, which can scan the
outputs of the microphones manually or automatically. The octave-band
analysis could be done with a 1921 Real-Time Analyzer, with the measured
equipment noise levels and background noise levels plotted on a 1522 Dc
Recorder.
The noise-certification tests for aircraft, as prescribed in Part 36 of the
Regulations of the Federal Aviation Administration ( FAA, 1969) require
extensive instrumentation and calculations. The noise at a number of points
must be recorded on magnetic tape, during specified landing and takeoff
procedures. The recording must also include acalibration signal. The response
characteristics of the recorder must be exceptionally good, since they must
meet the requirements of IEC-R179 (Precision Sound Level Meters).
The recorded noise is then analyzed every half second, into 1/3-octave
bands, by a 1921 Real Time Analyzer. These band levels are processed
successively by an associated small computer, by asimple recursive arithmetic
process, to produce the required dynamic response. Each 0.5 second set is
then stored in the computer. The levels are corrected for the calibration
results and the effective perceived noise level is calculated. This result is then
corrected for departures from the standard flight path and standard atmos-
pheric conditions.
The standard prepared by the Institute of Electrical and Electronic
Engineers for Airborne Noise Measurements on Rotating Electric Machinery,
IEEE No. 85, Feb. 1965, covers avariety of measurement procedures. These
include sound-level measurements and sound analysis in octave or third-
octave bands, at a point or at a number of points in the vicinity of the
machinery. When the sound-power level is required, it is calculated from the
band levels measured at certain specified points.
For sound-level and octave-band level measurements, the 1933 Precision
Sound-Level Meter and Analyzer would be preferred. Since measurements are
to be made at a number of points, the measurement procedure can be
speeded by the use of a number of microphones and a 1566 Multichannel
Amplifier. It can be automated still further by the use of a 1921 Real Time
Analyzer and a1522 Dc Recorder to plot the results of the analysis.
If 1/3-octave analyses are desired, the convenient combination of a1523
Level Recorder, with the 1523-P3 Stepped 1/3 Octave-band Analyzer plug-in,
is suggested. When it is used with a 1560-P42 Preamplifier and a 1560-P5
Microphone, band levels as low as about 34 dB re 20 .LN/m 2 can be recorded.
If lower-level recordings are needed, the recorder can be driven from a1551-C
Sound Level Meter or a1933 Precision Sound-Level Meter and Analyzer.

154
 NORMALIZED FREQUENCY- ege

Figure 6-27. Type 1952 Universal Filter; at left are low-pass and high-pass filter
characteristics.

system bandwidth can be reduced to decrease the relative intensity of

extraneous signals without affecting the important part of the signal to be
studied, or the effect on the signal of limiting the bandwidth needs to be
studied. Such applications are often handled by adding the 1952 Universal
Filter, shown in Figure 6-27, to asystem. It will perform as acontinuously
adjustable low-pass, high-pass, band-pass or band-reject filter over the fre-
quency range of 4to 60,000 Hz ( Kundert, 1968a).
Controlled bands of noise can be generated by the use of this filter with a
random-noise source. Such bands are often useful for psychoacoustic tests,
room-acoustics tests, and vibration tests.

6.12 TONE- BURST GENERATOR.

The combination of an oscillator and a tone-burst generator ( Skilling,
1968) is asource of electrical tone-burst waveforms that can be converted to
acoustical tone bursts by means of an earphone or a power amplifier and
loudspeaker. Such an acoustical signal is useful in room-acoustic measure-
ments, in psychoacoustic experiments, in studies on transducers and acousti-

Figure 6-28. Tone-burst

pattern from 1396-B Tone
Burst Generator. Input, 3
kHz; 16 cycles on; 16 cy-
cles off.

133
cal material properties, and in amplifier tests. It is particularly helpful in
locating stray reflections in anechoic chambers and in tracing sound-transmis-
sion paths.
The 1396-B Tone- Burst Generator, shown in Figure 6-28, produces an
abrupt on-off transition and is particularly useful for measurements of physi-
cal systems. The on and off times can be set in terms of the number of cycles
of the input wave or aseparate wave or in terms of time.
The GIS Model 929E Electronic Switch and Model 471 Interval Timer in
combination constitute a flexible means of controlling an audio signal,
including an abrupt tone burst or one with rise and decay times of 1to 500
msec. The signal, with a gradual rise and decay time, is more suitable for
psychoacoustic testing than one that is abrupt.

6.13 AUTOMATIC- LEVEL REGULATOR.

In the testing of response or transmission-loss characteristics in room and
building acoustics, it is often helpful to control or regulate the level of the
test signal to keep it constant. The 1569 Automatic Level Regulator ( Kundert
and Woodward, 1968), shown in Figure 6-29, is an important element in such
asystem. It can, for example, reduce alevel variation of 25 dB to avariation
of only 1 dB, when it is supplied with the proper reference signal and it is
inserted as part of the signal-supply loop. Such large initial variations can
occur because of the marked frequency irregularity of rooms and many
loudspeakers.

6.14 OSCILLOSCOPE.
A cathode-ray oscilloscope is auseful means of observing the waveform of
asound or vibration signal from asound-level meter or avibration meter. It
can be used to measure the peak amplitude of a wave and, after some
experience, the observer can, by adjusting the sweep frequency, tell some-
thing about frequency components by looking at the wave. In addition, the
oscilloscope makes possible the study of the instantaneous values of a
vibration motion. In contrast to the vibration analyzer and other wave
analyzers that present information in terms of frequency, the oscilloscope
presents information as a function of time. This representation is often of
great assistance in the solution of vibration problems. Because the oscillo-
scope presents information instantly and continuously, and because its fre-
quency response is not a limiting factor, it is useful in the study of any
vibration waveform.
For sound and vibration measurements an oscilloscope with slow sweep
rates, long-persistence screen, and dc amplifier is recommended. Many oscillo-

Figure 6-29. Type 1569 Automatic Level Regulator.

134
 e
i 1560-9531
LAL

1568
I

1560-9669

\
1521-8
(t -

1560- P42
1562
156C - i-53

Figure 7-6. Narrow-band sound or vibration analysis.

CAL

1562 1933

Figure 7-7. Impact-noise analysis in octave bands.

Cat I

11111 I tamp aIam= amm

0•0
00000

LEEe 000100010 ®

1562 1933 170 1923

Figure 7-8. Narrow-band high-speed vibration analysis, including correlation, transfer

function, and summation analysis as well as spectrum analysis.

7.2.4 Machinery Preventive Maintenance. Only one aspect of machinery

maintenance is considered here, namely, the relation of the vibration level of
a machine to its condition. That is, vibration measurements can guide in
predicting incipient failure of a machine, in deciding when cleaning, parts
replacements, and other maintenance procedures are necessary, and in deter-
mining the relation between vibration and the performance of the machine.
The best general measurement for this purpose is velocity, and it is
usually measured at the bearing housings. The 1553 Vibration Meter provides
for velocity measurements, and it can be used for quick surveys. But since
analysis of the vibration signal is almost always desirable, the 1933 Precision
Sound- Level Meter and Analyzer with avibration pickup and acontrol box is
achoice that reduces the total number of instruments required.

158
 Figure 6-31. Type 1562 Sound Level
Calibrator installed on asound level meter
(left). A Type 1557 Vibration Calibrator is
shown at right.

6.16 CALIBRATORS.
Much is to be gained from the use of an accurately calibrated acoustical- or
vibration-measurement system. When an accurate calibration is made, the
consistency of comparison measurements can be improved, acloser approach
to an allowed performance specification is possible, and careful attention to
measurement techniques will be repaid by more accurate measurements.

6.16.1 Sound- Level Calibrator. When an over-all acoustical check of a

system at several frequencies is desired, the sound-level calibrator can be used
(Gross, 1967a). It comprises a small, stabilized, and rugged loudspeaker
mounted in an enclosure that fits over the microphone of the sound-level
meter. The chamber is so designed that the acoustic coupling between
loudspeaker and microphone is fixed and can readily be repeated. The level is
high enough so that readings are unaffected by normal background noises.
The calibrator in Figure 6-31 includes its own battery-operated oscillator
to drive the transducer, which supplies aknown level at 125,250,500,1000,
and 2000 Hz to the microphone. This device permits aquick check of the
performance of an acoustic measurement system over the most important
frequency range for acoustical measurements. It is also invaluable for calibrat-
ing a measurement tape recorder and for supplying calibrating signals for
recording.

6.16.2 Audiometer Calibrators. Audiometry is one of the fields of meas-

urement where it is particularly important to have equipment in proper
operating condition and well calibrated. Comprehensive calibrations should
be done at least every 6 months, by the manufacturer or at some other
laboratory qualified to certify such calibrations. In addition, adaily monitor-
ing check on the audiometer and earphones can be done quickly and easily
with the 1562-Z Audiometer Calibration Set. If the audiometer is not used
daily, the monitoring check can be done before use of the audiometer.
The basic instrument in the 1562-Z is the 1565-B sound-level meter,
which, of course, is essential to ahearing-conservation program for checking

136
 One basic technique in this form of preventive maintenance is to compile
a history of the analyzed vibration levels at each bearing housing. When the
levels change noticeably, the situation is reviewed to see if it is reasonable or
if it is likely that deterioration of some structure has occurred. It can then be
used as one guide in deciding when and how machinery is to be overhauled.
Octave-band analyses have been found helpful here (Glew and Watson, 1971),
but 1/3-octave analyses are also used ( Bowen and Graham, 1967). Here, the
history is conveniently compiled in the form of aseries of recorded vibration
levels taken on a 1523 Level Recorder with a 1523-P3 Stepped 1/3-Octave-
Band Analyzer plug in. The remarkable range of that analyzer, 1 Hz to
80,000 Hz, means that the full range of the vibration meter, including the
important low-frequency range, is nicely covered.

1560-P53

Figure 7-9. Tenth octave-band vibration analysis with stroboscopic observation.

When many machines are being monitored, so that many analyses need to
be made and recorded, the combination of a1921 Real Time Analyzer and a
1522 Dc Recorder provides arapid way of producing the data to be studied
for maintenance procedures.
Fine detail of analysis may be helpful in investigating certain faults of
rotating machinery, and the narrower-band analyzers, including the T/D 1923
FFT Analyzers, are then appropriate instruments to use. These FFT Analyz-
ers also provide correlation measurements and waveform averaging (summa-
tion analysis), which are useful investigational tools in machinery mainte-
nance (see paragraphs 5.4.1 and 5.4.5).

I56C - C53 1560-P.3

1557 '933

Figure 7-10. Velocity measurement and analysis sytern.

159
=igure 6-33. Type 1531 Strobotac® electronic stroboscope (left). Type 1531-P2 Flash
)e/ay (attached to stroboscope), and Type 1536-A Photoelectric Pickoff.

nachine can be analyzed with the stroboscope under normal operating

•onditions.
This type of instrument can be used to measure the speeds at which
ibrations occur in most rotating or reciprocating machines. Displacements in
ibrating parts can often be measured accurately with the aid of a micro-
cope, if afine reference line is scribed on the part. This technique has been
tsed to confirm the calibration of vibration calibrators, and automotive
ngineers have used it to measure crankshaft whip and vibration.

6.17.1 Strobotace Electronic Stroboscope. The Strobotac® electronic

troboscope, shown in Figure 6-33, is asmall, portable stroboscope calibrated
o read speed directly in revolutions per minute (Holtje, 1966). The light
ource is astrobotron tube, mounted in aparabolic reflector. The frequency
4* an internal electronic pulse generator determines the flashing speed, which
an be adjusted by means of a direct-reading dial. Normal flashing ranF is
rom 110 to 25,000 per minute. Another model of the Strobotac is
vailable for flashing rates up to 150,000 per minute, and that model can be
perated from arechargeable battery. Speeds above and below this range can
emeasured by use of flashing rates that are simple multiples or submultiples
f the speed to be measured. As the flashing rate of the Strobotac is
ecreased below 600 per minute, the flicker becomes pronounced due to the
lability of the human eye to retain successive images long enough to give the
lusion of continuous motion.
Of especial use in vibration measurements is the provision for connecting
n external synchronizing signal to the Strobotac. Thus the light flashes can
etriggered directly by a 1536-A Photoelectric Pickoff, which uses aphoto-
ell to synchronize the stroboscope with repetitive mechanical motion. It
tquires no attachment to the device being observed, and thus can be used
ffectively with low-torque devices. The output of the photoelectric pickoff
tquires amplification to trigger the stroboscope; this is provided by the Type
531-P2 Flash Delay Unit, which also permits observation of the vibration at
ny point in its cycle.
The stroboscope can also be flashed by the output from one of the
ibration-pickup systems described in paragraph 6.5.1. For instance, avibra-

138
tion pickup can be used with asound-level meter or vibration meter to send
triggering impulses to the stroboscope. Filtering is necessary between the
measuring instrument and the stroboscope. An octave-band or anarrow-band
analyzer can be used for such filtering.

6.17.2 Strobolume The 1540 Strobolume® electronic stroboscope pro-

vides a very bright light over alarge area ( Miller, 1969). The maximum beam
width is 40° x 65°. It can be controlled by an accurate oscillator for the range
of 110 to 25,000 flashes per minute, by an oscillator/delay unit, or by a
control unit to respond to external signals. The light output and versatility of
the 1540 make it well suited for TV applications, such as video recordings of
rapidly-moving parts in mechanical devices.

6.17.3 Motion Analysis Set. The Type 1539-A Motion Analysis and
Photography set is arranged for visual analysis of a repetitive motion or
inspection of a process where the independent flashing-rate setting of the
Strobotac is not required and for high-speed photography with conventional
cameras. The major application areas for the motion analysis are in machinery
and metal working, including packaging, printing, textile, earthmoving ma-
chinery, metal products, shipbuilding, automotive manufacturing, ordnance,
chemical processing and aerospace.

6.17.4 Stroboscopic Applications. Stroboscopic techniques are widely

used for visual observation of vibration. The high-speed performance of fans,
propellers, and other rotating devices can be studied by means of the
slow-motion effect of the stroboscope, and sources of vibration and noise due
to misadjustments, misalignment, and wear can be readily detected. The
vibratory modes of turbine blades are checked as they are driven electromag-
netically, and the mode shapes are observed with the aid of an optical
magnifier under stroboscopic illumination. Similarly, the flapping of the
blades of a model helicopter rotor has been observed in slow motion by
stroboscopic illumination.

Figure 6-34. Combination of instrumentation forces to solve a vibration problem. Shown

are the Types 1538 and 1539 Stroboscopes, paired with a Type 1553 Vibration Meter,
to study the motion of parts on an electronic assembly being exercised on ashake table.

139
 The stroboscope can also be used to observe the motion of apparatus being
tested on ashaker, as in Figure 6-34. If the flashing rate is just slightly offset
from the frequency of the shaker, a slow-motion replica of the high-speed
vibration will result, so that the displacement can readily be observed. The
form of the motion can be seen, and one can often tell what section needs to
be strengthened and how damping material and damping devices can best be
applied.
When a rotating or reciprocating machine is brought up to speed or is a
variable-speed device, there may be resonant vibration modes of various parts
at certain speeds, known as critical speeds. If these parts are visible and can be
illuminated by a stroboscope, it is often possible to use the slow-motion
feature to check on the actual behavior of the part at resonance. One can see
if it is afundamental resonance or amultiple resonance with various sections
going in phase and others in phase opposition. This type of observation can
be of great assistance in the determination of the proper treatment to reduce
the resonant vibration.
TV cameras and receivers and video recording techniques offer agreater
degree of flexibility in the use of stroboscopic techniques, particularly for
remote observation.
Photographic recording of the stroboscopically illuminated motion makes
possible accurate measurements of the motion. The accurately timed, very-
short-duration light flashes provide the time scale and the almost instantane-
ous sampling of the motion.
For further details on the stroboscope and its uses consult F. T. Van Veen,
Handbook of Stroboscopy, General Radio Company, Concord, Mass., 1966,
and Charles E. Miller, Handbook of High-Speed Photography, General Radio
Company, Concord, Mass., 1967.

6.18 SYSTEMS, COMPUTERS, AND PERIPHERAL EQUIPMENT.

When anumber of instruments and devices are arranged to work together
to solve ameasurement problem, the combination is often called ameasure-
ment system. Engineering such asystem requires the planning of the over-all
design to yield the required output information from the selected inputs, the
selection of the individual parts of the system, the design and development of
the interconnection means (often called " interfacing"), and the testing and
debugging of the completed system.
Most of the items described in this chapter can be integrated into such
measurement systems. The block diagrams of Figure 6-35 and 6-36 illustrate
this statement. The source of the signal is often atransducer or abank of
transducers, or it may be arecorded signal. If there is more than one input
source, they will need to be selected or scanned. This signal is then measured
or analyzed, and the result is transferred to an output device or to acomputer
for further processing before it is sent to an output device.
Sometimes the interconnection of the various parts is simple, because they
have been designed to work together. But often careful attention to many
details is necessary in order to make the system as awhole work well, and the
development of electronic circuits to modify the output of one device to
satisfy the needs at the input of the cascaded device may be necessary. For
example, when an output is in analog form and it is to drive adevice that

140
 INPUT

MEASURE

REAL-TIME ANALYZER

Figure 6-35. System for acoustic power measurements.

INPUT SIGNAL CONDITIONING

n r
--

MICROPHONE PREA

I
I AN -
ANA
DIG TAC
I PICwuP PREAMP SCANNER ALIASING »-AIN TAL
I I FILTER CONTE PTV,

I RECORDER I I

_J L_ _J

CONTROL AND PROCESSING

r i HIGH SPEED
NAG TAPE MTERFACE
TERFACE READER,
STORAGE PUNCH

cow PUTE MTERFACE

ERFACE

IITERFACE
TERFAC

TELETYPE

COUPLER
INTERFAC

TAPE TYPEOUT

17]

OUTPUT

DIGITAL/
•NALOG
CONVERTER

DIGITAL ,
ANALOG
CONVERTER

RECORDER

CARD
PUNCH

L__
Figure 6-36. Computer controlled system for measurement
and analysis by digital techniques.

141
requires adigital form of input, it is obvious that some significant interfacing
electronics is necessary.
To illustrate the nature of the problems here is alist of some of the input
and output characteristics that need to be considered in the connection of
two devices.
Analog-to-analog connection
Impedance
Signal levels
Frequency response
Interaction
Grounding
Stray pickup
Dc levels
Power requirements
Connectors
Digital-to-digital connection
Logic levels
Polarity
Code
Number of bits
Parallel or serial
Clock rate
Pulse durations
Logic load
Interrupt logic
Noise immunity
Connectors
Buffer storage
Analog-to-digital connection
Those above plus
Analog-to-digital ( AID) converter characteristics
Digital-to-analog connection
Those above plus
Digital-to-analog ( D/A) converter characteristics
In addition to the equipment already described in this chapter that may be
used in systems, as already implied, adigital computer is often included in
larger systems because of the versatility it gives to the system ( Fichtenbaum,
1969; McAleer, 1971; Peterson and Fichtenbaum, 1969). It will be described
briefly. Other devices that are mainly used in systems and are often called
"peripheral equipment" will also be described briefly.
When a computer is part of a system, the program for the computer
("software") must fit the particular devices used in agiven system, and the
interconnecting hardware that permits the computer to control the devices
and to receive data from them or send data to them must also be included
(DEC, 1968, 1970).
6.18.1 The Computer in Acoustical Systems: We shall discuss the role of a
computer in ameasurement system in two parts, as aprogrammed calculator
and as asystem controller.
Computer as Calculator. Most engineers regard acomputer as ahigh-speed
programmable calculator. In this role it can play auseful part in acoustics,
because many of the results that we want are not given directly by the values

142
 Another standard, ASTM E336-67T, Tentative Recommended Practice
for Measurement of Airborne Sound Insulation in Buildings, specifies how
similar measurements should be made in the field. Since the conditions of the
measurement cannot be as well controlled as in the laboratory situation, the
results may be quite different. The measurement techniques are similar,
although not necessarily identical, but the instrumentation required is gener-
ally the same as outlined above.

7.4.3 Reverberation Time or Decay Rate. The rate at which sound decays
in an auditorium has been found to be a useful quantity in rating an
auditorium. If sound decays too rapidly, the room sounds " dead," and an
organ, for example, played in adead hall loses much of its appeal. A very live
room, that is, one in which the sound decays slowly, may be useless as a
lecture hall, because the individual syllables of speech are obscured by the
persisting sounds of previous syllables.
The more common name for the characteristic that rates sound decay is
"reverberation time," which is the time taken for the sound to decay 60 dB
after the source has stopped. This time is usually measured by exciting the
room with a band of noise, which requires a noise generator and apower
amplifier driving one or more loudspeakers ( see Figure 7-15). In an auditor-
ium or theater, the power amplifier and loudspeaker system for sound
reinforcement can usually be used as part of this measurement system.
This sound is then picked up by amicrophone feeding another 1/3-octave
filter, which drives agraphic level recorder with a40 or 50 dB potentiometer.
The pickup system is adjusted so that the level on the recorder is near full
scale. The sound is then suddenly turned off and the graphic level recorder
plots the decay. A straight line is drawn on the chart to fit the average slope
of the curve. From this slope, one can calculate the rate of decay and the
reverberation time. A number of refinements are often introduced into this
process, and these are discussed in the book by Davis and in papers in the
Journal of the Acoustical Society of America. These measurements are made
over a wide frequency range and in a number of representative places in the
hall.
A tape recorder can be of great help in this measurement ( Schultz, 1963).
(See Figure 7-16.)

POWER
AMPLIFIER

1362
1564

AUDITORIUM 1560- 952

CAl

1562
1523, - P I

Figure 7-15. Decay-rate and reverberation measurements for 1523 system.

164
a. Record
test signal.

b. Playback
signal and
record hall
response.

C. Playback
hall response
for analysis.
— 11

1935 1564 1521 -

Figure 7-16. Use of tape recorders to perform an acoustic analysis of an auditorium.

When it is necessary to simplify the testing procedure at ahall (if, for

example, an audience is present), an alternative procedure may give useful
information with relatively short testing time. The hall is excited by means of
a broad-band of " pink" noise (noise whose spectrum level decreases with
increasing frequency to yield constant energy per octave of bandwidth). The
sound in the hall is recorded on atape recorder with the noise on and during
the decay period when the noise is suddenly switched off. Later, in the
laboratory, this recorded signal is played back through a1/3-octave analyzer
and a graphic level recorder. This recording is repeated for the full range of
the desired settings of frequency of the analyzer. In this way, a complete
picture of the decay rate as a function of frequency can be obtained with
only one " exposure" at the hall. Some of the detail, particularly at lower
levels, may be obscured, however, by the background noise in the hall.

7.4.4 Response Testing. The equalization of asound system, to obtain

good uniformity of response with frequency, can improve the performance of
the system. The suggested procedure for measuring the response is as follows:
The sound system is driven by broadband pink noise. The sound produced is
then analyzed by arecording 1/3-octave analyzer ( Figure 7-17). This analysis
is carried out for anumber of representative places in the area to be covered
by the sound system.
The adjustments in the frequency response are then made on the basis of
astudy of the response curves. The fine details of the curves are ignored, and
the broader trends are used to decide how best to adjust the frequency-re-

165
 The program for the Computer (Software). The program for an instrumen-
tation system may include such widely varying functions as control of
measuring instruments, calculation of various predictors of human reactions,
comparisons with stored criteria, calculation of statistical measures of the
measured results, and the printout of a summary. In order to prepare this
program, the programmer must be familiar with the measurement task, the
characteristics of the equipment comprising the system, and the characteris-
tics of the computer. The function of the programmer/engineer is then one of
designing aprogram rather than designing instrumentation hardware.

Large-vs-Small Computers. For most instrumentation systems, a small

computer is adequate for the job. A small computer should then almost
always be used. But if alarge computer is already available in an organization,
it is often proposed that it be tied into the instrumentation system to do the
job. This procedure is usually impractical for one or more of the following
reasons:

1. In an instrumentation system, the timing of responses to events in the

system may be crucial. If the large computer is run on abatch-process-
ing basis, the use in instrumentation and control becomes impractical.
Either the instrumentation system must have top priority and complete
availability of the necessary computer facilities when the measurement
is in process or afast time-sharing system should be used.
2. Connection of an instrumentation system to alarge computer is diffi-
cult and expensive, particularly if the computer is remote, as in a
time-sharing system. The task of programming may be difficult, because
the instrumentation manufacturer may not be familiar with the particu-
lar large computer and the measurement task at hand, and the user's
engineers may not know how to program the system. The expense of
the special interconnection, the cabling, and the more involved pro-
gramming may exceed the cost of the small computer.
3. If the instrumentation system is tied to alarge computer, it fails when
the computer fails. With asmall computer in the system, the user can
do his own maintenance or even keep aspare system, if it is essential
that the down time be minimized.
4. With asmall computer as part of the system, it is under the complete
control of the user and not subject to limitation or regulation by the
department controlling the large computer.

When Should a Computer Be Used? Factors that affect the decision to

include acomputer in ameasurement system:
1. Quantity of data to be processed.
2. Rate at which data must be processed.
3. Complexity of the measurement function, for example, involved se-
quencing of instruments.
4. Calculations needed on the directly measured data to obtain required
output results, e.g., corrections, calculation of loudness, mean and
standard deviations of sets of data.
5. Use of system to perform different measurement tasks. When the basic
instruments can be programmed through the computer to perform the
required measurement tasks, the cost of the , computer with several
programs may be less than the alternative instrument systems, or at

145
 111111111111111111111111111111.

Figure 6-37. Typical computer-

controlled system built around
the 1921 Real- Time Analyzer.

least the possibility of using the computer for several tasks may contrib-
ute to justifying its cost.
6. The versatility of the computer system. It may permit one to avoid the
expensive development of specialized instrumentation for aparticular
job.
An Acoustical Measurement System. The particular system shown in
Figure 6-37 uses asmall computer with the 1921 system. It is programmed to
take a series of measurement samples of an input noise signal. The third-
octave levels for each noise sample are stored in the computer memory. The
program permits the calculation of the octave-band levels, the loudness level
by the method given in ANSI S3.4, and the speech-interference level. It can
also select the maximum loudness of any of the loudnesses calculated for the
noise samples.
The computer program for the test system is kept on apunched paper tape
and is loaded through atape reader on the teletypewriter into the computer
memory. In normal operation, the program is stored in the memory indefi-
nitely and is ready to run whenever the power switch is turned on. All
transfer of data and instruction between the operator and the computer
occurs through the teletypewriter.
The computer is connected to the 1921 system by cables and circuits
called " interface circuits." These circuits interpret the computer commands
to control the instruments, they perform any required shifting of voltage
levels, they accomplish intermediate storage of data, and they transfer the
measured data from the 1921 system to the computer.

146
frequency noise is much greater than that of isolating high-frequency noise,
and the knowledge of the level of low-frequency sounds and vibrations may
be an essential element in cost studies.
A related problem is that of locating astudio within an existing building.
Here, acareful survey of possible locations may lead to asignificant saving in
construction costs.

REFERENCES
L. L. Beranek ( 1949), Acoustic Measurements, John Wiley 8c Sons, Inc., New York.
K. A. Bowen and J. S. Graham ( 1967), " Noise Analysis: A Maintenance Indicator,"
Mechanical Engineering, Vol. 89, # 10, 1967, pp. 31-33.
D. Davis ( 1965), Acoustical Tests and Measurements, H. W. Sams, Indianapolis.
FAA ( 1969), Federal Aviation Regulations, Part 36-Noise Standards: Aircraft Type
Certification, Federal Aviation Administration, Department of Transportation.
Federal Coal Mine Health and Safety Act ( 1969), Public Law 91-173, 91st Congress,
S.2917, December 30, 1969.
C. A. W. Glew and D. C. Watson ( 1971), "The Octave Band Vibration Analyzer as a
Machinery Defect Indicator," Am Soc of Mech Engineers, Paper 71- DE-47, April
1971.
E. Meyer and H. Kuttruff ( 1962), " Progress in Architectural Acoustics," in Richardson
and Meyer, 1962, pp. 222-277.
OSHA ( 1970), " Part 1910- Occupational Safety and Health Standards," under authority
of Williams-Steiger Occupational Safety and Health Act of 1970 ( P.L. 91-596) ( 84
Stat. 1593, 1600; 29 U.S.C. 655, 657) Federal Register Vol. 36, # 105, May 29,
1971, Washington, D.C.
E. G. Richardson and E. Meyer, eds. ( 1962), Technical Aspects of Sound Vol ///,
Elsevier Publishing Company, Amsterdam.
T. J. Schultz, " Problems in the Measurement of Reverberation Time," Journal of the
Audio Engineering Society, Vol. 11, #4, October 1963, pp. 307-317.
T. J. Schultz and B. G. Watters ( 1964), " Propagation of Sound Across Audience
Seating", JAcoustical Society Am, Vol. 36, #5, May 1964, pp. 885-896.
WHPCA ( 1969), Occupational noise-exposure regulations under Section 50-204.10 of
the Walsh-Healey Public Contracts Act - Safety and Health Standards, Federal
Register, Vol. 34, #96, May 20, 1969, ammended July 1, 1969.

167
 Paper- Tape High-Speed Reader/Punch. When computer programs are
stored on paper tape, they are usually so long that they take 15 minutes or
more to read in on the tape reader attachment for the teletypewriter. If a
number of programs are to be run on the computer, it is usually worthwhile
to use ahigh-speed tape reader to read in tapes.
If data are to be stored on paper tape, or if the computer setup is to be
used in the development of programs, ahigh-speed punch may be worthwhile.
Digital Printers. Often aprinted record of the measured results is desired.
Various forms of printers are available. They range from the teletype, through
high-speed column printers and line printers.
Magnetic Digital Storage. Programs or data in digital form can also be
stored on magnetic tape. The transfer of digital signals between the magnetic
tape and the computer is much faster than for paper tape. The ordinary audio
tape recorder cannot be used for this purpose, however, since the require-
ments are quite different, particularly with respect to the drive mechanism,
the heads, the electronic systems and the tape. In order to ensure accuracy in
the transfer of the digital signals, the tape must be of exceptionally high
quality. Some redundancy is added to the digital signals when they are stored,
and this redundancy is used to check that the transfers are correct.
The magnetic-storage medium can also be in the form of ametal disk that
has athin magnetic-oxide coating on both sides. When it is in this form, any
stored program can be located more quickly than for storage on tape. But for
alarge storage capacity, disk storage is more expensive than tape storage.
Do not Fold, Staple, or Mutilate. Almost everyone closely associates the
:omputer with the punched-card form of storage of digital information. The
Dun ched card has the important advantage that readily readable information

:an be printed on it, and for that reason it appears almost indispensable for
;ome applications. The devices for punching and reading cards are more
expensive than those for paper tape, however, and for most of the acoustical
lata-processing applications the cards are not needed.

REFERENCES

itandards
INSI S1.4-1971 Sound- Level Meters
‘NSI S1.11-1966 Octave, Half-Octave, and Third-Octave Band Filter Sets
EC/123 ( 1961) General Purpose Sound-Level Meters
EC/179 ( 1965) Precision Sound Level Meters
EC/225 ( 1966) Octave, Half-Octave, and Third-Octave Band Filters

)ther
1. R. Arnold, H. C. Hill, and A. V. Nichols ( 1969), Modern Data Processing, John Wiley
& Sons, Inc., New York.
A. W. Basch ( 1964), " New Talents for the Graphic Level Recorder," General Radio
Experimenter, Vol. 38, #9, Sept, pp 15-24.
A. W. Basch ( 1969), " A Programmable High-Speed Dc Recorder," General Radio
Experimenter, Vol. 43, #5/6, May/June, pp. 16-21.
I. B. Bauer ( 1957), " Microphones for Sound-Level Meters,"J Acoust Soc Am, Vol. 29,
#12, Dec, pp. 1333 f.
,. L. Beranek ( 1949), Acoustic Measurements, John Wiley & Sons, Inc., New York.
I. A. Bonk ( 1967), " The New General Radio Microphone," General Radio Experi-
menter, Vol. 41, #5/6, May/June, pp. 3-8.
;. B. Booth ( 1963), " Vibration Generation," Chapter 9in Crandall, 1963.
C. A. Bowen and T. S. Graham ( 1967), " Noise Analysis, A Maintenance Indicator,"
Meth Engineering Vol. 89, # 10, pp. 31-33.

148
E. V. Carlson ( 1952), " A Ceramic Vibration Pickup," Trans IRE Prof Group on Audio,
PSA-10, Nov-Dec, pp. 2-8.
L. J. Chamberlain ( 1971), " A Simple Discussion of Time-Series Analysis," Sound and
Vibration, Vol. 5, #4, April, pp. 18-25.
S. H. Crandall, ed. ( 1963), Random Vibration, Vol. 2, The MIT Press, Cambridge, Mass.
D. Davis ( 1965), Acoustical Tests and Measurements, H. W. Sams, Indianapolis.
A. Dranetz and A. W. Orlacchio ( 1961), " Pickup Characteristics," Chapter 16, Part Iin
Harris and Crede, 1961.
DEC ( 1968), Introduction to Programming 1969, Digital Equipment Corporation, May-
nard, Mass.
DEC ( 1970), PDP8/E Small Computer Handbook 1971, Digital Equipment Corporation,
Maynard, Mass.
S. V. Djuric ( 1972), " Electret Condenser Microphones For Measurement," J Acoust
Soc Am, Vol 51, # 1, Part 1, Jan, p 129.
FAA ( 1969), Federal Aviation Regulations, Part 36 - Noise Standards: Aircraft Type
Certification, Federal Aviation Administration, Department of Transportation.
J. J. Faran, Jr. ( 1968), " Random Noise Generators," General Radio Experimenter, Vol.
42, # 1, Jan, pp. 3-14.
Federal Coal Mine Health and Safety Act ( 1969), Public Law 91-173, 91st Congress,
S.2917, December 30, 1969.
M. L. Fichtenbaum ( 1969), " Computers in Instrument Systems," General Radio Auto-
matic-Impedance-Measurement Seminars (unpublished).
C. A. W. Glew and D. C. Watson ( 1971), " The Octave Band Vibration Analyzer as a
Machinery Defect Indicator," Am. Soc. of Mech. Engineers, Paper 71-DE-47, April.
R. L. Grason ( 1970), " Audiometric Measurement: 150 Years of Applied Research,"
General Radio Experimenter, Vol. 44, #10-12, Oct/Dec, pp. 3-7.
E. E. Gross, Jr. ( 1960), " Little Dithers," General Radio Experimenter, Vol. 34, # 11/12,
Nov/Dec, pp. 3-5.
E. E. Gross, Jr. ( 1966), " A Standard Earphone Coupler for Field Calibration of
Audiometers," General Radio Experimenter, Vol. 40, # 10, Oct, pp. 15-20.
E. E. Gross, Jr. ( 1967a), " New Five- Frequency Sound-Level Calibrator," General Radio
Experimenter, Vol. 41, # 5/6, May/June, pp. 9-13.
E. E. Gross, Jr. ( 1967b), " Field Calibration of Audiometers," General Radio Noise
Measurement, Vol 1, # 1, March, pp. 3-5.
E. E. Gross, Jr. "( 1968), " New Audiometer Calibration Set," General Radio Experi-
menter, Vol 42, #4, April, pp. 20-22.
C. M. Harris and C. E. Crede ( 1961), Shock and Vibration Handbook, McGraw-Hill, New
York, Vol. I, Chapters 12 through 17.
M. S. Hawley ( 1955), " The Condenser Microphone as an Acoustic Standard," Bell
Laboratories Record, Vol. 33, # 1, Jan, pp. 6-10.
M. C. Holtje ( 1966), " Flash- A New Strobotac Electronic Stroboscope," General Radio
Experimenter, Vol. 40, #4, April, pp. 3-9.
G. W. Kamperman ( 1958), " A Portable Magnetic Tape Recorder for Acoustical Measure-
ments," Noise Control, Vol. 4, # 1, Jan, pp. 23-27.
W. R. Kundert ( 1963), " New Performance, New Convenience with the New Sound and
Vibration Analyzer," General Radio Experimenter, Vol. 37, #9/10, Sept/Oct, pp. 1-9.
W. R. Kundert ( 1966), " A One-Percent-Bandwidth Wave Analyzer," General Radio
Experimenter, Vol. 40, #9, Sept, pp. 10-18.
W. R. Kundert ( 1968a), " A Universal Filter for Low- Frequency Work," General Radio
Experimenter, Vol. 42, #4, April, pp. 14-19.
W. R. Kundert ( 1968b), " A Calibrated Spectrum Synthesizer," General Radio Experi-
menter, Vol. 42, # 10, Oct, pp. 5-10.
W. R. Kundert and C. A. Woodward ( 1968), " A Versatile Level Regulator for Swept- Fre-
quency Sound and Vibration Testing." General Radio Experimenter, Vol. 42, #4,
April, pp. 10-13.
W. R. Kundert and E. R. Marteney ( 1971), " A Modern Portable Sound Analysis
System," 7th ICA, Budapest.
W. R. Kundert, J. A. Lapointe, and G. R. Partridge ( 1969), " New Generation Acoustical
Analyzers," General Radio Experimenter, Vol. 43, #5/6. May/June, pp. 3-11.
J. A. Lapointe ( 1969), " Some Notes on Digital Detection," General Radio Experi-
menter, Vol. 43, #5/6, May/June, p 12 f.
H. T. McAleer ( 1971), " A Look at Automatic Testing," IEEE Spectrum, Vol 8, #5, May,
pp 63-78.

149
 E. R. Marteney ( 1970), " A Big Little- Brother Preamplifier," General Radio Experi-
menter, Vol. 44, # 10-12, Oct/Dec, p. 8.
W. P. Mason ( 1961), " Properties of Piezoelectric and Piezoresistive Materials," Chapter
16, Part 2in Harris and Crede.
E. Meyer and H. Kuttruff ( 1962), " Progress in Architectural Acoustics," in Richardson
and Meyer, pp. 222-277.
C. D. Mee ( 1964), The Physics of Magnetic Recording, North Holland, Amsterdam.
C. E. Miller ( 1969), " Detailed Viewing in Ambient Brightness," General Radio Experi-
menter, Vol. 43, #9/10, September/October 1969, pp. 3-6.
C. E. Miller ( 1970), " Using Stroboscopy," Machine Design, Vol. 42, 30 April 1970, 14
May, 1970. ( two parts).
J. Ormondroyd, R. B. Allnutt, F. Mintz, and R. D. Specht ( 1950), " Motion Measure-
ments," Handbook of Experimental Stress Analysis, M. Hetenyi, ed., John Wiley &
Sons, Inc., New York, Chapter 8, pp. 301-389.
OSHA ( 1970), Williams-Steiger Occupational Safety and Health Act of 1970 ( 84 Stat
1593, 1600; 29 U.S.C. 655, 657) Occupational Safety and Health Administration,
Part 1910 - Occupational Safety and Health Standards- Federa/ Register, May 29,
1971, Washington, D.C., Vol. 36, # 105, Part II.
G. R. Partridge ( 1971), " A Noise Exposure Meter,"J Acoust Soc Am, Vol. 41, # 1 ( part
1), Jan, p. 129.
A. P. G. Peterson ( 1967), "The Tape Recorder in Acoustical Measurements," Sound and
Vibration, Vol. 1, # 10, Oct, pp. 14-20.
A. P. G. Peterson and M. L. Fichtenbaum ( 1969), "The Computer in Acoustical
Measurement Systems," General Radio Real-Time Sound and Vibration Measure-
ment Seminar ( unpublished).
G. Rasmussen ( 1960), " Pressure Equalization of Condenser Microphones and Perform-
ance at Varying Altitudes," B & K Technical Review, #1, pp. 3-23.
G. Rasmussen ( 1963), " Miniature Pressure Microphones," B & K Technical Review, #1,
pp. 3-10.
E. G. Richardson and E. Meyer, eds. ( 1962), Technical Aspects of Sound Vol. III,
Elsevier Publishing Company, Amsterdam.
G. M. Sessler and J. E. West ( 1966), " Foil-
Electret Microphones," J Acoust Soc Am,
Vol. 40, #6, Dec, pp. 1433-1440.
T. J. Schultz and B. G. Watters ( 1964), " Propagation of Sound Across Audience
Seating," JAcoust Soc Am, Vol 35, #5, May, pp. 885-896.
J. K. Skilling ( 1968), " Redesigned Tone-Burst Generator Has Customer-Suggested Fea-
tures," General Radio Experimenter, Vol. 42, # 10, Oct, pp. 14f.

WHPCA ( 1969), Occupational noise-exposure regulations under Section 50-204.10 of

the Walsh-Healey Public Contracts Act - Safety and Health Standards, Federal
Register, Vol. 34, #96, May 20, 1969, amended July 1, 1969.
J. Tall ( 1958), Techniques of Magnetic Recording, MacMillan, New York.
D. G. Tucker and B. K. Gazey ( 1966), Applied Underwater Acoustics, Pergamon Press,
Oxford, Chapter 5.
F. T. Van Veen, ( 1966), Handbook of Stroboscopy, General Radio Company, Concord,
Mass.
C. A. Woodward ( 1965), " A New, Low-Noise Preamplifier," General Radio Experi-
menter, Vol. 30, #6, June, pp. 1-6.
C. A. Woodward and E. E. Gross, Jr. ( 1968), " A Precision Sound-Level Meter," General
Radio Experimenter Vol. 42, #4, April, pp. 3-6.

150
nished with the earlier Types 1551-A, 1551-B, and 759-B Sound Level
Meters, and in the Type 1555-A Sound-Survey Meter, however, gradually
dissolves if the humidity is too high (above about 84%). The Rochelle-salt
crystal unit in the microphone, however, is protected by acoating, so that it
is relatively unaffected by high humidity. Nevertheless, it is wise to avoid
unnecessary exposure. A Rochelle-salt microphone should not be stored for
long periods in avery dry atmosphere, since it can dry out.

NOTE
Rochelle-salt-crystal microphones are no longer supplied by Gen-
eral Radio Company.

A condenser microphone is not damaged by exposure to high humidity,

but its operation can be seriously affected unless proper precautions are
taken. For proper operation it is essential that very little electrical leakage
occur across the microphone. The exposed insulating surface in the micro-
phone has been specially treated to maintain this low leakage, even under
conditions of high humidity. In spite of the precaution, the leakage may
become excessive under some conditions. Then it may be advisable to keep
the microphone at a temperature higher than the ambient temperature to
reduce the leakage. In climates where the humidity is normally high, it is
recommended that the microphone itself be stored at atemperature above
ambient to avoid condensation.
An electret microphone can stand for long periods the normal variations in
temperature and humidity without significant change in sensitivity. But, if
the humidity is normally high, an electret microphone should be stored in a
small jar containing silica gel. A relative humidity of 90% is the maximum
recommended for the electret.
High or Varying Temperature. Although most noise measurements are
made indoors at average room temperatures, some measurement conditions
expose the microphone to much higher or lower temperatures. When these
conditions are encountered, it is essential to know the temperature limita-
tions of the equipment.
The microphones supplied with the 1565 and 1551-C Sound-Level Meters
will withstand temperatures of —40° to +60 °C without damage. Even at 95 °C,
apermanent sensitivity loss of only about 0.5 dB may occur. In contrast, the
maximum safe operating temperature for Rochelle-salt crystal microphones,
formerly furnished with some sound-level meters, is about 45 °C ( 113 °F). At
55.6 °C ( 132 °F), the Rochelle-salt crystal is permanently changed. It is, there-
fore, not safe to put aRochelle-salt microphone in the trunk or back of acar
that is to be left standing in the sun.
The normal operating temperature range for the condenser microphone
systems is from — 30° to +65°C. They will withstand higher temperatures
without damage, but alimit of 80°C is recommended for the preamplifiers.
The electret condenser microphone should be limited to — 25° to + 55°C
(131° F).
Fortunately, it is usually possible to keep the sound-level meter itself at
more reasonable temperatures. Its behavior at extreme temperatures is limited
by the batteries. Temperatures of even 55°C ( 130°F) will result in much-
shortened battery life. Operation below — 10°F is not ordinarily possible
without special low-temperature batteries.

172
7.2 DEVICES THAT ARE NOISY OR VIBRATE EXCESSIVELY.
7.2.1 Product Noise and Test Codes. Specifications of acceptable noise
limits for products are becoming relatively common. These specifications are
usually given as maximum sound levels or maximum octave-band levels or
sometimes third-octave band levels at certain measuring points. Some specifi-
cations also include the measurement of radiated acoustic power.
Various engineering groups and trade associations have standardized test
codes for measuring the noise from certain devices, for example, transform-
ers, cooling towers, electric motors, fans and blowers, etc. These codes are
often referenced as a part of a specification in order to standardize the
measurement procedure to be used in checking for compliance to amaximum
noise requirement. A representative list of test codes is given in the standards
section of the Appendix.
A- Weighted Sound Levels. A simple example of noise testing is the check
for compliance by a manufacturer for a customer who requires that the
A-weighted sound level at 3feet from any major surface of a motor be less
than say 55 dB. He may also specify that the motor be mounted on a hard
reflecting surface in an essentially anechoic space. Here, the A-weighted

ii
sound level needs to be measured and asound-level meter with amicrophone
will do the job.

CAL

I
I
111M1'

OR 1933

Figure 7-1. System to measure "A"-weighted sound levels.

The 1565 Sound-Level Meter may be adequate for this test, and it is
generally wise to include a 1562 Sound-Level Calibrator as part of the
measurement system. If measurements below 40 dB may be required, the
1551-C Sound Level Meter should be substituted.
In many instances, however, it can be worthwhile to use a1933 Precision
Sound-Level Meter even for this simple measurement, because of the im-
proved accuracy for the A-weighted measurement.
Octave-Band Analysis. Some customers may specify the maximum allow-
able octave-band levels under certain measurement conditions. The 1933
Precision Sound Level Meter and Analyzer is the appropriate instrument to
use, because it provides the octave-band analysis at high accuracy and awide
range of sensitivity levels.
Again, a 1562 Calibrator should be included as a check on the accuracy
of the measurement.
For estimates of probable customer reaction to the noise of aproduct, an
octave-band analysis of the noise is the most widely used measurement. The
band levels are used to calculate loudness level or perceived noise level. If
competitors' products are measured in the same way, either procedure should
permit one to rank the units in order of acceptability with good reliability.

152
 1562 1953

Figure 7-2. Octave-band analysis.

Acoustic Power Output. The use of acoustic-power output for rating

noisy devices is widely recognized as the best approach for certain measure-
ments. Acoustic power is calculated from the results of a number of sound-
pressure level measurements, usually octave or 1/3-octave-band levels. The
procedure requires a controlled environment, usually an anechoic room, a
reflecting floor in an anechoic room, or areverberation room. Under certain
conditions the requirements on the environment may be relaxed. The instru-
mentation used here covers awide range.
For example: It can be a group of microphones scanned with a 1566
Multichannel Amplifier and analyzed with a 1921 Real-Time Analyzer. The
1921 may be adjusted to read the sound power level in each 1/3-octave band
directly. If the calculation of directivity factor or other data reduction is
required, asmall computer may be connected to the 1921. Calculated results
can be printed out on a Teletype. In place of agroup of microphones, it can
be a moving microphone that scans agiven area as its output is analyzed on a
1921 Real Time Analyzer, with the averaged output plotted on a 1522 Dc
Recorder. It can be a1933 Precision Sound Level Meter and Analyzer with its
microphone on a tripod. An analysis at each of a number of microphone
positions is made. Here, the results are recorded and calculated by hand.
Product Noise and Test Codes. In order to be more specific, some
examples of instrumentation for certain test codes will be given.
The American Society of Heating, Refrigerating, and Air-Conditioning
Engineers ( ASHRAE) has prepared a set of standards* for measuring the
noise radiated by heating, refrigerating, and air-conditioning equipment.
These standards are used by the Air-Conditioning and Refrigeration Institute
(ARI) to obtain the basic data in their standards for rating the equipment.
The preferred basic instrument for this use is the 1921 Real-Time
Analyzer. It provides the highly selective third-octave filters and the long
averaging times required by the standards.
A group of microphones scanned with a 1566 Multichannel Amplifier, or
a microphone that moves over a long path as the sound-pressure-squared is
integrated in the 1921, may be appropriate.
The Compressed Air and Gas Institute (CAGI) and the European Com-
mittee of Manufacturers of Compressed Air Equipment ( PNEUROP) have
prepared a " CAGI-PNEUROP Test Code for the Measurement of Sound from

*See Appendix VII for list of standards and codes.

153
 8.2.2 Effects of Room and Nearby Objects.
Effect of Observer and Meter Case. As mentioned in the previous section,
the observer can affect the measured data if he is close to the microphone.
When measurements are made in alive room and not close to asource, the
effect is usually not important. But if measurements are made near asource,
it is advisable for the observer to stand well to the side of the direct path
between the source and the microphone. For precise measurements in avery
dead room, such as an anechoic chamber, the instruments and the observer
should be in another room with only the source, the microphone, the
extension cable, and aminimum of supporting structure in the dead room.
For many measurements, however, it is most convenient to be able to
carry the sound-level meter around. When held in the hand, the sound-level

10

5 •••••••••

A
o A WA
-5 o
w -10
Ui
to EFFECT OF OBSERVER
510
o
5

1020
100 200 1000 2000 4000
FREQUENCY IN HERTZ

10

o
A

EFFECT 0F OBSERVER
— L_L.1à .
b E_P
.... .Wtft)
....„4..

..—,

Inn InnÉl ,Inete, •,•

FREQUENCY IN HERTZ

Figure 8-3. Effect on frequency response as aresult of using the microphone directly on
the swivel post of the instrument (in two positions) without an observer present, and the
extent of the effect with an observer present. A single-frequency, plane, acoustic wave
was used in an anechoic chamber, and the decibel values are the differences between the
response under the conditions noted and the response of the microphone alone.

177
 1560-P5

CAL

1562

1560 -P42 OR 1921 WITH 1522

Figure 7-3. Production-line testing instrumentation.

7.2.2 Production- Line Testing. Ideally, many devices should be tested for
noise output on the production line. Noise measurements on the production
line are often possible, but hardly ever in an ideal manner. That is, precision
acoustical testing usually requires a large, isolated, echo-free space, which
would not ordinarily be considered for inclusion as part of aproduction line.
Nevertheless, useful noise measurements can often be made with relatively
simple procedures, although the accuracy of rating may be significantly
reduced compared with that possible with an ideal measurement.
In this discussion we shall consider briefly several possible solutions to
this problem, ranging from the elaborate to the simple. For some expensive
devices where the noise level is exceptionally important, for example, large
power transformers, the required very large, isolated, echo-free chambers have
been used to test each unit as it is produced. When the device is not so large
and low frequencies are not important, areasonable-size anechoic chamber
with refrigerator type doors can be used.
Although the acoustic environment is an important consideration for all
the noise measurements discussed in this chapter, the requirements of produc-
tion testing make the control of the environment a more difficult problem
than it is in aresearch and development laboratory.
A massive, tight, resiliently mounted enclosure is necessary to avoid
pickup of ambient stray noise that will affect the measurement. For the same
reason the access door must be one that seals exceptionally well. Then, in
order to get the required echo-free behavior, extensive treatment of the inside
is necessary.*
An enclosure with hard walls can also be used in some instances. Here the
design should be such as to make it areverberation room.
A much simpler technique is sometimes satisfactory for production-line
screening of noisy devices. This approach depends on a vibration measure-
ment that has been correlated with the acoustic noise. For example, acoustic
measurements of a number of samples may show that the noisy ones are
invariably noisy in one or two octave bands, say the bands at 500 and 1000
Hz. Then a measurement of the vibration of these samples may show what
vibration levels are acceptable in these bands. Some exploration of the vibra-
tion of the various surfaces of the device will be necessary to find the critical

*Anechoic chambers of various sizes are manufactured by, for example, the Eckel
Corporation, 155 Fawcett Street, Cambridge, Mass.

155
band is necessary and possible. The spectrum of the background noise is
usually different from that of the noise to be measured, and the corrections
in each band will be different.
lf this difference between background level and total noise level is small,
an attempt should be made to lower the background level. Usually the first
step is to work on the source or sources of this background noise to reduce
the noise directly. The second step is to work on the transmission path
between the source and the point of measurement. This step may mean
simply closing doors and windows, if the source is external to the room, or it
may mean erecting barriers, applying acoustical treatment to the room, and
opening doors and windows, if the source is in the room. The third step is to
improve the difference by the method of measurement. It may be possible to
select a point closer to the apparatus, or an exploration of the background
noise field may show that the measuring position can be shifted to a
minimum of this noise. The latter possibility is more likely when an analysis
is being made and the background level in aparticular band is unusually high.
It may also be possible to point the microphone at the apparatus to obtain an
improvement at high frequencies (see Figure 8-2); it may be necessary to use
adirectional microphone; or it may be desirable to use avibration pickup.

8.2.3 Instrument Precautions.

• Low Noise Levels — Effect of Circuit Noise. When low noise levels are to
be measured, the inherent circuit noise may contribute to the measured level.
This effect is usually noticeable in the range below 40 dB when a small
microphone is used or aceramic microphone is used on the end of avery long
cable. If the microphone is directly on the sound-level meter, the level at
which this effect may be important is below 30 dB, if the C weighting is used
or even lower if the A or B weighting is used. To measure the circuit noise the
microphone may be replaced by awell-shielded capacitor with acapacitance
equal to that of the microphone. A correction can then be made for this
noise, if necessary, by the same procedure as outlined for background noise in
paragraph 8.2.2. If the circuit noise is comparable to the noise being meas-
ured, some improvement in the measurement can usually be obtained by use
of an octave-band analyzer. The circuit noise in each band should be checked
also to see if correction is necessary.
Whenever low noise levels must be measured and extension cables are used,
the Type 1560-P42 or P40 Preamplifier should be used at the microphone.
Hum Pickup. When noise is measured near electrical equipment, acheck
should be made that there is no appreciable pickup of electro- magnetic field
in the sound-measuring system. The procedure depends on the directional
character of the field. The orientation of the instruments should be changed
to see if there is any significant change in level. If an analyzer is used, it
should be tuned to the power-supply frequency, usually 60 Hz, which would
be the 63-Hz band for the octave-band analyzer, when this test is made. If no
analyzer is included, the C-weighting should be used in this test to make the
effect of hum most noticeable, and a good-quality pair of earphones, with
tight-fitting ear cushions, should be used to listen to the output of the
sound-level meter.
If a dynamic microphone is used, tests should be made with different
orientations of the microphone, with the microphone disconnected, and with

180
variety of speeds, with the usual slow-scan techniques. One can, however, run
the engine for areasonable period at each of a number of selected nominal
speeds and record short samples, say five seconds, at each of these speeds to
form aseries of tape loops. Each loop is played back and is analyzed. Since
the inertia of the rotating system is often so large that serious fluctuations in
speed do not occur in the short interval of the tape loop, the engineer obtains
aseries of frequency spectra that can be related to shaft speed. He may then
be able to deduce much about the noise producing mechanisms from the
relations between amplitude, frequency, and shaft speed.
The analysis of intermittent sounds or signals can be a helpful step in
tracking down the sound sources. By means of tape recording, one can
monitor the noise from adevice for long periods, to catch these intermittent
sounds, which can then be separated out for analysis.
Background noise may make noise studies of machines impractical when
acoustically isolated rooms are not available. Often, however, the background
noise is much less during lunch periods or outside normal working hours,
particularly early in the morning, and measurements may then be practical.
Even during such periods acomplete study of the noise may be awkward or
inconvenient, but, if tape recordings can be made during the quiet periods,
the recorded signal can be analyzed at any convenient time.

CAL

ÍI

t=4

1562

Figure 7-5. Noise recording assembly.

The full range of analysis equipment is helpful in the product-develop-

ment phase. The detailed analyses, and the wide variety of techniques
available, make the 1923 FFT Analyzers invaluable in tracking down the
cause and sources of troublesome components of the noise. The other
analyzers provide varying degrees of detail in the analysis, and often one can
select the one that is most suitable for aparticular job from among the 1568,
1564, 1523 with -P3 plug-in, 1921, 1923, and 1933.
For any of these studies, the 1933 Precision Sound Level Meter and
Analyzer is helpful for making the basic reference measurements of overall
level, A-weighted level, and even octave-band analysis for checks on ratings as
the noise-control procedures are instituted. It also provides the impulse mode
for measuring noise from typewriters, trippers, chain drives, riveters, and the
like.
If vibration reduction is the prime goal, vibration pickups should, of
course, be used to supply the signals to the analyzers. But even if noise
reduction is the desired goal, the reduction is often accomplished by reducing
the vibration of various parts of the device. Here vibration pickups should be
used, or the motion should be studied with stroboscopic observation of the
moving parts. This procedure is described in Chapter 9.

157
 e
i 1560-9531
LAL

1568
I

1560-9669

\
1521-8
(t -

1560- P42
1562
156C - i-53

Figure 7-6. Narrow-band sound or vibration analysis.

CAL

1562 1933

Figure 7-7. Impact-noise analysis in octave bands.

Cat I

11111 I tamp aIam= amm

0•0
00000

LEEe 000100010 ®

1562 1933 170 1923

Figure 7-8. Narrow-band high-speed vibration analysis, including correlation, transfer

function, and summation analysis as well as spectrum analysis.

7.2.4 Machinery Preventive Maintenance. Only one aspect of machinery

maintenance is considered here, namely, the relation of the vibration level of
a machine to its condition. That is, vibration measurements can guide in
predicting incipient failure of a machine, in deciding when cleaning, parts
replacements, and other maintenance procedures are necessary, and in deter-
mining the relation between vibration and the performance of the machine.
The best general measurement for this purpose is velocity, and it is
usually measured at the bearing housings. The 1553 Vibration Meter provides
for velocity measurements, and it can be used for quick surveys. But since
analysis of the vibration signal is almost always desirable, the 1933 Precision
Sound- Level Meter and Analyzer with avibration pickup and acontrol box is
achoice that reduces the total number of instruments required.

158
 One basic technique in this form of preventive maintenance is to compile
a history of the analyzed vibration levels at each bearing housing. When the
levels change noticeably, the situation is reviewed to see if it is reasonable or
if it is likely that deterioration of some structure has occurred. It can then be
used as one guide in deciding when and how machinery is to be overhauled.
Octave-band analyses have been found helpful here (Glew and Watson, 1971),
but 1/3-octave analyses are also used ( Bowen and Graham, 1967). Here, the
history is conveniently compiled in the form of aseries of recorded vibration
levels taken on a 1523 Level Recorder with a 1523-P3 Stepped 1/3-Octave-
Band Analyzer plug in. The remarkable range of that analyzer, 1 Hz to
80,000 Hz, means that the full range of the vibration meter, including the
important low-frequency range, is nicely covered.

1560-P53

Figure 7-9. Tenth octave-band vibration analysis with stroboscopic observation.

When many machines are being monitored, so that many analyses need to
be made and recorded, the combination of a1921 Real Time Analyzer and a
1522 Dc Recorder provides arapid way of producing the data to be studied
for maintenance procedures.
Fine detail of analysis may be helpful in investigating certain faults of
rotating machinery, and the narrower-band analyzers, including the T/D 1923
FFT Analyzers, are then appropriate instruments to use. These FFT Analyz-
ers also provide correlation measurements and waveform averaging (summa-
tion analysis), which are useful investigational tools in machinery mainte-
nance (see paragraphs 5.4.1 and 5.4.5).

I56C - C53 1560-P.3

1557 '933

Figure 7-10. Velocity measurement and analysis sytern.

159
 CAl.

r
1560-P53 1560-P23

1523 - P3
.557

Figure 7-11. Recording system for compiling history of 1/3 octave vibration velocity
levels for continuing machinery preventive maintenance program.

AL

1.6r- P54

o
.557

1531

Figure 7-12. In-place balancing assembly.

One common source of trouble in machinery is rotational unbalance. A

measurement system to help in-place balancing consists of aType 1560-P54
Vibration Pickup, the Type 1564 Sound and Vibration Analyzer, Type
1531-P2 Flash Delay, and the Type 1531 Strobotac.

7.3 ENVIRONMENTAL (COMMUNITY) NOISE.

7.3.1 Hearing Damage Risk from Noise Exposure. As described in Chap-

ter 4, the noise near some machines is intense enough to cause permanent
hearing damage, if the exposure continues for long periods. As explained, the
main technique for checking potentially dangerous areas is to measure the
A-weighted sound level with a standard sound-level meter. If the sound is
impulsive, the sound-level reading should include apeak reading taken on a
1933 Precision Sound-Level Meter and Analyzer. Alternatively, a 1556-B
Impact Noise Analyzer connected to the 1565 sound-level meter is available
to measure this peak value.
A more convenient and satisfactory measurement of the noise exposure is
possible with a 1934 Noise Exposure Monitor, which combines the observed
levels according to the current regulations and also monitors the peak level.
These measurement procedures are the ones currently used under the
regulations of the Williams-Steiger Occupational Safety and Health Act of
1970 ( OSHA, 1970), the Walsh-Healey Public Contracts Act, as amended, and
the Federal Coal Mine Health and Safety Act of 1969.
As pointed out in Chapter 4, if potentially hazardous noise exists in a
plant, audiometric examinations of exposed personnel and work to control
the noise are necessary. The audiometric examinations can be contracted out,
or with the use of a 1703 Automatic Recording Audiometer, they can be

160
done at the plant, as described in Chapter 4. Other instrumentation is useful
in noise control, which is discussed in Chapter 9and Section 7.2.

7.3.2 Local Noise Ordinances — Area Noise. Some cities and towns
regulate the maximum noise levels permitted at lot boundaries, with the
limits set according to the district. Most of these ordinances now use the
A-weighted sound level measured on a sound-level meter although some
specify octave-band levels.
In the 1971 Chicago Noise Ordinance, for example, the noise level in
business and commercial districts at the lot boundary is not to exceed 62 dB
(A). In residential areas the noise level coming from a residence is not to
exceed 55 dB ( A) at the lot boundary. In manufacturing, districts the noise is
measured at the district boundaries and the limits range from 55 to 66 dB
(A).
The sound-level meter should be used with awindscreen for this monitor-
ing. Since the operations in afactory can vary considerably during the day
and night cycle, some plants may require monitoring with arecorder on the
output of asound-level meter for long periods of time. A 1565 Sound-Level
Meter and a 1523 Level Recorder with a1523-P1 Amplifier plug-in or a 1933
Precision Sound Level Meter and Analyzer and a simple, dc strip chart
recorder can be used.
The extensive use of air-conditioning units, particularly those with out-
door heat exchangers, has made noise monitoring more important in residen-
tial areas. Air conditioners can be particularly bothersome at night when
some wish to have their bedroom windows open, and monitoring the noise on
ahot night is particularly appropriate.

7.3.3 Motor-Vehicle Noise. Some states and some cities and towns in the
USA and many other countries have ordinances or laws that set permissible
limits on noise from motor vehicles. These are also generally specified in
terms of the A-weighted sound level at some distance from the vehicle.
Chicago sets limits on the vehicles as sold, as well as in operation.

7.3.4. Powered Equipment Used Outdoors. Much construction and indus-

trial equipment is used outdoors, and many power tools, power lawnmowers,
riding tractors, etc. are used in residential areas. These are also being regulated
in an attempt to reduce the noise nuisance, again mainly by the specification
of maximum A-weighted sound levels.

7.3.5 Airport Noise. Noise of busy airports is of great concern to those

living nearby. In some areas vigorous community action has resulted, and
much effort is now being devoted to studies of how to control and reduce
this noise impact. Some airports have extensive noise-monitoring systems. By
imposing maximum limits on the noise level permitted for the aircraft that
use the airport, the airport operator can bring the noise problem under better
control. By gradually lowering those limits as quieter aircraft are developed,
he can effect further improvements.
California is requiring noise monitoring with maximum noise level limits
for all airports that have a noise problem. A-weighted sound levels, as
determined with the slow dynamic characteristic, are used. The A-weighted
levels are summed over time on an energy basis to obtain a noise-exposure
level ( NEL), referenced to 20 µN/m 2 and one-second duration. This noise-
exposure level is used in different ways to obtain limits on the allowable

161
levels, either as asingle event or as adaily community noise equivalent level
(CNEL). The accumulation for CNEL is adjusted to give more weight to the
noise occurring between 1900 and 2200 hours and most weight to noise
occurring between 2200 and 0700 hours. The annual CNEL at different
locations is used to determine the noise impact area according to the bound-
ary at which the annual CNEL is equal to aset value and according to the
land use. The criterion value set for the CNEL is to be lowered in subsequent
years.
These requirements on the noise monitoring systems show that some-
thing appreciably more complex than a sound-level meter and recorder is
required, and specialized systems have been developed for this purpose. A
noise-monitoring terminal for this use should contain a microphone, an A-
weighting network, squaring, averaging, integrating, conversion and timing cir-
cuits, and an output recording or logging device. This terminal can be used as
a separate monitoring device, or, similar terminals with the addition of con-
version circuits and transmission lines can be tied in as part of a computer-
controlled system to monitor the noise over a wide area surrounding an
airport.

7.4 ARCHITECTURAL ACOUSTICS.

A wide variety of measurements are made in the field of architectural
acoustics ( Meyer and Kuttruff, 1962). Some of these are formalized by
standards, and others require individual judgment in deciding what needs to
be measured and how it needs to be measured. Detailed analysis of all these
measurements is beyond the scope of this book, but useful references will be
cited in the brief summaries given here. In particular, Davis ( 1965) gives
helpful discussions of instrumentation and details of measurement.

7.4.1 Sound Absorption. A method of test for the sound absorption of

acoustical material in reverberation rooms is described in the American
Society for Testing and Materials C423-66 (also ANSI S1.7-1970). This
standard specifies the requirements for the reverberation room, the test signal
at test frequencies of 125, 250, 500, 1000, 2000, and 4000 Hz, the mounting
of the test specimen, and the measurement procedure. One possible arrange-
ment of instruments is anoise generator supplying an octave-band analyzer,

o §
POWER
AM PLIE PER
1382

CAL
560-P5

1523, - P1 1564

52

Figure 7-13. A systein for sound absorption measurements.

162
followed by a power amplifier, which drives some loudspeakers that can be
switched on and off. A microphone is connected by means of acable to a
1/3-octave analyzer, which drives agraphic level recorder.
The particular instruments chosen may depend on other measurements to
be made. Thus, the 1521-B recorder could be substituted for the 1523 re-
corder with the 1523-P1 preamplifier plug-in. The octave-band filter is not
essential; it helps in overriding background noise.
The basic procedure is to measure the decay rate of sound, which is a
band of noise, with and without the acoustical material in the room. These
two measurements permit one to calculate the sound absorption of the
specimen.

7.4.2 Sound Transmission Loss. The American Society for Testing and
Materials E90-70 gives the recommended practice for the Laboratory Meas-
urement of Airborne Sound Transmission Loss of Building Partitions. Here,
two reverberant rooms with acommon wall, the wall under test, are required.
A diffuse noise field is set up and measured in one room and is also measured
in the other. From these two measurements and ameasurement of the sound
absorption in the second room (see paragraph 7.4.1) the transmission loss can
be calculated.
One possible arrangement of instruments is anoise generator followed by
a power amplifier, which drives loudspeakers in the first room. A microphone
in each room is connected by means of cables to a 1/3-octave analyzer, so
that the level in each room can be monitored. A sound-level calibrator is
essential in order to compare the sensitivities of the two microphone systems.

DIFFUSING VANES

POWER
AMPLIFIER TEST
MATERIAL

1933
CAL 1560-P5 TEST ROOMS 15 60 9531

r§ql

1562 1564

Figure 7-14. A system for sound transmission measurements.

Since agood space average of the sound-pressure in each room is desired,

a number of microphone positions may be needed in each room. Rather than
moving a microphone from one position to another, agroup of microphones
can be scanned by a 1566 Multichannel Amplifier. One 1566 can be used for
both rooms. With the 1562 Sound-Level Calibrator as asource, the sensitivity
of each transducer can be equalized by the individual-channel gain adjust-
ments in the 1566. If a 1921 Real-Time Analyzer is used for the 1/3-octave
measurements, the squared outputs of the microphones can be combined in
the detector system as they are scanned. The set of microphones in one room
would be scanned and then the set in the second room.

163
varies approximately as the volume of the room, the effect in small, treated
rooms is usually negligible.
The relation given above is shown graphically in Figures 8-8 and 8-9, where
Figure 8-8 applies to the nondirectional or simple source or to adirectional
source in the direction having adirectivity factor of 1 ( Q=1), and Figure 8-9
applies to the directions having the labeled values of directivity factor.

METERS
•5 0.2 0.5 2 5 10 zo
1 , 1 I I 1i

R•50 ft 2

4.6
1Ó00 2
93 ,,,2

200f 1
2

OPEN AIR 19.6m 2

Rrm 1
5000 2
I
116ne
1000ft 2
93m 2
2000ft 2

166rn 2

50001, 2

465, 2
10.0001, 2
930. 2
0,0001, 2

1860. 2

I
0.5 07 2 3 4 5 7 10 20 30 40 50 70 100
DISTANCE FROM ACOUSTIC CENTER
OF A NONDIRECTIONAL SOURCE IN FEET • r

Figure 8-8. Chart shoWing the sound-pressure level, SPL, relative to the power level,
PWL, for a nondirectional source for different values of the room constant, R, as a
function of the distance from the source.

• Reverberant Field. The graphs of Figure 8-8 and 8-9 show that close to the
source the sound-pressure level tends to vary with the distance from the
source as it does under free-field conditions ( R = 00). But far from the source
the sound-pressure level becomes independent of the directivity and the
distance to the source. This region is called the reverberant field. Here, the
level is determined by the acoustic power radiated by the source and the
acoustic characteristics of the room. The region over which the transition
between the free-field behavior and the reverberant field occurs is determined
by the directivity factor and the room constant.
Actual Room Behavior. In a well-designed reverberation room, the be-
havior on the average is similar to that shown in the figures. Most other rooms
have characteristics that on the average fall between that reverberant behavior
and the free-field sound-pressure level decrease of 6 dB for each doubling of
the distance ( Ogawa, 1965; Gober and Lubcke, 1966; Yamamoto, 1961,
Peutz, 1968).
In flat rooms (i.e., rooms whose ceilings are low relative to room length
and width), the sound-pressure level at adistance from the source tends to
decrease a fixed amount, but less than 6 dB, for each doubling of distance.
The decrease depends on the sound absorption in the room. In very long
rooms or halls, the sound-pressure level tends to decrease afixed number of

194
a. Record
test signal.

b. Playback
signal and
record hall
response.

C. Playback
hall response
for analysis.
— 11

1935 1564 1521 -

Figure 7-16. Use of tape recorders to perform an acoustic analysis of an auditorium.

When it is necessary to simplify the testing procedure at ahall (if, for

example, an audience is present), an alternative procedure may give useful
information with relatively short testing time. The hall is excited by means of
a broad-band of " pink" noise (noise whose spectrum level decreases with
increasing frequency to yield constant energy per octave of bandwidth). The
sound in the hall is recorded on atape recorder with the noise on and during
the decay period when the noise is suddenly switched off. Later, in the
laboratory, this recorded signal is played back through a1/3-octave analyzer
and a graphic level recorder. This recording is repeated for the full range of
the desired settings of frequency of the analyzer. In this way, a complete
picture of the decay rate as a function of frequency can be obtained with
only one " exposure" at the hall. Some of the detail, particularly at lower
levels, may be obscured, however, by the background noise in the hall.

7.4.4 Response Testing. The equalization of asound system, to obtain

good uniformity of response with frequency, can improve the performance of
the system. The suggested procedure for measuring the response is as follows:
The sound system is driven by broadband pink noise. The sound produced is
then analyzed by arecording 1/3-octave analyzer ( Figure 7-17). This analysis
is carried out for anumber of representative places in the area to be covered
by the sound system.
The adjustments in the frequency response are then made on the basis of
astudy of the response curves. The fine details of the curves are ignored, and
the broader trends are used to decide how best to adjust the frequency-re-

165
measure the sound-pressure levels, the method of calculating acoustic power,
and the requirements on the characteristics of the space in which the
measurement is to be made.
Four different types of environment are considered in the discussion of
sound-power measurements, that is, free field or anechoic, free-field above a
reflecting plane, reverberation room, and asemireverberant field. The choice
among these is determined by many factors, most of which will become
evident from the descriptions that follow. The influence of development in
instrumentation are not described in the procedures, but they need to be
considered also. In particular, the fewer microphone positions required for
the reverberant room measurement is no longer as significant afactor as it
was. The techniques for scanning a number of microphone outputs and
summing on a pressure-squared basis, as provided by the 1566 Multichannel
Amplifier and the 1921 Real-Time Analyzer, now simplify the anechoic
measurement or the free-field measurement above a reflecting plane to the
point that their other advantages make them more attractive than formerly.
• Measurement Procedure& The source characteristics are obtained by use of
the principles discussed earlier in this chapter.* Generally, the following
characteristics must be determined:
1. The total sound power radiated by the source, as expressed by the
power level, as afunction of frequency.
2. The directional characteristics of the source, as expressed by the
directivity factor, as afunction of direction and frequency.
• Measurements Around the Source. If free-field conditions can be closely
approximated, the power level and directivity can be calculated from the
sound-pressure levels measured at a number of points. These measurements
are made at points at equal distances from the source and all around the
source. The points can be considered as being on the surface of ahypothetical
sphere surrounding the source. The radius of this sphere should be at least
twice the largest dimension of the source but not less than 2ft ( 0.6 m).
If the equivalent of afree field is produced by extensive treatment of the
surfaces of aroom, the hypothetical measurement sphere should not be closer
to the absorbent surfaces than 1/4 wavelength corresponding to the center
frequency of the lowest frequency band of interest. Since anechoic chambers
built with wedges have the wedges about one-fourth wavelength long at the
lowest frequency of interest, one can readily estimate the minimum dimen-
sions for an anechoic chamber. For a noise source less than a foot in
maximum dimension, the wall-to-wall inside distance should then be at least
one wavelength plus 4 ft ( 1.2 m). The following table gives this value for
some limiting frequencies.
X+4 X+1.2
(Hz) (ft) ( m)

100 15.3 4.6

125 13.0 4
160 11.1 3.4
200 9.6 2.9

Theoretically, the sound-pressure levels over the entire surface of the

sphere should be measured. The practical procedure for approximating this

*The procedures outlined here and in subsequent sections are similar to those given in
ANSI S1.2-1962, " Standard Method for the Physical Measurement of Sound," and that
should be consulted for specific details on the standard method.

197
frequency noise is much greater than that of isolating high-frequency noise,
and the knowledge of the level of low-frequency sounds and vibrations may
be an essential element in cost studies.
A related problem is that of locating astudio within an existing building.
Here, acareful survey of possible locations may lead to asignificant saving in
construction costs.

REFERENCES
L. L. Beranek ( 1949), Acoustic Measurements, John Wiley 8c Sons, Inc., New York.
K. A. Bowen and J. S. Graham ( 1967), " Noise Analysis: A Maintenance Indicator,"
Mechanical Engineering, Vol. 89, # 10, 1967, pp. 31-33.
D. Davis ( 1965), Acoustical Tests and Measurements, H. W. Sams, Indianapolis.
FAA ( 1969), Federal Aviation Regulations, Part 36-Noise Standards: Aircraft Type
Certification, Federal Aviation Administration, Department of Transportation.
Federal Coal Mine Health and Safety Act ( 1969), Public Law 91-173, 91st Congress,
S.2917, December 30, 1969.
C. A. W. Glew and D. C. Watson ( 1971), "The Octave Band Vibration Analyzer as a
Machinery Defect Indicator," Am Soc of Mech Engineers, Paper 71- DE-47, April
1971.
E. Meyer and H. Kuttruff ( 1962), " Progress in Architectural Acoustics," in Richardson
and Meyer, 1962, pp. 222-277.
OSHA ( 1970), " Part 1910- Occupational Safety and Health Standards," under authority
of Williams-Steiger Occupational Safety and Health Act of 1970 ( P.L. 91-596) ( 84
Stat. 1593, 1600; 29 U.S.C. 655, 657) Federal Register Vol. 36, # 105, May 29,
1971, Washington, D.C.
E. G. Richardson and E. Meyer, eds. ( 1962), Technical Aspects of Sound Vol ///,
Elsevier Publishing Company, Amsterdam.
T. J. Schultz, " Problems in the Measurement of Reverberation Time," Journal of the
Audio Engineering Society, Vol. 11, #4, October 1963, pp. 307-317.
T. J. Schultz and B. G. Watters ( 1964), " Propagation of Sound Across Audience
Seating", JAcoustical Society Am, Vol. 36, #5, May 1964, pp. 885-896.
WHPCA ( 1969), Occupational noise-exposure regulations under Section 50-204.10 of
the Walsh-Healey Public Contracts Act - Safety and Health Standards, Federal
Register, Vol. 34, #96, May 20, 1969, ammended July 1, 1969.

167
Chapter 8

Techniques, Precautions,
and Calibrations
8.1 INTRODUCTION.
The previous chapter was designed to help in deciding what measurements
need to be made for a given acoustic or vibration problem. This chapter
discusses how to make the measurements. Other chapters provide help in
interpreting the results of the measurements.
The goal is to make valid measurements ( Stein, 1962). In order to achieve
this goal, it is helpful to recognize that the results of a measurement are
determined by anumber of factors, among which are the following:
1. The phenomenon being measured.
2. The effect of the measurement process on the phenomenon being
measured.
3. The environmental conditions.
4. The characteristics of the transducers and instruments being used at the
time they are used.
5. The way the transducers and instruments are used.
6. The observer.
Although many useful measurements are made by those with little back-
ground in acoustics, the chances of making valid measurements are increased
as the understanding of these factors becomes more thorough. Thus agood
knowledge of vibration and acoustics, of transducers, of instruments, and of
measurement techniques, is helpful in making noise measurements. In this
chapter we shall, therefore, provide information that is particularly pertinent
for measuring noise.
Even when one does not need to measure noise according to astandard
procedure, it is often wise to try to do so if an appropriate standard can be
found. The standards have been prepared to help obtain valid data. They are
useful guides for the inexperienced, and they help the experienced to keep in
mind the required steps in a measurement procedure. They help to make
comparisons of measured results more meaningful.
Those who prepare the standards try to recognize as many of the problem
areas as they can and they attempt to set the requirements and procedures to
bring them under control. But, often significant compromises have to be
made because not enough is known to resolve the problem, or the available
instrumentation may be inadequate. As the state of the art advances, the
standards can be improved correspondingly. It is, therefore, important to use
the most recent standards.
The general standard ANSI S1.13-1971, " Standard Methods for the Meas-
urement of Sound Pressure Levels," is particularly recommended.
A thorough study of the instruction books supplied with the instruments
to be used will often make it clear how to make the most effective use of the

169
istruments. Practice in their use on familiar sounds is also helpful, and
coustical calibrating signals are particularly good for this purpose.
As implied in the listing of factors that determine the measurement results,
he use of aprecision instrument does not guarantee that ameasurement will
c valid or accurate. When measurements are done properly, however, a
'recision instrument will help to yield more consistent results than is possible
rith a less precise instrument. Better measurement techniques then will be
:ss limited by instruments, and improved results can be obtained more
eadily.
An obvious but important rule in any measurement task is to review the
esults to see if they are reasonable. If they are not, try to track down
mssible sources of trouble, particularly simple things like poor connections,
dugs in the wrong places, no power, low batteries, controls set incorrectly,
lamaged equipment, stray grounds and pickup. If nothing can be found that
:an be corrected to bring the data into line, perhaps the data only seem
tnreasonable because of limited understanding of the phenomena or of the
neasurement process.
The results of a noise measurement may be a key factor in resolving a
mise problem. In addition, the experience and data often help in doing a
)etter job on another noise problem. Careful records of noise measurements
:an be valuable for future reference on subsequent problems, and this
mssibility should be kept in mind in tackling anoise problem.
A recognition of the accuracy limitations of acoustic and vibration meas-
irements is important, in order to be reasonable in the approach to a
measurement problem. Thus, consistency to 0.1 dB or better is attainable in
Dnly afew laboratory calibration procedures in acoustics and not in general
icoustical measurements. Field calibrations of sound-level meters at one
frequency with acalibrator may be consistent to 0.5 dB or slightly better. A
consistency of 1dB is difficult in general measurements, even under carefully
controlled conditions, but it is amore reasonable goal than 0.1 dB.

8.2 SOUND MEASUREMENTS.

Most of the applications discussed in the previous chapter require a
measurement of either sound-pressure level as afunction of frequency or of
sound level. These quantities are measured at asingle point or at anumber of
points that are determined by the conditions of the application.
The basic procedure for measuring the sound level or the sound-pressure
level at a given point is to locate the sound-level-meter microphone at that
point and to note the reading of the sound-level meter. Some preliminary
exploration of the sound field is usually necessary to determine that the point
selected is the correct one, and this exploration is discussed later in this
chapter. Other practical details regarding this measurement are also given in
this chapter, but the actual manipulation of the individual instrument con-
trols is discussed in the instruction books that are furnished with the instru-
ments.
We shall discuss the choice of microphone and auxiliary apparatus, the
effects of extraneous influences, the recording of adequate data, the calibra-
tion of the instruments, and the interpretation of the data. Much of this
discussion is necessary because no ideal instrument or combination of instru-
ments and accessories is available that would be suitable for all conditions.

170
 measurements are usually made in octave bands. The measured levels
are averaged on an energy basis for each band.
b. The unknown source is operated in place of the standard source. The
sound-pressure levels are measured at the same points as before and
averaged for each octave band.
c. For each octave band the difference in average level, between the
standard and the unknown, is applied to the known power level of the
standard source.
Requirements for Standard Source. The standard source should produce a
stable and reproducible sound. Such sources have been developed for the
Compressed Air and Gas Institute and for the fan and blower industry
(Hardy, 1959).
Requirements for Room. The measurement room should be large and its
characteristics should approach those of areverberant room. No obstructing
object should be in the immediate vicinity of the source or the microphone
positions.

• 8.3.6 Predicting Noise Levels. When the acoustic power output and the
directivity pattern of a device are known, the noise levels that it will produce
under a variety of conditions can be predicted (on the average) with fair
accuracy. These predictions are based on the principles discussed earlier in
this chapter.
If anoisy device is placed in aroom that is not anechoic, it is desirable to
measure the decay rate of sound, D, in the room; and then the following
formula, adapted from one by Young, can be used to predict the average level
of sound in that part of the room where the reverberant field dominates:
SPL = PWL — 10 log V — 10 log D + 47.8

where V is the volume of the room in cubic ft. PWL is the source power level
and the constant 47.8 varies with atmospheric pressure (to determine the
variation add 0.5 dB to the values shown in Figure 8-15).
Close to the source, the level is almost as if free-field conditions existed.
The level decreases with increasing distance from the source and the average
approaches the reverberant field level. Here, standing waves will exist and it is
only the average level that can ordinarily be predicted. At points less than 1/4
wavelength from ahard wall, the level will be higher than the average in the
reverberant field. Very near ahard wall the increase may be as much as 3dB;
very close to an edge, 6dB; and right at the vertex of acorner, 9dB.
When the decay rate in the room cannot be measured, it can be estimated
from a detailed knowledge of the room and its surface conditions. The
procedures are given in books on architectural acoustics. There the calcula-
tion procedure is normally given for reverberation time, T. The decay rate, D,
is then easily obtained as follows:

A related procedure is to calculate the room constant, from the character-

istics of the room. Then the equation given there can be used to estimate the
sound-pressure level. The discussion in paragraph 8.3.4 should also be con-
sidered in order to modify the prediction to fit the actual conditions.
As described earlier, the near-field behavior of asource may not be closely

204
nished with the earlier Types 1551-A, 1551-B, and 759-B Sound Level
Meters, and in the Type 1555-A Sound-Survey Meter, however, gradually
dissolves if the humidity is too high (above about 84%). The Rochelle-salt
crystal unit in the microphone, however, is protected by acoating, so that it
is relatively unaffected by high humidity. Nevertheless, it is wise to avoid
unnecessary exposure. A Rochelle-salt microphone should not be stored for
long periods in avery dry atmosphere, since it can dry out.

NOTE
Rochelle-salt-crystal microphones are no longer supplied by Gen-
eral Radio Company.

A condenser microphone is not damaged by exposure to high humidity,

but its operation can be seriously affected unless proper precautions are
taken. For proper operation it is essential that very little electrical leakage
occur across the microphone. The exposed insulating surface in the micro-
phone has been specially treated to maintain this low leakage, even under
conditions of high humidity. In spite of the precaution, the leakage may
become excessive under some conditions. Then it may be advisable to keep
the microphone at a temperature higher than the ambient temperature to
reduce the leakage. In climates where the humidity is normally high, it is
recommended that the microphone itself be stored at atemperature above
ambient to avoid condensation.
An electret microphone can stand for long periods the normal variations in
temperature and humidity without significant change in sensitivity. But, if
the humidity is normally high, an electret microphone should be stored in a
small jar containing silica gel. A relative humidity of 90% is the maximum
recommended for the electret.
High or Varying Temperature. Although most noise measurements are
made indoors at average room temperatures, some measurement conditions
expose the microphone to much higher or lower temperatures. When these
conditions are encountered, it is essential to know the temperature limita-
tions of the equipment.
The microphones supplied with the 1565 and 1551-C Sound-Level Meters
will withstand temperatures of —40° to +60 °C without damage. Even at 95 °C,
apermanent sensitivity loss of only about 0.5 dB may occur. In contrast, the
maximum safe operating temperature for Rochelle-salt crystal microphones,
formerly furnished with some sound-level meters, is about 45 °C ( 113 °F). At
55.6 °C ( 132 °F), the Rochelle-salt crystal is permanently changed. It is, there-
fore, not safe to put aRochelle-salt microphone in the trunk or back of acar
that is to be left standing in the sun.
The normal operating temperature range for the condenser microphone
systems is from — 30° to +65°C. They will withstand higher temperatures
without damage, but alimit of 80°C is recommended for the preamplifiers.
The electret condenser microphone should be limited to — 25° to + 55°C
(131° F).
Fortunately, it is usually possible to keep the sound-level meter itself at
more reasonable temperatures. Its behavior at extreme temperatures is limited
by the batteries. Temperatures of even 55°C ( 130°F) will result in much-
shortened battery life. Operation below — 10°F is not ordinarily possible
without special low-temperature batteries.

172
 Microphones are usually calibrated at normal room temperatures. If a
microphone is operated at other temperatures, its sensitivity will be some-
what different and acorrection should be applied. The correction for sensitiv-
ity for the ceramic microphone is only about —.01 dB per degree Celsius, so
that for most purposes the correction can be neglected.
Hum Pickup. Dynamic microphones are sometimes used for measurement
purposes because they are readily used with long cables. The development of
modern preamplifiers, such as the 1560-P42 Preamplifier, makes the use of
dynamic microphones unnecessary. But if they are used, care must be taken
to avoid hum pickup, which is the induction of undesired electrical signals
from the external magnetic field of equipment such as transformers, motors,
and generators. Ceramic and condenser microphones are relatively free from
this undesirable effect.
Long Cables. For the most accurate sound measurements, only the micro-
phone should be put into the sound field, and the measuring instruments and
the observers should not be near the point where the sound-pressure is to be
measured. For this reason, and also for the situations when it is impossible or
impractical for the observer to be near the microphone, an extension cable is
ordinarily used to connect the microphone to the instruments. If the micro-
phone is attached directly to apreamplifier, long cables can be used without
any deleterious effects. Condenser and electret microphones should almost
always be used directly on apreamplifier.
When amicrophone is used directly with an extension cable, acorrection
for loss in sensitivity is necessary. This correction is readily determined by the
use of a Type 1562 Sound-Level Calibrator. (See paragraph 8.2.3.) The
correction is about 7 dB when a 25-ft cable ( 650 pF) is used between a1-in.
ceramic microphone and the instrument, so that 7dB should be added to the
indicated level to obtain the level at the microphone. For longer cables the
correction is greater. For Rochelle-salt microphones, the correction is a
function of the temperature of the microphone; values are given in the
instruction manuals for instruments using such microphones.
Wind Effects. The microphone should also be kept out of any appreciable
wind, if possible. Wind on the microphone produces anoise, which is mainly
of low frequency as shown in Figure 8-1. This added noise may seriously
upset the measurement, particularly for high wind speeds, since the noise
increases with wind speed. If it is not possible to avoid wind on the
microphone, a wind screen should be used. The 1560-9521 Wind Screen
reduces the wind noise significantly, as shown in Figure 8-1, without a
serious effect on the frequency response.
Since the wind noise is mainly of low-frequency, the use of A-weighting
reduces the over-all wind-induced noise level markedly. It is, therefore, pos-
sible that a useful A-weighted level can be measured, even though aflat or
C-weighted level of the noise source is obscurred by wind noise. But even here
the wind screen is adesirable addition.
Direction of Arrival of Sound at the Microphone. Some microphones are
designed to be directional at all frequencies. That is, the response of the
microphone depends on the direction of arrival of the sound wave. Most of
the microphones used for sound measurements, however, are essentially
omnidirectional at low frequencies (below about 1kHz). At frequencies so
high that the size of the microphone is comparable to the wavelength of the
sound in air, even these microphones will show directional effects. This

173
 100
I
25mph- 40 km/h
1
90
))

80
NO WIND SCREEN

70 +

83dB
OVERALL
BAND LEVEL - dB

60
WIND SCREEN

50 +

40

30

o 0

20 WIND NOISE REDUCTION

Figure 8-1a. Wind noi:

o
spectrum- flat weightir
10

25 50 100 200 400 1000 2000 4000 10,000

f- Hz ( 1/3- OCTAVE)

80 1 1
25mph-40 km/h

— +
70
I
NO WIND
SCREEN

60 +
BAND LEVEL - dB

50 56 - dBIAI

7,,••••
+ _ + + +
WIND
40 SCREEN

30

+
+
20

Figure 8-1b.
spectrum-A weighting
10

o
25 50 100 200 400 800 1000 2000 4000 10,000

f- Hz ((/ 3 OCTAVE)

174
 +10

+8 FREE- FIELD PERPENDICULAR

RESPONSE CHARACTERISTIC

til I
+6

RANDOM- INCIDENCE
+4
RESPONSE CHARACTERISTIC

0°

GRAZING- INCIDENCE
-4 RESPONSE CHARACTERISTIC

1020
50 70 100 200 500 700 1k 2k 5k 7k 10k 20k

FREQUENCY IN HERTZ
a.

12

10
PERPENDICULAR
INCIDENCE

RANDOM
INCIDENCE

GRAZING INCIDENCE

620 50 70 100 200 500 700 lk 2 5 7 10 20

FREQUENCY IN HERTZ
b.
Figure 8-2. Typical response characteristics for
the ceramic microphones, (al 1- in., ( b) Win.

change in response with direction should be considered in positioning the

microphone for a measurement. The extent of these variations is shown by
the frequency response characteristics of the different microphones ( see
Figure 8-2). The microphone is usually positioned so that the response to the
incident sound is as uniform as possible.
When sound-pressure level is measured in a reverberant room at a point
that is not close to a noise source, the sound arrives at the microphone from
many different directions. Then the orientation of the microphone is not
critical, and the response is assumed to be that labeled "random incidence."
Under these conditions, nevertheless, it is usually desirable to avoid having
the microphone pointing at anearby hard surface, from which high-frequency
sounds could be reflected to arrive perpendicular ( 00 incidence) to the plane
of the diaphragm. ( For all the microphones used in the General Radio Sound

175
Measurement System this perpendicular incidence is along the axis of cylin-
drical symmetry of the microphone. This axis is used as the 0° reference line.)
If this condition cannot be avoided, the possibility for errors from this effect
can be reduced by some acoustic absorbing material placed on the reflecting
surface.
When measurements are made in areverberant room at varying distances
from a noise source, the microphone should generally be oriented so that a
line joining the microphone and the source is at an angle of about 70° from
the axis of the microphone. When the microphone is near the source, most of
the sound comes directly from the source and a 70° incidence response
applies. On the other haild, near the boundaries of the room the incidence is
more nearly random and the random-incidence response applies. These two
response curves are nearly the same, so that there is little change in the
effective response characteristic as the microphone is moved about the room.
This desirable result would not be obtained if the microphone were pointed
at the noise source.
If, however, asource is to be measured in anearly anechoic or free-field
space, the use of a microphone with a uniform response for perpendicular
incidence may be preferred. Then the microphone can be pointed at the
source, and the directional behavior will help to reduce the effects of
extraneous noises. Although this type of microphone is acceptable for inter-
national standards on sound-level meters, it is not acceptable for ANSI
S1.4-1971.
Position of Microphone. In previous sections of this chapter some com-
ments have been made on various aspects of the problem of placing the
microphone in the most satisfactory position for making the noise measure-
ment. In general, the location is determined by the type of measurement to
be made. For example, the noise of a machine is usually measured with the
microphone placed near the machine according to the rules of atest code, or
if its characteristics as a noise are desired, a comparatively large number of
measurements are made according to the methods and the placement given in
paragraph 8.3.5.
It is important to explore the noise field before deciding on a definite
location ( paragraph 8.3.4) for the microphone.
Many measurement locations may be necessary for specifying the noise
field, particularly if the apparatus produces anoise that is highly directional.
Further discussion of directionality is given in 8.3.2.
If the noise level is measured for calculation of the speech-interference
level or loudness level or for determination of deafness risk, it is important to
explore the noise field to make sure that the measurement made is representa-
tive. The possible effects of obstacles in upsetting the distribution of sound,
particularly at high frequencies, should be kept in mind during this explora-
tion.
At first thought, it seems logical, when measurements regarding noise
exposure are made, to mount the microphone at the operator's ear. Actually,
because of the variables introduced by the effect of the operator's head being
close to the microphone, this technique is not used, except in certain
scientific tests with special probe microphones. All ratings of speech-interfer-
ence, loudness, and deafness risk are based on ameasurement with no person
in the immediate vicinity of the microphone. The microphone should, how-
ever, be about where the operator's ear would normally be.

176
 8.2.2 Effects of Room and Nearby Objects.
Effect of Observer and Meter Case. As mentioned in the previous section,
the observer can affect the measured data if he is close to the microphone.
When measurements are made in alive room and not close to asource, the
effect is usually not important. But if measurements are made near asource,
it is advisable for the observer to stand well to the side of the direct path
between the source and the microphone. For precise measurements in avery
dead room, such as an anechoic chamber, the instruments and the observer
should be in another room with only the source, the microphone, the
extension cable, and aminimum of supporting structure in the dead room.
For many measurements, however, it is most convenient to be able to
carry the sound-level meter around. When held in the hand, the sound-level

10

5 •••••••••

A
o A WA
-5 o
w -10
Ui
to EFFECT OF OBSERVER
510
o
5

1020
100 200 1000 2000 4000
FREQUENCY IN HERTZ

10

o
A

EFFECT 0F OBSERVER
— L_L.1à .
b E_P
.... .Wtft)
....„4..

..—,

Inn InnÉl ,Inete, •,•

FREQUENCY IN HERTZ

Figure 8-3. Effect on frequency response as aresult of using the microphone directly on
the swivel post of the instrument (in two positions) without an observer present, and the
extent of the effect with an observer present. A single-frequency, plane, acoustic wave
was used in an anechoic chamber, and the decibel values are the differences between the
response under the conditions noted and the response of the microphone alone.

177
 The effect of fastening by means of double-sided tape was generally less
han 10% deviation from the reference condition at all frequencies up to the
•esonance at 3200 Hz. In some instances, the deviation over the range to
1200 Hz was only about 2%. The variability was probably aresult of changes
ncontact adhesion obtained with different samples of the tape.
Plasticene as afastening means, even as thick as 1/16-in, showed very good
eproduction of the reference performance, being within 2to 5% up to 4000
.

U. In one instance, amarked departure from the reference performance was

'ound even at 500 Hz, and this was quickly traced to the fact that the pickup
iad come loose from the plasticene. This example illustrates the importance
)f careful inspection of the fastening during a test, particularly when one
:annot check the performance independently.
The response of the pickup when held to a smooth, flat, steel plate by
neans of the Type 1560-P35 Permanent Magnet Clamp is shown in Figure
l-21. Up to 5kHz, the response is very similar to the reference response. One
hould fasten the pickup carefully to the magnet so that no rocking motion is
)ossible, and the magnet itself should be placed on asmooth surface so that
t, too, will not rock; otherwise, serious errors may result.
25

20

15
dB

I
0
IN
RESPONSE

5
M60- P35
O PERMANENT _
MAGNET CLAMP

-5 REFERENCE

10

-I 5
200 500 1000 2000 5000 1Q000 20,000
FREQUENCY IN Hz Œffa
=igure 8-21. Frequency response of vibration pickup attached by Type 1560-P35
2ermanent-Magnet Clamp.

8.4.3 Stray Effects.

Effect of the Pickup on the Vibration. The mass added by the pickup to
he vibrating surface being measured changes the motion of that surface. If
he added mass is much smaller than that of the vibrating surface and is
:losely coupled to it, the effect is small except near resonant modes. Thus, it
simportant to have alightweight pickup.
One can often judge the effect of adding the mass of a pickup by noting
he difference in behavior with the pickup fastened and with another mass
:qual to that of the pickup, in addition to the pickup. If the difference is
iegligible for these two conditions, the effect of the pickup is usually
inimportant. Under certain conditions near the resonant vibration frequency
4* the device under test, even asmall mass can shift the resonance enough to
.ffect the motion at the original resonant frequency by alarge amount.
When it is possible to change the excitation rate or frequency so that
esonance with the pickup in place is re-established, the behavior at the new
esonant point will often be sufficiently similar to the resonance behavior

212
 Effect of Background Noise. Ideally, when anoise source is measured, the
measurement should determine only the direct air-borne sound from the
source, without any appreciable contribution from noise produced by other
sources. In order to ensure isolation from other sources, the measurement
room may need to be isolated from external noise and vibration. As atest to
determine that this requirement has been met, the American National Stand-
ard Method for the Physical Measurement of Sound, S1.2, specifies the
following:
"If the increase in the sound pressure level in any given band, with the sound
source operating, compared to the ambient sound pressure level alone, is 10
dB or more, the sound pressure level due to both the sound source and
ambient sound is essentially the sound pressure level due to the sound source.
This is the preferred criterion."
If the background noise level and the apparatus noise level are steady, a
correction can be applied to the measured data according to the graph of
Figure 8-4. The procedure is as follows: After the test position has been
selected according to the test code and after exploration of the field as
outlined in paragraph 8.2.1, the background noise level is measured in the test
position. Then the sound level is measured with the apparatus operating. The
difference between the sound level with the apparatus operating and the
background level determines the correction to be used. If this difference is
lo

0o 2 3 4 5 6 7 8 9 10
dB DIFFERENCE BETWEEN TOTAL NOISE El BACKGROUND ALONE

Figure 8-4. Background noise correction for sound-level measurements.

less than 3 dB, the apparatus noise is less than the background noise; and the
level obtained by use of the correction should be regarded as only indicative
of the true level and not as an accurate measurement. If the difference is
greater than 10 dB, the background noise has virtually no effect, and the
reading with the apparatus operating is the desired level. An example of a
situation intermediate between these two is as follows: The background noise
level is 77.5 dB, and the total noise with the machine under test operating is
83.5 dB. The correction, from the graph of Figure 8-4, for a 6.0-dB differ-
ence, is 1.2 dB, so that the corrected level is 82.3 dB.
When apparatus noise is analyzed, the background noise level in each band
should also be analyzed to determine if the correction for the level in each

179
band is necessary and possible. The spectrum of the background noise is
usually different from that of the noise to be measured, and the corrections
in each band will be different.
lf this difference between background level and total noise level is small,
an attempt should be made to lower the background level. Usually the first
step is to work on the source or sources of this background noise to reduce
the noise directly. The second step is to work on the transmission path
between the source and the point of measurement. This step may mean
simply closing doors and windows, if the source is external to the room, or it
may mean erecting barriers, applying acoustical treatment to the room, and
opening doors and windows, if the source is in the room. The third step is to
improve the difference by the method of measurement. It may be possible to
select a point closer to the apparatus, or an exploration of the background
noise field may show that the measuring position can be shifted to a
minimum of this noise. The latter possibility is more likely when an analysis
is being made and the background level in aparticular band is unusually high.
It may also be possible to point the microphone at the apparatus to obtain an
improvement at high frequencies (see Figure 8-2); it may be necessary to use
adirectional microphone; or it may be desirable to use avibration pickup.

8.2.3 Instrument Precautions.

• Low Noise Levels — Effect of Circuit Noise. When low noise levels are to
be measured, the inherent circuit noise may contribute to the measured level.
This effect is usually noticeable in the range below 40 dB when a small
microphone is used or aceramic microphone is used on the end of avery long
cable. If the microphone is directly on the sound-level meter, the level at
which this effect may be important is below 30 dB, if the C weighting is used
or even lower if the A or B weighting is used. To measure the circuit noise the
microphone may be replaced by awell-shielded capacitor with acapacitance
equal to that of the microphone. A correction can then be made for this
noise, if necessary, by the same procedure as outlined for background noise in
paragraph 8.2.2. If the circuit noise is comparable to the noise being meas-
ured, some improvement in the measurement can usually be obtained by use
of an octave-band analyzer. The circuit noise in each band should be checked
also to see if correction is necessary.
Whenever low noise levels must be measured and extension cables are used,
the Type 1560-P42 or P40 Preamplifier should be used at the microphone.
Hum Pickup. When noise is measured near electrical equipment, acheck
should be made that there is no appreciable pickup of electro- magnetic field
in the sound-measuring system. The procedure depends on the directional
character of the field. The orientation of the instruments should be changed
to see if there is any significant change in level. If an analyzer is used, it
should be tuned to the power-supply frequency, usually 60 Hz, which would
be the 63-Hz band for the octave-band analyzer, when this test is made. If no
analyzer is included, the C-weighting should be used in this test to make the
effect of hum most noticeable, and a good-quality pair of earphones, with
tight-fitting ear cushions, should be used to listen to the output of the
sound-level meter.
If a dynamic microphone is used, tests should be made with different
orientations of the microphone, with the microphone disconnected, and with

180
1. F. Hopkins and N. R. Stryker ( 1948), " A Proposed Loudness- Efficiency Rating for
Loudspeakers and the Determination of System Power Requirements for Enclosures,"
Proc IRE., Vol 36, # 3, March, pp. 315-335.
t. T. Jenkins ( 1954), " A New Sound Integrator," Bell Labs Record, Vol. 32, September,
pp. 331-335.
/. 0. Knudsen and C. M. Harris ( 1950), Acoustical Designing in Architecture, John
Wiley & Sons, Inc., New York.
). Lubman ( 1968), " Fluctuations of Sound with Position in a Reverberant Room," J
Acoust Soc Am, Vol 44, #6, December, pp. 1491-1502.
Ogawa ( 1965), "The Applicable Limit of Beranek's Formula and Ishii's Formula of
Sound Pressure Level Distribution in a Room," J Acoust Soc Jap, Vol 21, # 3, May,
pp. 137-140.
/. M. A. Peutz ( 1968), " The Sound Energy Density Distribution in a Room," Reports
6th Intl Cong. on Acoustics, August, Paper E-5-2, pp. E-165-168.
1 K. Rasanen and B. M. Wigle ( 1967), "Accelerometer Mounting and Data Integrity,"
Sound and Vibration, Vol. 1, # 11, November, pp. 8-15.
L M. Robinson ( 1961), " Arrangement of 24 Points on a Sphere," Math Annalen, Vol
144, pp. 17-48.
'. K. Stein ( 1962), Measurement Engineering, Stein Engineering Services, Tempe, Ari-
zona.
L V. Waterhouse and R. K. Cook ( 1965), " Interference Patterns in Reverberant Sound
Fields II,"J Acoust Soc Am, Vol 37, # 3, March, pp. 424428.
L Waterhouse and D. Lubman ( 1970), " Discrete versus Continuous Space Averaging in a
Reverberant Sound Field,"J Acoust Soc Am, Vol. 48, # 1 ( Part 1), July, pp. 1-5.
t. J. Wells and F. M. Wiener ( 1961), " On the Determination of the Acoustic Power of a
Source of Sound in Semi-Reverberant Spaces," Noise Control, Vol. 7, #1, January-
February, pp. 2149.
Yamamoto ( 1961), " On Distribution of Sound Energy Along a Corridor," J Acoust
Soc Jap, Vol. 17, #4, December, pp. 286-292.
t. W. Young ( 1959), " Sabine Reverberation Equation and Sound Power Calculations,"J
Acoust Soc Am, Vol. 31, #7, July, pp. 912-921.
t. W. Young ( 1962), " Can Accurate Measurements be Made with a Sound- Level Meter
Held in Hand?" Sound, Vol. 1, # 1, January- February, pp. 17-24.

216
noted. These levels can be entered on the data sheet as, say, 85-91 dB or 88 ±
3dB.
When an average sound-pressure level is desired and the fluctuations are
less than 6 dB ; a simple average of the maximum and minimum levels is
usually taken. If the range of fluctuation is greater than 6 dB, the average
sound-pressure level is usually taken to be 3dB below the maximum level. In
selecting this maximum level, it is also customary to ignore any unusually
high levels that occur infrequently.
The " SLOW" meter speed should be used to obtain an average reading
when the fluctuations on the " FAST" position are more than 3or 4dB. On
steady sounds the reading of the meter will be the same for either the
"SLOW" or " FAST" position, while on fluctuating sounds the "SLOW"
position provides along-time average reading.
A more detailed discussion of this problem is given in succeeding para-
graphs.
Tones and Beats. The indicated sound level of aconstant-amplitude pure
tone is steady, and so is that of a mixture of tones, unless at least two
components are close together in frequency. Examples of sounds that have a
constant indicated sound level are transformer hum and noise from some
rotating electrical machinery. When the combined noise of several machines is
measured, the indicated level is also constant, unless the speed of the
machines is such that some of the major noise components are only afew
cycles apart in frequency. In this situation an audible beat, aperiodic rise and
fall in amplitude, occurs, and the indicated level also rises and falls.
Varying-Speed Sources. Machinery that operates at avarying speed usually
produces anoise that fluctuates in level. If the speed varies periodically, the
level will also vary periodically. This variation results because the noise pro-
duced by the machine varies with speed, because the response of the room in
which the measurement is made varies with frequency, and, if an analyzer is
used, because the response of the measurement system varies with frequency.
If the machine speed varies erratically, the noise level will also vary
erratically, and the behavior may be similar to that of random noise.
Random Noise. The indicated sound level of a random noise, such as that
produced by jets, blowers, combustion chambers, ventilating systems, etc., is
not steady. In fact, all sounds contain some random noise energy, and most
have enough so that the indicated level fluctuates noticeably. The extent of
the fluctuation is aclue to the nature of the sound.
The fluctuations in level are ordinarily not aresult of erratic behavior of
the measuring equipment, but rather reflect the irregularities in the process of
noise production. This process can often be considered as acombination of
many sources that produce sound at random time intervals. The measurement
of such noises can be treated on asimplified statistical basis that is satisfac-
tory for almost all sounds.
Average Energy Level of a Random Noise. When a random noise is
measured, the first important result that is desired is the long-time average
energy level. This concept leads to taking the average of the fluctuating
pointer reading. If the fluctuations are less than about 2dB, this average can
be easily and confidently estimated to afraction of adecibel. If the fluctua-
tions cover arange of 10 dB or more, the average is much less certain.
The extent of the meter fluctuation depends on the meter characteristic.
The slower the movement, the smaller are the fluctuations. Thus, if the

182
fluctuations exceed 3or 4 dB for the " FAST" meter position, the "SLOW"
meter position should be used.
When the fluctuations are large, the nature of the source or sources should
be considered. If the noise-generating mechanism shifts from one mode to
another, it may be desirable to characterize the noise level by more than one
average value. This choice is obvious for adishwasher, for example, where the
wash, rinse, and dry cycles differ significantly. But this choice may also be
useful in machines where the mode shifts more rapidly.
If the noise is random and the fluctuations are large only because the
effective noise bandwidth is small (see paragraph 5.6.5) the average value
should be obtained on an energy basis.
Calibration and Corrections. Satisfactory noise measurements depend on
the use of measuring equipment that is kept in proper operating condition.
Although the instruments are reliable and stable, in time the performance of
the instruments may change. In order to ensure that any important changes
will be discovered and corrected, certain simple checks have been provided
for the General Radio line of sound-level equipment, and these will be
discussed in this section. These checks can be made as routine maintenance
checks, and some of them should usually be made before and after any set of
noise measurements.
In addition to these routine checks, more complete calibration of the
system may be desirable for accurate measurements, particularly above 1000
Hz. These calibrations are also discussed in this section.
Electrical Circuit Calibration. The Type 1933 and 1551-C sound-level
meters and the analyzers have built-in calibration circuits for checking ampli-
fier gain. In each case the gain of the amplifier is compared with the
attenuation of astable, resistive attenuator.
This test does not check the sensitivity of the microphone and the
indicating instrument; these tests are discussed in the next section. The
indicating instrument is rugged and relatively unaffected by temperature
changes. Its temperature coefficient is about —0.02 dB/°F (- 0.036 dB/°C).
Acoustical Calibration at Preferred Frequencies. The Type 1562 Sound-
Level Calibrator provides an over-all system calibration at 125, 250, 500,
1000, and 2000 Hz. If arecord is kept of the calibration of amicrophone as a
function of frequency, any significant change in the relative calibration is
readily noticed. If such a change occurs, the microphone and the calibrator
should be checked as soon as possible. Here, it is useful to have more than
one microphone on hand, so that asecond microphone can be used if the first
is damaged; at the same time, a consistency check on its calibration can help
to ensure that the calibrator has not been damaged.
In the interests of maintaining accuracy in sound measurements, another
calibration service is provided for owners of General Radio instruments. If
these instruments are brought in to one of the General Radio offices, the level
will be checked by means of an acoustic calibrator. This calibration will
usually show if the instrument is operating correctly. If there is a serious
discrepancy, the situation will have to be handled as a regular service problem.
The calibrator can also be used to measure the microphone cable correc-
tion. The procedure is as follows:
a. After the noise measurement has been made, the calibrator is put on
the microphone, with the microphone at the end of the cable, and a
level reading is taken on the sound-level meter.

183
 b. The microphone is removed from the end of the cable and put directly
on the sound-level meter. The calibrator is put on the microphone at
the sound-level meter and asecond level reading is taken.
c. The difference between these two level readings is the cable correction.
For high accuracy, it is usually essential to have a calibration of the
microphone response characteristic as a function of frequency. When this
calibration is available and an analysis of anoise is made, correction can be
made for the microphone f requency-response characteristic. This correction

can be applied only if the noise is analyzed or if the sound is dominated by a

component of known frequency, as, for example, in the measurement of
loudspeaker response. Otherwise, one must check the uniformity of response
of asystem to be sure that the measured level of anoise is correct.
Calibration At High Frequencies. The accurate calibration of amicrophone
at high frequencies in terms of sensitivity vs frequency requires elaborate
facilities. Only a few laboratories (e.g., The National Bureau of Standards)
offer such calibration as a regular service. General Radio Company will
calibrate response vs frequency only for those microphones that it supplies.
Such calibration is supplied with all microphones manufactured by General

Radio Co.
At General Radio, afree-field perpendicular-incidence calibration is made
by comparison with laboratory standard condenser microphones ( ANSI
S1.12-1967 Specifications for Laboratory Standard Microphones), according
to the methods given in S1.10-1966, Calibration of Microphones. The work-
ing-standard microphones are periodically compared with acondenser micro-
phone that has been calibrated at the National Bureau of Standards. They are
also periodically calibrated on an absolute basis by the reciprocity method.
Since the sound-level meter standard is based on a random-incidence
specification, the perpendicular incidence calibration is automatically con-
verted to the random incidence calibration in the plotting procedure, so that
calibrations supplied by General Radio are for random-incidence response
and data for converting to perpendicular-incidence or grazing-incidence cali-

brations are included.

Correction For Frequency- Response Characteristic. It is customary to set
the calibration of an acoustical measurement system to indicate the correct
level at 400 or 500 Hz. At other frequencies, the differences between true
and indicated levels, as determined from acalibration curve, can be applied as
corrections to the results of anoise measurement. At frequencies above 1000
Hz, the directional characteristic of the microphone should be taken into
account, and the particular curve that corresponds to the actual angle of
incidence used should be used for the corrections.
Comparison Tests Among Different Sound- Level Meters. When measure-
ments are made on the same noise with two different sound-level meters, it is
commonly found that the readings differ by a significant amount. The
preceding material in this chapter should indicate most of the possible sources
of discrepancy between the two. Differences in the microphone characteris -
tics are usually the chief cause of this discrepancy. For example, if one
sound-level meter uses a dynamic microphone and the other uses aceramic
microphone, and if the noise contains strong-low frequency components,
large differences can occur because of the generally poorer low-frequency
response of the dynamic microphone. When these effects are understood,
most of the discrepancies are readily explained.

184
 Table 8 1
ANSI* SLM CALIBRATION TOLERANCE— C- WEIGHTING

Tolerance —dB
Frequency Type I Type II Type Ill
Hz Precision General Purpose Survey

10 t2.5
12.5 t2
16 t2
20 ±2 +3,—...
25 t1.5 +2,-2.5 +3,-4.5
31.5 t1.5 +1.5,-2 +2.5,-3
40 t1 +1,-1.5 +2,-2.5
50 to 630 tl t1 *2
800 t1 t1 t2.5
1000 ±1 t1.5 t2.5
1250 t1 ±1.5 t2.5
1600 t1 ±2 ±3
2000 t1 t2.5 t3.5
2500 tl +3.5,-3.0 t4
3150 t1 +4.5,-3.5 t4.5
4000 t1 +5,-4 ±5
5000 +1.5,-2 +5.5,-4.5 ±6
6300 +1.5,-2 +6,-5 t7
8000 +1.5,-3 ±6 ±7
10000 +2,-4 +6,— co +7,—.
12500 +3,-6
16000 +3,—.

*ANSI s1.4-1971.
Another factor that can contribute to this discrepancy concerns the
average level. For purposes of meeting certain tolerances, the average level of
an instrument made by one manufacturer may be set slightly differently from
that made by another.
If the instruments are not operating properly, or if standing waves are not
averaged out, serious discrepancies can, of course, be expected.
In order to set an upper limit to these differences among sound-level
meters, the American National Standard Specifications for Sound-Level Me-
ters, S1.4-1971, sets certain tolerances on the prescribed frequency character-
istics. Representative values for C weighting are shown in Table 8-1.
Incidentally, when a sound-level meter is calibrated with a 1562 Sound-
Level Calibrator, these tolerances should be considered in comparing the
actual sound-level meter reading with the expected calibration level.
Effect of C Weighting on Band Levels. In principle, the response of the
equipment supplying an analyzer should be as uniform as possible to obtain
true pressure levels. Sometimes the "C" weighting is used for octave-band
measurements. If this is done for instruments meeting the latest ANSI and
International standards, there will be small differences in level at the low- and
high-frequency ends, compared with the levels that would be obtained with a
more nearly uniform response, because of the specified roll-offs in response
for the C weighting. Thus, the C-weighted octave-band level is less, by about 3
dB for the bands centered at 31.5 and 8000 Hz, and by 0.8 dB for the bands
at 63 and 4000 Hz, than with the uniform response weighting (flat). The
shifts in level for the bands in between are too small to be significant.
Check On Over-All Level. When an octave-band analysis has been made, it
is good practice to check that the sum of the individual band levels on an

185
 The results shown in Figure 9-1c illustrate that abarrier is mainly effective
at high frequencies, and there it produces only amoderate reduction in noise
level.
The novice in this field sometimes assumes that the materials used for
sound absorption can also be used alone for sound isolation. If we build an
enclosure solely of these materials mounted on alight framework, we would
typically find the result shown in Figure 9-1d. Only at high frequencies do we
have anoticeable reduction in level, and even there it is asmall reduction.
A more satisfactory enclosure is built of more massive and rigid construc-
tional materials. Assume that we enclose the machine by awell-sealed, heavy,
plasterboard structure. Then we might observe the result shown in Figure
9-2a. Here an appreciable reduction is obtained over the middle- and high-fre-
quency range. The enclosure is not as effective as it might be, however,
because two important factors limit the reduction obtained. First, the vibra-
tion of the machine is carried by the supports to the floor and then to the
whole enclosure. This vibration then may result in appreciable noise radia-
tion. Second, the side walls of the enclosure absorb only asmall percentage of
the sound energy.
The addition of a suitable vibration isolation mounting will reduce the
noise transmitted by solid-borne vibration. This effect is illustrated in Figure
9-2b. Here we see anoticeable improvement over most of the audio spectrum.
When the sound absorption within an enclosure is small, the noise energy
from the machine produces a high level within the enclosure. Then the
attenuation of the enclosure operates from this initial high level. The level
within the enclosure can usually be reduced by the addition of some sound-
absorbing material within the enclosure, with the result that the level outside
the enclosure is also reduced. This effect is shown in Figure 9-2c, which
should be compared with Figure 9-2b.
If even more noise reduction is required than that obtained by the one
enclosure, asecond, lined, well-sealed enclosure can be built around the first.
The first enclosure is supported within the second on soft vibration mounts.
Then a noise reduction of the magnitude shown in Figure 9-2d can be
obtained.

9.4 SUMMARY OF NOISE REDUCTION PROCEDURES.

The approach to anoise reduction problem can be summed up as follows:
•1. Consider the source.
( Can aquieter machine by substituted?
Can the noise energy be reduced?
Can auseful change be made in the directivity pattern?
Are resilient mounts of any use here?
Can amuffler be used?
2. Consider the path from the source to the listener.
Can the source or the listener be readily moved?
Is acoustic treatment auseful solution?
Should barriers be erected?
Is atotal enclosure required?

9.5 VIBRATION REDUCTION.

The basic procedure for vibration reduction will be described briefly.
Many specialized techniques have been developed also, and a complete
summary of these is impractical here. More extensive information on vibra-

222
of use cannot be duplicated, the noise measurements may not be sufficient to
predict the expected behavior, because of the difference in transmission of
noise energy through the supports. The usual alternative is to use a very
resilient mounting, so that the transmission of energy to the support is
negligible.

8.2.5 Effects of Atmospheric Pressure. Some acoustical measurements are

affected by atmospheric pressure and temperature. The output of the 1562
Calibrator, for instance, varies somewhat with pressure, and the rated refer-
ence level occurs at astandard atmospheric pressure of 1013 millibars. If the
pressure when the calibrator is used is significantly different from 1013
millibars, a correction should be made. The altitude where the calibrator is
used is usually the most significant factor in determining the average atmos-
pheric pressure, and a chart for correcting for this effect is included in the
calibrator instruction book. Since the output of most other sound sources is
affected by the pressure, achart relating height to average pressure is included
here ( Figure 8-5). The actual variation in output with pressure for practical
sources is usually between that shown by the corresponding decibel scale on
the right and one-half that value. Thus, for altitudes up to 2km ( 6560 ft), the
change in output with altitude is generally less than 2dB.
The variation of atmospheric pressure at agiven location from day to day
is usually less important, but for careful measurements where fractions of a
decibel are being considered, the actual atmospheric pressure should be

§
HEIGHT IN FEET
o ,0
o
o 5
CS' o- o,.
0 _ c4 W)
1200 T
i i

1000 -o

800 -2
PRESSURE IN MILLIBARS

-4 ,,,

_
_1
600 1.1.1
0

--6 5
w
o

-
400 -8
-

i\
HO
300

--t-I2
250 —
0 2 4 6 8 10
HEIGHT IN KILOMETERS
ABOVE SEA LEVEL

INCHES OF Hg
10 15 20 25 30
1, 1 11 ,1 I , I t
i I t ill 1 r l I
T
I I i i
200 300 400 500 600 700 800 900 I000 1100
MILLIBARS

Figure 8-5. Corrections for sound pressure level

for atmospheric pressure at various altitudes.

187
 noted. The pressure can be obtained from the local weather bureau, and a
correction for the difference in altitude between the point where the acous-
tical measurements are made and the weather bureau may be necessary. This
correction is readily estimated from the chart.

8.3 SOUND FIELDS.

The behavior of sound in various environments has been described briefly
at various places in this book. We shall now discuss it in more detail in order
to explain how sound-power measurements are made.
We shall begin with a discussion of a simple source under idealized
conditions. Then we shall point out various factors that alter the idealized
conditions and discuss in general what the important effects are.

8.3.1 Simple Source in Free Field.

Point Source. Any vibrating object will radiate sound into the air. The
amount of sound radiated depends on the amplitude of vibration of each
vibrating part, the area of each part, and the time pattern of the vibrations,
including the relative time pattern compared with that of the other parts.
The simplest form of source is a sphere that vibrates uniformly over its
entire surface. We can think of this source as around balloon with air in it.
We periodically pump some more air into it and then let the same amount of
air out. If the surface of the balloon then expanded and contracted uniform-
ly, the balloon would be asimple, spherical source. This source radiates sound
equally in all directions from an apparent center, which is the center of the
balloon. It then is a " point" source, insofar as sound radiation is concerned.
Free Field. If such a point (or spherical) source is in the air far from any
other objects, including the ground, the sound pressure produced by the
source is the same in every direction at equal distances from the point source.
Furthermore, the sound pressure is halved for each doubling of distance from
the point. This change is usually expressed as a decrease in sound-pressure
level of 6 dB. The sound field produced under these idealized conditions is
called afree sound field or, simply afree field because it is uniform, it is free
from all bounding surfaces, and it is undisturbed by other sources of sound.
Power Level in Free Field. Under free-field conditions, asingle measure-
ment* of the sound-pressure level at aknown distance from apoint source is
enough to tell us all about the sound field radiated by the source. For
example, we can then predict the level at any other point, since the sound
pressure varies inversely as the distance from the source. We can also com-
pute the total sound power radiated by the point source. This calculation is
usually made in terms of the power level re 10 12 watt (PWL) of the source
(paragraph 2.3). Then the required relation to the sound-pressure level ( SPL)
is:
PWL = SPL + 20 log r + 0.5 dB
where ris the distance in feet from the point source to the point where the
sound-pressure level is measured. This relation is correct for apoint source in
a free field at normal room temperature and barometric pressure, that is,
20°C and 1013 millibars. At other temperatures and pressures, the correction
shown in the graph of Figure 8-6 applies. This correction is usually unimpor-
tant.

*The concept of apoint source is an idealized one. It is not reasonable to assume that an
actual source is a true point source, so that one should never be content with a single
measurement.

188
 /013 rniiiibes
o
0

800 ,

• `" rnb
6
l
-3 5°0 rnb

I
-50 o 50 100 150
TEMPERATURE- C
-51
0 100 ISO 200 250 300
TEMPERATURE- F
Figure 8-6. Corrections for temperature and barometric pressure to be applied when the
equations relating power level (PWL) and sound-pressure level (SPL) are used. The
correction is to be added to, if positive, or subtracted from, if negative, the sound-pres-
sure level computed by the equation from the power level. If the power level is to be
computed from agiven sound-pressure level, the correction should be subtracted from, if
positive, or added to, if negative, the given sound-pressure level before the numerical
value is substituted in the equation.

As an example, suppose that we measure asound-pressure level of 73.5 dB

re 20 µN/m 2 at adistance of 20 ft from apoint source. Then

PWL = 73.5 + 20 log 20 + 0.5 = 100 dB re lo' 2 W.

The value for 20 log rcan be found from atable of logarithms or from the
decibel tables in the Appendix, where the columns labeled as pressure ratios
should be used for this distance.
The power level can be converted to actual acoustic power in watts as
explained in paragraph 2.3. For the example above, the 100 dB corresponds
to an acoustic power of . 01 W.
We can also use the equation to predict sound-pressure levels at any
distance in the free field if we know the acoustic power radiated. Thus, this
point source radiating . 01 W, corresponding to a power level of 100 dB re
10 -12 ,produces asound-pressure level of 100 —20.5 = 79.5 dB re 20121•1/m 2
at 10 ft from the source.

8.3.2 Directional Source in Free Field.

Directional Source. In actual practice, noise sources are not as simple as
point sources. The sound is not radiated uniformly in all directions, either
because the shape of the sound source is not spherical, or because the
amplitude and time phase of the vibrations of the different parts are not
uniform, or both. The net result is that more sound is radiated in some
directions than in others.
Sound-Pressure Contours. In other words, the sound-pressure level for a
given distance is different in different directions. As an example, let us
observe the sound field surrounding alarge 60-cycle power-distribution trans-
former, as shown in Figure 8-7. The contours around the transformer corres-
pond to the indicated values of sound-pressure level. This source is obviously
directional, since the contours are not circular.

189
 58db

lo ,

Figure 8-7. Simplified contours of

equal sound-pressure level around
alarge power- distribution
transformer.

58db

When such a directional sound source is far from any other objects,
however, it behaves in some ways like a point source. For example, the
sound-pressure level decreases 6 dB for each doubling of distance, provided
we start our measurements at a distance away from the source that is several
times the largest dimension of the source, and provided we move directly
away from the sburce. From the example of the transformer in Figure 8-7, we
see that, at distances greater than several times the length of the transformer,
the contours are similar in shape and the levels decrease approximately 6dB
for each doubling of distance. In actual practice this idealized behavior is
upset by the effects of variation in terrain, atmospheric conditions, and the
interference of nearby objects.
Near Field and Far Field. We can also see that at locations close to the
transformer the sound-level contours are different in shape from those at a
distance. Furthermore, there is no apparent center from which one finds the
6-dB drop for each doubling of distance. Consequently, this "near-field"
behavior cannot readily be used to predict the behavior at adistance. The
differences between the near field and "far field" can be described in part as
follows: Assume we have asource in which one part moves outwardly while
another moves inwardly and vice versa. The air pushed away by one part will
then tend to move over to compensate for the decrease in air pressure at the
inward moving part. If the air can move over quickly enough, there will be
considerable motion of air between the two parts, without contributing much
to radiation of sound away from the source. The time factor in this motion of
air can be expressed as arelation between the distance to be covered and the
wavelength of the sound in air. The wavelength, X, at normal temperature is as
follows:
1130 344
X 'It, ft --- —i---
f meters
f —
where fis the frequency in hertz and 1130 f/s is the speed of sound. Then, in

190
order that the near field effect should not be very important, one should be
at least one wavelength away from the source. This dimension should be
determined on the basis of the lowest frequency of interest. For the example
of the 60-Hz transformer, the lowest frequency of sound is 120 Hz, which
corresponds to awavelength of about 10 ft.
Another factor that enters into the differences between the near field and
far field behavior is the way the sound waves spread out from asource. The
sound waves from a large source vary with distance differently from waves
produced by a small source. But at adistance of several ( 3to 4) times the
largest dimension of the radiating source, "spherical spreading" is said to
exist, and the behavior is then nearly independent of the size of the source.
• Directivity Factor. When we are interested in sound-pressure levels beyond
the immediate vicinity of the source, any sound can be treated as apoint
source, provided we introduce a directivity factor. This factor takes into
account the variation in sound-pressure level with direction to the source.
This directivity factor, which is a function of direction and frequency, is
usually labeled Q. It can be expressed as the ratio of two acoustic powers.
One of these powers is that which would be radiated by apoint source, in
order to produce the observed sound-pressure level in the specified direction.
The other power is the total acoustic power radiated by the actual source.
• Sound-Pressure Level for a Directional Source. When we know this direc-
tivity factor for the direction of interest, we can use it, in the earlier equation
for apoint source, as a multiplying factor on the power. Expressed in terms
of level the new equation is as follows:

SPL = PWL + 10 log Q — 20 log r — 0.5 dB.

This equation relates the power level of the source, the sound-pressure level in
a given direction at a distance rfeet from the source, and the directivity
factor for that direction. (This equation is also subject to the minor correc-
tions for temperature and pressure shown in Figure 8-6).
For example, let us assume that an auto horn whose measured power level
is 104 dB is sounded. We are interested in the sound-pressure level at a
distance of 20 ft in the horizontal plane of the horn, but at an angle of 20°
from the principle axis of the horn. Along this direction of 20° from the axis
the directivity factor is 5, say. Then we have

SPL .-.- 104 + 10 log 5 — 20 log 20 — 0.5 = 84.5 dB

at 20 ft in the required direction.

8.3.3 Simulated Free Field. The free-field condition does not occur in
practice, because of the effects of sound reflected from the ground or floor,
from nearby objects, and from walls and ceiling. The result of these reflec-
tions is that the sound-pressure level measured at adistance from the source is
different from that predicted by the free-field equations. The reflections can
be reduced by acoustic absorbing materials applied to the reflecting surfaces.
By the proper design and application of this treatment, one can produce in a
room alimited space having the essential characteristics of afree field over a
wide frequency range. Many such rooms, called "anechoic" or "free-field"
rooms, have been built and are described in the literature. When accurate
measurements of the radiated sound power and directivity are required, the
measurements should be made in such an environment.

191
 8.3.4 Effect of Reflections in a Room. The sound that a noise source
radiates in aroom is reflected by the walls, floor, and ceiling. The reflected
sound will again be reflected when it strikes another boundary, with some
absorption of energy at each reflection. The net result is that the intensity of
the sound is different from what it would be if these reflecting surfaces were
not there.
Close to the source of sound there is little effect from these reflections,
since the direct sound dominates. But far from the source, unless the
boundaries are very absorbing, the reflected sound dominates, and this region
is called the reverberant field. The sound-pressure level in this region depends
on the acoustic power radiated, the size of the room, and the acoustic
absorption characteristics of the materials in the room. These factors and the
directivity characteristics of the source also determine the region over which
the transition between reverberant and direct sound occurs.
A second effect of reflected sound is that measured sound does not
necessarily decrease steadily as the measuring position is moved away from
the source. At certain frequencies in aroom with hard walls, marked patterns
of variations of sound pressure with position can be observed. Variations of
up to 10 dB are common and, in particular situations, much more can be
found. These variations are usually of the following form: As the measuring
microphone is moved away from the source, the measured sound pressure
decreases to a minimum, rises again to amaximum, decreases to aminimum
again, etc. These patterns are called standing waves. They are noticeable
mainly when the sound source has strong frequency components in the
vicinity of one of the very many possible resonances of the room. They also
are more likely to be observed when afrequency analysis is made; and the
narrower the bandwidth of the analyzer, the more marked these variations
will be.
In aroom, the spacing from one minimum in sound pressure to another is
on the average greater than one-half wavelength.
Reverberation Room. If aroom has very little sound absorption, the room
is said to be "live" or reverberant. Sound from asource in such aroom will be
reflected many times, as it bounces back and forth on the surfaces of the
room. At any one point in this room the sound will have arrived there from
many directions because of the many reflections. If the room dimensions are
properly proportioned and certain other design features are included, the
flow of sound energy in all directions can be made nearly equally probable,
and the field is then said to be diffuse. This type of room is called a
reverberation room, and it is widely used for the measurement of the
sound-absorption of materials, as well as for sound-power measurements,
when the directivity characteristics are not required.
Ordinary rooms. The sound field in an ordinary room cannot be described
in detail. The acoustical boundary conditions of ordinary rooms are extraor-
dinarily complicated, and most sound sources are also complicated. The result
of this complexity is that one can attempt only an average-type of descrip-
tion. Even arough approximation can be useful, however, and we shall review
briefly some of the work on room characteristics as it applies to the sound
produced by asource in aroom.
. Room Constant. In order to simplify the analysis of the effect of the
room, it is assumed that enough measurements are made so that any stand-
ing-wave patterns can be averaged out. A number of other assumptions are

192
 Supports should be located to avoid cross coupling from one mode of
vibration to another. Such a requirement ordinarily means that the line of
action of the support should pass through the center of gravity of the device
being supported,'
The foundation, the isolating suspension system, and the supported struc-
ture will have, individually and in combination, resonant modes at frequen-
cies higher than the first natural resonance. Sometimes these higher modes
cause trouble, because the isolation is reduced from that normally expected
(Plunkett, 1958).
l'he usual commercial vibration isolators include sufficient damping so
that effects of the higher-order resonances in the isolator are not seriou-e But
the isolation is usually significantly less at high frequencies than one would
expect on the basis of the simple idea of aweight supported on aspring.
Torsional vibration is isolated by the use of flexible couplings, flexible
shafts, and belts. These, too, include some damping, and they also introduce
resonant modes of torsional vibration in conjunction with the rotational
inertia of the coupled system.
Multiple isolators need careful design in order to be effective. When two
isolator units are used in cascade, serious effects that interfere with satisfac-
tory isolation may occur ( Skudrzyk, 1959).

9.5.5 Maintenance. When maintenance of proper performance or accept-

able noise and vibration levels is the goal, symptoms are used as aguide to
discover the source of any trouble that may develop and to decide on the
remedy. Before these symptoms are reviewed, it is also helpful to keep in
mind the many ways that machine performance is affected by changes that
occur with time. A systematic classification of the sources of these changes
should serve to point up the many possibilities that exist. They are:
1. Wear
2. Erosion

3. Corrosion
4. Aging
Curing
Crystalization and fatigue
Solidifying of grease or packing
Loss of adhesion or bonding
5. Inelastic behavior
Parts stressed out of shape
Bent parts
Increased tolerances
6. Loosening of fastenings

7. Broken or damaged parts

8. Incorrect or inadequate lubrication
9. Foreign matter
Dirt chips, dust, grit
Contaminants
Humidity
Ice accumulation
Paint and other finishes

229
varies approximately as the volume of the room, the effect in small, treated
rooms is usually negligible.
The relation given above is shown graphically in Figures 8-8 and 8-9, where
Figure 8-8 applies to the nondirectional or simple source or to adirectional
source in the direction having adirectivity factor of 1 ( Q=1), and Figure 8-9
applies to the directions having the labeled values of directivity factor.

METERS
•5 0.2 0.5 2 5 10 zo
1 , 1 I I 1i

R•50 ft 2

4.6
1Ó00 2
93 ,,,2

200f 1
2

OPEN AIR 19.6m 2

Rrm 1
5000 2
I
116ne
1000ft 2
93m 2
2000ft 2

166rn 2

50001, 2

465, 2
10.0001, 2
930. 2
0,0001, 2

1860. 2

I
0.5 07 2 3 4 5 7 10 20 30 40 50 70 100
DISTANCE FROM ACOUSTIC CENTER
OF A NONDIRECTIONAL SOURCE IN FEET • r

Figure 8-8. Chart shoWing the sound-pressure level, SPL, relative to the power level,
PWL, for a nondirectional source for different values of the room constant, R, as a
function of the distance from the source.

• Reverberant Field. The graphs of Figure 8-8 and 8-9 show that close to the
source the sound-pressure level tends to vary with the distance from the
source as it does under free-field conditions ( R = 00). But far from the source
the sound-pressure level becomes independent of the directivity and the
distance to the source. This region is called the reverberant field. Here, the
level is determined by the acoustic power radiated by the source and the
acoustic characteristics of the room. The region over which the transition
between the free-field behavior and the reverberant field occurs is determined
by the directivity factor and the room constant.
Actual Room Behavior. In a well-designed reverberation room, the be-
havior on the average is similar to that shown in the figures. Most other rooms
have characteristics that on the average fall between that reverberant behavior
and the free-field sound-pressure level decrease of 6 dB for each doubling of
the distance ( Ogawa, 1965; Gober and Lubcke, 1966; Yamamoto, 1961,
Peutz, 1968).
In flat rooms (i.e., rooms whose ceilings are low relative to room length
and width), the sound-pressure level at adistance from the source tends to
decrease a fixed amount, but less than 6 dB, for each doubling of distance.
The decrease depends on the sound absorption in the room. In very long
rooms or halls, the sound-pressure level tends to decrease afixed number of

194
 METERS
0.5 2 5 20
+10 I 1 I I I

0.8

R•50ft 2

4.
6m2

R.5000 2
À
46m 2

0 .1/
4 / 1 2 4 8
0,000f1 2
30m 2

-25
05 07 2 3 4 5 7 10 20 30 40 50 70 100
DISTANCE FROM ACOUSTIC CENTER
OF A SOURCE IN FEET • r

Figure 8-9. Chart showing the sound-pressure level, SPL, relative to the power level,
PWL, for a directional source as a function of the distance from the source. The relation
is shown for three different values of the room constant, R, and for six different values
of the directivity factor Q.

decibels for aconstant increment in the distance from the source.

In the vicinity of local obstructions and other areas with marked changes
in acoustic characteristics, the sound-pressure level can change markedly.
In order to illustrate one of these effects, we shall reproduce a set of
measurements made on a rectangular studio with an acoustic noise source
covering the range from 300 to 600 Hz ( Ogawa, 1965). The studio was 13.5
m by 24 m by 4.2 m and the average absorption coefficient at 500 Hz was
0.25. The room constant is then about 320 m2 or about 3400 ft 2 .Figure
8-10 shows the results of measurements in this room. The crosses are the
observed sound-pressure levels and the smooth curve is plotted to correspond
with the room-constant.
The calculations from the simple formula tend to overestimate the level at
aconsiderable distance from the source. The average trend of the sound-pres-
sure level is a drop of about 3.5 dB for a doubling of the distance. The
marked departure from the behavior expected, on the basis of the simple
formula, is aresult of the fact that the average absorption was relatively high
and the height was relatively low compared to the other dimensions, but it is
similar in shape to that of many large, general, office areas.
Much more complicated formulas can yield values in closer agreement with
the measurements, but the calculations become impractically tedious unless
they are programmed on a computer. The simple formula is still useful for a
preliminary estimate of the expected behavior, particularly if the absorption
is small and if no one room dimension is markedly different from the others.

8.3.5 Measurement of Acoustic Power. A noise rating is often intended to

make possible the prediction of the noise level that the apparatus will

195
 O
\

\
2 \ 4

\
3
\F
4 1 \ +
+
5 1
SPL- dB

6 + ‘•,...1

7 4.
.41.
RELATIVE

ROOM
.......
e I + 4 I.
.....
..... ......
CONSTANT
...,
+
9 11- 4
+ +
10 +1 4 1
+ +- +
Il 11 1 1
4 +
12 1 + +
+++ 44
13 1H + 11 I
+++ 4 +
14 -F + + 1 1 1 1 1
+ +
-15
2 3 4 5 7 10 20
DISTANCE FROM SOURCE - METERS

Figure 8-10. Plot of actual measurements made in aroom with an absorption coefficient
of 0.25 vs calculated performance (dashed line). (Ogawa, 1965.)

produce when installed. In order for the rating to be adequate for this
purpose, the total acoustic power radiated by the source and the acoustic
directivity pattern of the source should be included as part of the rating. We
shall explain in this section how the power and directivity can be determined,
but first we shall discuss the limitations of the usual method of noise rating.
For example, an air compressor may be rated by the manufacturer as
producing a noise level of 85 dB at adistance of 5 ft. This level may have
been calculated by an averaging of afew sound-level readings 5 ft from the
compressor. When it is installed and the level is measured, the new level may
be, say, 90 dB at 5ft. Naturally, the purchaser feels that he should complain
because the machine was incorrectly rated; perhaps he returns the compres-
sor, or he decides that he can no longer trust the manufacturer. Actually, the
manufacturer may have been entirely correct in his noise measurements, but
the rating was inadequate. The difference of 5 dB may have been caused by
incorrect installation, but usually such adifference is aresult of the acoustical
characteristics of the factory space. By the use of an adequate rating system
and a knowledge of acoustical room characteristics, it would have been
possible to predict this effect.
Another part of this problem is the prediction of levels at places in the
factory other than at the measurement distance. For example, the nearest
worker may be 20 ft away, and the level at adistance of 20 ft is then more
important than at 5 ft. Again, a knowledge of the acoustic power radiated
and the acoustical characteristics of the factory space will be needed to
predict the probable level at this distance.
The procedure suggested here for determining the power and directivity is
based on measurements of the sound-pressure level at a number of points
around the noise source. The measurement of sound-pressure level has already
been discussed. We shall discuss here the selection of the points at which to

196
a.

1
00 110

90 100

5 90

:<> 60

50 '

-T— ,*. ---- r -

lb 'do •
7,9*;
lO Kio idoo 000
FREQUENCY IN Hz FREQUENCY IN Hz

C.

WTT_TIU; "MT

e
î

5 80

'
0 70

Figure 10-1. Noise reduction possible

through careful acoustical analysis
followed by corrective action.0

.= 0

H i 11 I
•7- -errr .... -
io loo
FREQUENCY IN Hz

Then some exploration of the office floor for a 37-Hz component indicated
nothing significant. An analysis of the vibration of the pump structure
produced the result shown in Figure 10-1b. These measurements satisfied the
one engineer that his explanation based on vibration was probably incorrect.
Since the 37-Hz component was so dominant, it could be assessed as a
pure-tone. The equal-loudness contours of Figure 3-3 indicated that its
loudness level was about 70 phons. Although the rating procedures of
Chapter 3 apply generally to broad-band noise, it is obvious from those
ratings that a significant drop in level for this very loud tone would be
necessary in order to make it acceptable. If it could be lowered to 40 phons,
it would probably be unobjectionable. This drop would require adecrease in
sound-pressure level to 70 dB.
Many possibilities for correcting the annoying condition were considered,
for example, changes in the offices so that they would not resonate at the
troublesome frequency, a change in compressor speed, and rerouting of the
air intake. The solution adopted was to add a pneumatic filter to the air

234
 exploration is to select a number of points at which measurements will be
made. Areas on the sphere are then associated with these points. These areas
have the measurement points as their centers, and the extent of each area is
determined by the nearness of the other measuring points. In the process of
making the basic measurements, the microphone should be moved around to
determine the variation in sound-pressure level within each area. If the
variations in sound-pressure level within any one area are greater than 2dB, it
is advisable to select additional measuring points in that area. However, if no
attempt is being made to obtain an accurate picture of the directivity pattern,
the extent of the variation can be noted. Then, provided the variation is less
than 6 dB, the average level can be used as arepresentative value for the area.
Uniformly Distributed Measuring Points. The calculations for the radiated
power are simplified if the measuring points are uniformly distributed on the
surface of the sphere. Because of the nature of the geometric pattern, only six
such sets of points are possible. These six sets have 2, 4, 6, 8, 12, and 20
uniformly distributed points. The locations for the sets of 8, 12, and 20
Points are shown in Figures 8-11 through 8-13. These are now generally used,
although a different orientation with respect to the ground plane may be
found desirable for some particular applications. The areas associated with
the sets of 8, 12, and 20 points are regular spherical triangles, regular
spherical pentagons, and regular spherical triangles, respectively.

+ ( 0,0.8165, 10.5774)

o
(
- 0. 8165, 0, ± 0.5774)

10.8(65, 0, .± 0 5774)

+ ( 0, - 0.8(65, 1:0.5774)

Figure 8-11. Plan view of eight points uniformly distributed on asphere of unit radius.
Coordinates are given in terms of distances from center along three mutually perpendicu-
lar axes (x, y, z). The "i" refers to the existence of two points, one above the x- y
reference
only the foplane andabove
ur points one bthe
elow. When
plane used.urements are to be made on a hem i
are meas sphere

Other sets of points that may be useful are those that correspond to the
vertices of an Archimedean serniregular polyhedron. The most interesting of
these have 24 (R. M. Robinson, 1961), 48, and 60 points. Although these
points are
points.
because th
notstribution
e di
uniformly distributed, they are all of equal importance,
of points

around any one po i

nt is the same for all

# Hemispherical Measurements. When the device to be tested is normally

mounted on a concrete foundation, on the ground, or on awall, it is often
desirable to test it while it is so mounted. The environment that approximates

198
heavily treated with acoustical tile. In addition, to avoid troubles from open
windows in the summer bypassing the wall barriers, the room was air-condi-
tioned. The air conditioning made it unnecessary to open the windows, and it
made it possible to wear the earmuffs with comfort even in the summer.

10.3 REDUCTION OF FLUTTER IN A TAPE RECORDER.

Measurements of the flutter (variations in tape speed) of atape recorder
showed strong components at 15, 30, and 75 Hz and other minor compo-
nents. Although it was expected that eccentricities in the drive produced
some of this flutter, acheck on the effects of vibration was made.
A Type 1560-P54 Vibration Pickup was mounted on the tape deck and
connected through a Type 1560-P40 Preamplifier to aType 1564 Sound and
Vibration Analyzer. The analyzed vibration showed strong components at 30
and 75 Hz, but very little at 15 Hz. The 30 Hz component corresponded to
the motor speed. Although the motor and attached flywheel had been
balanced before being mounted, the vibration was easily reduced by rebalanc-
ing of these in place. After balancing, the flutter component at 30 Hz was
negligible, and the 75 Hz component was markedly reduced also.
The 15-Hz component corresponded to the capstan speed, and it was the
largest remaining component. The fact that there was no appreciable vibra-
tion at this frequency seemed surprising at first. The flutter could be reduced
significantly by the placing of eccentric weights on the capstan flywheel; but
then the vibration of the tape deck increased markedly at 15 Hz. Obviously,
this flutter component was caused by eccentricity in the capstan, and the
vibration introduced was canceling the effects of eccentricity.
Measurements at the capstan bearing showed that the vibration at 75 Hz
was a maximum there. What was happening here, apparently, was that the
capstan flywheel and shaft structure had avibration resonance at 75 Hz that
was excited either by the fifth harmonic of the capstan rotation frequency or,
more likely, by a combination of amultiple of the motor rotation frequency
and the capstan rotation frequency. Because of the resonance, eery little
energy was required to produce asignificant vibration. This mode could be
reduced significantly by a change in the resonant frequency, but the balanc-
ing of the motor and flywheel had already reduced it so that the 15-Hz
component was the only significant one remaining.

10.4 AN OIL- PUMP PROBLEM.

An oil pump, used in aproduction setup to supply oil at high pressure to a
number of hydraulic presses, was so noisy that the workmen objected to
using it. This pump had been installed to speed up production with new
presses, but the men preferred to use an earlier production method because it
was not then necessary to use the noisy pump. The problem was to find out
what should be done to make the noise less objectionable.
In this example, it was assumed that the pump itself could not be modified
to reduce the noise, since correcting basic design faults would be a major
problem. Errors in alignment or looseness of mounting, as the source of the
high noise levels, however, should be taken into consideration. On that basis,
the apparent procedure was to investigate these possibilities, to measure the
noise produced by the machine, to measure the background noise level, and
then to decide what recommendations should be made.

236
in also be used, but then 4 of the set must be weighted by a factor of
ne-
half (or, 3 dB is subtracted from the levels at these four points). See
igure 8-13.
The reflecting plane affects the sound-radiation characteristics of the
urce (ANSI S1.2-1962; Baade, 1964; Ellison et al, 1969). The near-field
Dnditions are extended and the directivity pattern is more complex, corn-
ared with the conditions of the same source in afree field. Because of these
ffects, the hypothetical test hemisphere centered about the source should
ave a radius at least twice the average distance of the source from the
:fleeting plane, but not less than twice the maximum dimension of the
urce of 2ft ( 0.6 meter).
The reflecting plane should extend beyond the farthest microphone posi-
ion a distance at least one wavelength that corresponds to the center
:•equency of the lowest frequency band to be used. The effects of other
bstacles and reflecting surfaces should be minimized by keeping them away
rom the source and the microphone positions.
Because the reflecting plane introduces a plane of symmetry, the measur-
ig points that are at equal heights above that plane have some redundancy.
'his redundancy makes them less useful for providing agood space-average
ound-pressure level than for the same number of points with no two of them
tthe same height. A double rotation of the set of 20 points of Figure 8-13
nakes it possible to produce such a set of 10 for the hemisphere, and a
epresentative set of coordinates is given in Figure 8-14. This set of points is
lot included in the standard.

(0 158, 0.898, 0.410)

(0.775,0.550, 0.313)
(-0.257, 0.498,0 828)

l-0.834, 0 315,0 452)

(0.738, - 0.067, 0.671)

(.0 100, - 0 099, 0 990)

(-0.834, - 0.399, 0.381) (0.775, - 0 603, 0.198)

(-0.257, - 0.652, 0 713)

0.158,-0.961, 0.224

Figure 8-14. Plan view of 10 points distributed on ahemisphere of unit radius.

When the hemisphere is used, the procedure for calculating power is th

same as that described for the sphere. But 3dB should be subtracted from th
power level finally obtained, because the area of the hemisphere is jus
one-half that of the sphere.
• Rotation of Source. Another way of simplifying the calculations is t
rotate the source, with the microphones placed on the surface of ahypothet
cal sphere surrounding the source, so that the projections of their positior
on the axis of rotation are uniformly distributed. A variation of this methoc
practiced by the Bell Telephone Laboratories (Jenkins, 1954), calls for th
rotation of aset of microphones about astationary source.

200
• Calculation of Power Level. If exploration shows that the basic set of
points yields representative data, the calculations of the power level and
directivity factor can be made. For auniformly distributed set of points, first
calculate the average level on a power basis. If the total range of sound-pres-
sure levels is less than 6dB, asimple arithmetical average is usually adequate.
The accurate method for any situation is as follows: Convert the decibel
readings at each of the points of measurement to power ratios by using the
tables in the Appendix, add these power ratios, and convert back to adecibel
level. Then, subtract the decibel value corresponding to apower ratio numeri-
cally equal to the number of levels used (for 8, 12, and 20 readings subtract
9, 10.8, and 13 dB, respectively). The result is then the average level, which
we shall call SPL Provided free-field conditions exist, the power level is then
calculated from the equation:
PWL = SPL + 20 log r + 0.5 dB
where r is the radius of the measuring sphere, in feet. When the rotating
source or rotating microphones are used, the average energy during acom-
plete rotation as well as for all the microphone positions should be taken, and
the corresponding average sound-pressure level used in the above formula.
• Calculation of Directivity Factor. After the average sound-pressure level,
SPL, has been determined, the directivity factor can also be calculated. If it is
desired for a particular direction, the sound-pressure level on the measuring
sphere corresponding to that direction, SPL I,is measured. The difference
between this level and the average level is called the directional gain, DG 1.
Thus,
DG 1 = SPL 1 —SPL dB
To determine the directivity factor, Q, convert the DG 1 value in decibels
into a power ratio by using the decibel tables in the Appendix. Thus, a
directional gain of — 2dB corresponds to adirectivity factor of 0.63.
Effect of Room on Measurements. The space in which power level and
directivity are to be determined must be carefully considered. As explained
previously, the measurement should ordinarily be made in an anechoic
chamber. Sometimes the measurement can be made outdoors, far from other
objects. If the device under test is normally mounted on the ground, this
outdoor measurement may be ideal, provided that the location is free from
interfering objects and the background noise level is low enough.
• Requirements on Room Characteristics If the measurement is to be made
in aroom, it should be alarge room, with extensive acoustic treatment. Large
acoustic absorption is particularly important if the directivity characteristics
must be accurately determined. In order to obtain satisfactory results in
moderate-sized rooms, extraordinarily good acoustic treatment must be used.
Many of these special anechoic chambers have been built.*
• Sound Source in a Reverberant Room. All sources that radiate sound as
discrete tones, or as very narrow-band components, and all sources whose
directivity must be determined, should be measured by the above free-field
procedure. The total power radiated by a source, whose sound energy is
distributed over awide band of frequencies, can, however, be determined in a
reverberant room — that is, aroom with hard walls, floor, and ceiling.

•Anechoic chambers of various sizes are manufactured by, for example, the Eckel
Corporation, 155 Fawcett Street, Cambridge, Mass.

201
• Measurements in a Reverberant Room. In a reverberant room, sound
power can be determined from measurements of average sound pressure in
the room and of the total absorption. The absorption is determined from a
measurement of the rate at which atransient sound in the room decays. The
procedure is as follows: The sound source in the room is turned on and the
sound is allowed to reach a steady value. The sound is picked up by the
microphone of a sound-level meter whose output is recorded on agraphic-
level recorder. The sound source is abruptly turned off, the sound in the
room decays, and this decay is plotted by the graphic level recorder. The
slope of the decay curve, in dB per second, is the rate of decay, D.
For a highly reverberant room, that is, where D is small (say 50 dB/s or
less), the sound power level of the source is then given by the following
expression.*
PWL = SPL + 10 log V + 10 log D — 47.3
where V is the volume of the room in cubic feet and SPE is the average
sound-pressure level in the reverberant field. The numerical value of 47.3 in
the above formula varies with atmospheric pressure, as shown in Figure 8-15.
For most measurements at sea level the value of 47.3 can be used.

48

o
LJ

—2 47
F- °
Z °
a CC

cr0—
zz
46
owce
co
LJ

LJ
cr 45
600 700 800 900 1000 1100
MILLIBARS
1 1 I I I I I I
18 20 22 24 26 28 30 32
IN. Hg
ATMOSPHERIC PRESSURE

Figure 8-15. Variation of numerical constant in the equation relating power level and
sound-pressure level for areverberant room.

• Room Requirements. In order for the measurement to be accurate, the

room must satisfy the following conditions:
1. If the source has abroad spectrum and the measurements are made in
octave bands, the smallest dimension of the room should be at least
equal to awavelength at the center frequency of the lowest octave band
of interest.
2. No two dimensions of the room should be alike or have aratio that is
an integer. A ratio of 1: 3 VT: Yzifor the height, width, and length is
often recommended.
3. The walls of the room should be hard and smooth. Large, hard objects
should be near the boundaries of the room to help diffuse the sound.
4. The average absorption coefficient of the room should not be greater
than .06. This requirement will lead to a decay rate that is less than
*The procedures given are essentially those of ANSI S1.2-1962, "Physical Measurement
of Sound," but the standard gives it in more detail. The standard gives the expression in
terms of the reverberation time. (See also Young, 1959; Hardy, 1959; and Wells and
Wiener, 1961).

202
walls and floor, by replacement of metal with nylon pinion gears in the drive
assembly, and by other minor modifications of the looms. The important
reductions in level were obtained in the octave bands from 250 to 8000 Hz.
Judd and Spence ( 1969) show how significant reductions in noise from
electric motors can be obtained by selection of the proper type, by the use of
a treated fan shroud, or by full enclosures. The differences in level were
mainly in the octave bands above 250 Hz, except for the full enclosure where
20- to 25-dB reductions in level were obtained over the full audible range.
Torpey ( 1969) describes how power-generating equipment can be made
quieter by avariety of techniques. Among these are sound-attenuating ducts
for air intake and exhaust and covering surfaces with sound-absorbing mater-
ials. His experiments were based mainly on measurements of the noise with
octave or 1/3-octave-band analyzers. He used an analyzer with an even
narrower band to find that avery high 250-Hz octave-band level was caused
by the tone produced by agenerator cooling fan with afundamental blade
pulsing frequency in that band.
Hudson and Crocker ( 1970) show the effects of abarrier in reducing noise
from an air-cooled refrigerant compressor. Noise reductions of from 7.5 to 16
dB(A) were achieved, and they proved adequate in satisfying complaints.
Excessive vibration of a cooling tower ( Buscarello, 1968) was traced to
tower resonance excited by impulses from rotating fan blades. Stroboscopic
observation showed that the tip of each fan blade rose about 3in. as it passed
over a nearby drive shaft. The resulting impulses occurred 900 times per min,
which was the resonant frequency of the tower. The excessive vibration was
avoided by stiffening the tower to raise its resonant frequency.
Many examples of noise or vibration reduction are given in Chapter 11 of
the All-IA Industrial Noise Manual (1966), in Chapter 7cf the Foundry Noise
Manual ( American Foundrymen's Society, 1966), in Harris ( 1957) and in
Harris and Crede ( 1961).
REFERENCES
American Foundrymen's Society ( 1966), Foundry Noise Manual, American Foundry-
men's Society, Des Plaines, Ill, 2nd edition.
AIHA ( 1966), Industrial Noise Manual, American Industrial Hygiene Association, 14125
Prevost, Detroit, Mich 48227, 2nd edition.
A.E.W. Austen and T. Priede ( 1958), " Origins of Diesel Engine Noise," Proceedings of
the Symposium on Engine Noise and Noise Suppression, London, Institution of
Mechanical Engineers, Oct, pp 19-32.
L. L. Beranek, ed. ( 1971), Noise and Vibration Control, McGraw-Hill, New York.
R. T. Buscarello ( 1968), " Practical Solutions for Vibration Problems," Chem Eng, Vol
75, Aug 12, pp 157-166.
M. Feinberg ( 1965), " New Methods Simplify Analysis of Vibration Isolation Systems,"
Part I, Mach Des, Vol 37, #18, Aug 5, pp 142-149.
C.M. Harris, cd ( 1957), Handbook of Noise Control, McGraw-Hill, New York.
C. M. Harris and C. E. Crede ( 1961), Shock and Vibration Handbook, McGraw-Hill, New
York.
R. R. Hudson and M. J. Crocker ( 1970), " The Practical Problems Associated with
Quieting a Refrigerant Compressor," Applied Acoustics Vol. 3 #4, Oct, pp 251-258.
S. H. Judd ( 1969), " Engineering Control of Furnace Noise," Amer Indust Hyg Assoc J,
Vol 30, #1, Jan-Feb, pp 35-40.
S. H. Judd and J. A. Spence ( 1969), " Noise Control for Electric Motors," Amer Indust
Hyg Assoc J, Vol 30, #6, Nov-Dec, pp 588-595.
L. N. Miller and I. Dyer ( 1958), " Printing Machine Isolation," Noise Control, Vol 4, #4,
July, pp 21-23.
R. O. Mills ( 1969), " Noise Reduction in aTextile Weaving Mill," Amer Indust Hyg Assoc
J, Vol 30, #1, Jan-Feb, pp 71-76.
P. J. Torpey ( 1969), " Noise Control of Emergency Power Generating Equipment," Amer
Indust Hyg Assoc J, Vol 30, #6, Nov-Dec 1969, pp 588-595.

241
 measurements are usually made in octave bands. The measured levels
are averaged on an energy basis for each band.
b. The unknown source is operated in place of the standard source. The
sound-pressure levels are measured at the same points as before and
averaged for each octave band.
c. For each octave band the difference in average level, between the
standard and the unknown, is applied to the known power level of the
standard source.
Requirements for Standard Source. The standard source should produce a
stable and reproducible sound. Such sources have been developed for the
Compressed Air and Gas Institute and for the fan and blower industry
(Hardy, 1959).
Requirements for Room. The measurement room should be large and its
characteristics should approach those of areverberant room. No obstructing
object should be in the immediate vicinity of the source or the microphone
positions.

• 8.3.6 Predicting Noise Levels. When the acoustic power output and the
directivity pattern of a device are known, the noise levels that it will produce
under a variety of conditions can be predicted (on the average) with fair
accuracy. These predictions are based on the principles discussed earlier in
this chapter.
If anoisy device is placed in aroom that is not anechoic, it is desirable to
measure the decay rate of sound, D, in the room; and then the following
formula, adapted from one by Young, can be used to predict the average level
of sound in that part of the room where the reverberant field dominates:
SPL = PWL — 10 log V — 10 log D + 47.8

where V is the volume of the room in cubic ft. PWL is the source power level
and the constant 47.8 varies with atmospheric pressure (to determine the
variation add 0.5 dB to the values shown in Figure 8-15).
Close to the source, the level is almost as if free-field conditions existed.
The level decreases with increasing distance from the source and the average
approaches the reverberant field level. Here, standing waves will exist and it is
only the average level that can ordinarily be predicted. At points less than 1/4
wavelength from ahard wall, the level will be higher than the average in the
reverberant field. Very near ahard wall the increase may be as much as 3dB;
very close to an edge, 6dB; and right at the vertex of acorner, 9dB.
When the decay rate in the room cannot be measured, it can be estimated
from a detailed knowledge of the room and its surface conditions. The
procedures are given in books on architectural acoustics. There the calcula-
tion procedure is normally given for reverberation time, T. The decay rate, D,
is then easily obtained as follows:

A related procedure is to calculate the room constant, from the character-

istics of the room. Then the equation given there can be used to estimate the
sound-pressure level. The discussion in paragraph 8.3.4 should also be con-
sidered in order to modify the prediction to fit the actual conditions.
As described earlier, the near-field behavior of asource may not be closely

204
related to the far-field behavior that constitutes the radiated sound power. As
aresult, aknowledge of the radiated sound power does not ensure that one
can estimate reliably the level near amachine. When it is important to know
the sound level very close to a machine, as is often the situation for
noise-exposure measurements, the actual sound level should be measured at
the desired point. As explained earlier, exploration is desirable here, because
close to the machine the sound level may vary markedly with position.
The sound-pressure level produced by the source is also affected by its
position in the room — that is, if it is suspended in the middle of the room, or
mounted on the floor, wall, or ceiling, or in acorner. It is often very difficult
to predict the exact effect, however (Waterhouse and Cook, 1965). Ordinarily
the level is higher when the source is very near ahard surface than when it is
in the middle of the room. And, as explained earlier, if the source is generally
mounted on a hard surface it should be measured that way, so that the effect
on the source is taken into account. Then the levels in another room can be
predicted with better accuracy.

8.3.7 Record of Measurements. One important part of any measurement

problem is obtaining sufficient data. The use of data sheets designed specifi-
cally for a noise problem helps to make sure that the desired data will be
taken and recorded; a sample data sheet is shown in Figure 8-16. The
following list of important items may be found helpful in preparing data
sheets of this type:
1. Description of space in which measurements were made. Nature and
dimensions of floor, walls, and ceiling. Description and location of
nearby objects and personnel.
2. Description of device under test ( primary noise source). Dimensions,
name-plate data and other pertinent facts including speed and power
rating. Kinds of operations and operating conditions. Location of
device and type of mounting.
3. Description of secondary noise sources. Location and types. Kinds of
operations.
4. Type and serial numbers on all microphones, sound-level meters and
analyzers used. Length and type of microphone cable.
5. Positions of observer.
6. Positions of microphone. Direction of arrival of sound with respect to
microphone orientation. Tests of standing-wave patterns and decay of
sound level with distance.
7. Temperature of microphone.
8. Results of maintenance and calibration tests.
9. Weighting network and meter speed used.
10. Measured over-all and band levels at each microphone position. Extent
of meter fluctuation.
11. Background over-all and band levels at each microphone position.
Device under test not operating.
12. Cable and microphone corrections.
13. Date and time.
14. Name of observer.
When the measurement is being made to determine the extent of noise
exposure of personnel, the following items are also of interest:
1. Personnel exposed — directly and indirectly.

205
 Plant- Noise Survey Form DATE F Ift09

LOCATION 80t StevENECI

METER e : 64-47
Floor, eoPt. Eiroidisa DIAGRAM
(Show amt..,
NUMBER OF PERSONNEL EXPOSED 1 location with
an X I
WEARING EAR PROTECTION , Fer
1443eLl.ti P R
VOs no
*I on
OPERATOR: J. I?, L&COan•r
* 2 on 89.6: 92
SIGNED: P. G.
Betio Ore 14. e 97 98
1505 PROCEDURE CHECK LIST

t REMOVE WHITE CAP.

2. SELECT A WEIGHTING. SLOW METER

RESPONSE OR C INFIGHTING. FAST
RESPONSE FOR IMPACT NOISE/ I I-if c) 402A,
3. TURN ATTENUATOR KNOB TO GET
P x Itç'
'
SCALE READING
SCALE FOR IMPACT NOISE/.
IOR TO 140 dB FULL
R
4. ADD READING TO KNOB SETTING TO
GET TOTAL AB.
mike 5' Z" oef el..,
NOTES "--Ope,...DP' ros b.cde oncl rtk be.Locc.., m 11.111 g
— Ste
6- er:ovres
-vp t..- co-t
e'- •••••• 04 ; lebrie,1
,uss r,0 I In.m
— 25 ,....i.ics/c1. 4-avc....›e„ 1,.)orsr c.o.sc uscol •
— 1.3.54 eme.Junsa net aiwaélrk s on dor.n e in,t.J oe non

dB Al
TOYAL
L OURA 10
IN HOURS
PEA DA

«k,E NONE
IIII

olp KIM

KM

Mil

ha;
ie
;!!ili . •I.
grillf .i a a ffNiri
e r i n . ail
mii:
l
ii 11
n _

Tr
90 lir.1.1111.1
UNO

90
R • 80
FM
IPAPACT C
iie.44kilIoci«..,,so
UNDER I. AS PEAR
tor..1 4.
83 C=Il
re=
:I .. SA WITH
1E0 AIS PULL
EYER ^ SCALE RANGE
WEIGHT AS E E HOER
M:EL. MAX — /
et dec.) TOTAL 11:

,Ir«::".•:::,:r.,=°,7
RECOMMENDATIONS:. COFIti.N.V.e. 030. Of ear protect:0A

GENERAL RADIO ,,, OAKEN AVENUE coNc.0110 MASSAC.USE Its 01/42

ferry, No J227 A

Figure 8-16. A sound-survey data sheet.

vibration pickup should be placed. Sometimes, however, the correct pickup
location is not obvious and some exploration of the vibration pattern of the
device being studied is necessary. Furthermore, the pickup must be correctly
oriented, and this too sometimes requires exploration.
Fastening a pickup to a device is usually a simple task, if the device is
much larger than the pickup and if the important vibration frequencies are
below 1000 Hz. Otherwise, difficulties may arise because of the mechanical
problem of fastening the pickup at the desired point, because the pickup
seriously affects the motion to be measured, or because the method of
attachment affects the performance of the pickup.

8.4.2 Choice and Use of Pickup.

Range of Levels. A very wide range of vibration levels can be covered by
the pickups available. The most sensitive pickup can be used with the
vibration meter or the control box and sound-level meter to measure down to
.01 in/s 2 or .0003 m/s 2 .At the other extreme, the Type 1560-P53 pickup
can be used up to 1000 g. Special pickups are manufactured for even higher
accelerations.
Frequency Range. The Type 1560-P52 Pickup is particularly well suited
for low-frequency vibration measurements. When connected to the Type
1553 Vibration Meter, this pickup can be used at frequencies down to 1Hz;
with the Type 1564 Sound and Vibration Analyzer it can be used down to
2.5 Hz. The other pickups can also be used down to 2.5 Hz, if they are
connected to aType 1560-P42 or 1560-P40 Preamplifier.
The Type 1560-P53 Pickup has a resonance at about 28,000 Hz, which
imposes a practical upper frequency limit of about 20,000 Hz. The other
pickups in the General Radio line are limited to a maximum frequency of
about 2000 Hz.
Orientation of Pickup. The piezoelectric accelerometers used in General
Radio vibration-measuring instruments are most sensitive to vibrations in the
direction perpendicular to the largest flat surface on the pickup. This direc-
tion'is the one for which the rated sensitivity applies. The sensitivity in other
directions varies approximately as the cosine of the angle with respect to this
rated direction, with a minimum of about 5% ( or less) of rated sensitivity,
when vibrated in adirection perpendicular to the rated one.
For accurate results, the pickup must be properly oriented with respect to
the direction of motion. In practice, this orientation is usually not critical,
however, because sensitivity changes slowly with direction, there being adrop
of only about 2% for a10-degree change in orientation.
The direction of maximum vibration at a point is often obvious from the
structure that is vibrating. That is, it is usually in the direction of least
stiffness. But this rule is sometimes misleading, because of the many possible
resonant modes of vibration, some of which are perpendicular to the obvious
direction of least stiffness. Such amode can be strongly excited if close to the
frequency of a component of the driving force. Furthermore, the nature of
the motion may favor one mode of vibration rather than another.
When it is important to be certain of the direction of motion, one can
measure the motion along three mutually perpendicular axes. Often one can
select these so that only one of these components of motion is significant,
and that will determine the choice of direction. Otherwise, they must be
combined vectorially to yield aresultant total; then, one needs to know the

207
•elative phase of the components. To determine phase, sums and differences
:an be measured with two pickups, as explained later, or another set of three
measurements can be made along mutually perpendicular axes that are
•otated from the first step. With two sets of measurements, one can sort out
:he possible combinations and calculate the direction of the total motion.
Xten it is simple to obtain the direction of the maximum motion by
experiment.
Except for simple harmonic motions, this resultant direction is of signifi-
:ance only as afunction of frequency. Then an analyzer is essential so that
me can determine the motion for the individual components.
When one attempts to measure vibration in a direction that is not the
Erection of the total vibration at the point of measurement, the orientation
s more critical, because the vibration in the other directions will provide
;ome signal in the output. It is often impractical to measure a directional
eomponent that is less than 5% of the total vibration at apoint.
The above procedure does not lead to a measurement of the rotational
libration about a point. This type of measurement can be made with a
orsional vibration pickup or by the technique discussed in paragraph 8.4.5.
Hand-Held Pickup. When one must explore avibration pattern or make a
pick check of the vibration amplitude, one is tempted to hand hold the
Mckup against the device being measured. If the device is massive and is
ibrating with a significant amplitude, this technique can be useful for
iequencies below about 1000 Hz. There are enough serious limitations to
his technique, however, so that it should not generally be expected to yield
Lccurate or highly reproducible results ( Gross, 1965).
When the pickup is held by hand, a test probe, a pointed metal rod, is
*astened to the pickup to facilitate applying the probe to the desired point.
Che motion is transmitted along the rod to the pickup, and the motion in the
lirection of the rod actuates the pickup.
Because the test probe adds another element to the pickup, the response is
lifferent from that of the pickup alone. Typical relative frequency response
:haracteristics are shown for two types of probe in Figures 8-17 and 8-18.
lore than one response run is shown to indicate the variability that can
)ccur. Note resonance in the range from 1400 to 2000 Hz introduced by the
ong (6 3/8-in.) probe, and the one above 2000 Hz for the probe with the
hort conical tip.
Unless the device being tested is massive, the force, mass, resilience, and
lamping introduced by the hand may seriously alter the motion, and another
nethod of applying the pickup should be tried.
Some vibration is applied to the pickup by tremor of the hand. This
ibration is made up chiefly of components below 20 Hz, and the peak-to-
ieak order of magnitude is 5 in./s 2 acceleration, 0.2 in./s velocity, and 10
nils displacement, when the pickup is held against a relatively stationary
urface. These values will be appreciably attenuated by a low-frequency
utoff at 20 Hz, such as is obtained on the " DISP-20 CUTOFF" position of
he vibration meter. Then the observed peak-to-peak displacement is of the
it-der of 0.2 mil.
This tremor sets alower limit to the vibration that can be observed when
he pickup is hand-held against the vibrating device. One should not attempt
o use a hand-held pickup down to the levels quoted above unless some
iltering is introduced to reduce the low-frequency response.

208
 25 i
LONG PROBE ( 6 3/8) I
20 ¡r
14\
15
RESPONSE IN dB

_
10

5
.._..-,','•'
o ...----
REFERENCE
5 1
\
10
\
—I 5
200 500 1000 2000 00
FREQUENCY IN Hz

Figure 847. Frequency response of vibration pickup mounted on hand - held long (6
3/8-in.) probe. Several sample responses are shown. The curve labeled REFERENCE is
the frequency response of the pickup without the probe.

25

20
SHORT PROBE (5/8) / 1
1
I/
15 / 1

/
RESPONSE IN clB

10 1

o
REFERENCE 1
5 1
\
10

—1
200 500 1000 2000 5000 Iq000 a4OOO
FREQUENCY IN Hz

Figure 8-ia Frequency response of vibration pickup mounted on hand-held short

(518-in.) probe. Several Sample responses are shown.

Pickup Fastening Methods. Pickups are fastened to avibrating surface by

many different methods. For greatest accuracy the fastening should be as
direct and as rigid as possible ( Rasanen and Wigle, 1967; Gross, 1965). But if
the acceleration is less than gravity, if only atemporary fastening is desired,
and if only low frequencies are present, simple fastenings are adequate. These
may be plasticene or double-sided adhesive tape placed between the base of
the pickup and aflat surface at the point desired. If the surface is horizontal,
flat, and smooth, the pickup may be wrung to the surface with athin film of
petroleum jelly. Another simple technique, useful on magnetic materials, is to
fasten amagnet to the pickup and then attach the magnet to the surface to be
measured.
At high accelerations, these simple fastenings are not satisfactory, and a
stud or bolt must be used to hold the pickup directly against the surface
being measured.

209
 The performance of some pickups is affected by any attachment to the
pickup body other than to the reference surface, so that apickup should not
be attached by clamps to the body of the pickup.
When the pickup is to be permanently installed, the use of an adhesive,
such as a dental cement, Eastman 910, or an epoxy cement, is often
advisable. For best results, one should be careful to use only athin layer, so
that the elastic characteristics of the bonding cement will not affect the
behavior of the pickup.
For maintenance tests, it is often convenient to fasten avery smooth flat
iron disk to bearing housings with avery hard epoxy cement. The disk should
be pressed as tight as possible against the housing. Then a magnetic attach-
ment can be used, again with athin film of silicone grease or petroleum jelly
to ensure good contact.
The fastening should be rigid, so that the pickup does not move signifi-
cantly with respect to the surface to which it is fastened. Any rocking
motion, or looseness that might lead to chattering, should be prevented. If
the fastening alone is not adequate to prevent this looseness, the use of some
plasticene in addition may be helpful. When fastening, even by bolts, the use
of a lubricant or petroluem jelly is advisable to ensure close contact between
the pickup and the fastening surface, without putting undue strain on the
pickup.
When the surface is not smooth or flat, the pickup is sometimes mounted
on abracket. For low vibration frequencies ( below afew hundred hertz), the
bracket can readily be made stiff enough so that it does not seriously affect
the behavior of the pickup.
The procedure for obtaining agood connection between the pickup and
the vibrating surface is illustrated by the specifications of MIL-STD- 740B
(SHIPS).

Transducers shall be attached as follows:

(a) Transducers shall be attached to blocks, which are to be brazed or

welded to equipment, or subbase, as close as possible to the mounting points
of the equipment to be tested.
(b) The blocks shall be made of steel and shall be as small as possible. The
block surfaces on which transducers are mounted shall be plane and shall have
a surface finish of 125 micro-inches rms or better and be mutually perpendic-
ular within one degree.
(c) Three holes in the mounting blocks shall be drilled and tapped to a
depth of at least 1/4 inch with 10-32 NF threads to accommodate triaxial
arrays of transducers which shall be attached to the blocks with insulated
steel studs. The holes shall be perpendicular to the finished surfaces within
plus or minus 1degree.
(d) Just before transducers are mounted on a block, all mating surfaces
shall be cleaned of all dirt, grease, and other foreign matter in preparation for
mounting. The surfaces of the attachment area and the studs shall be lightly
covered with clean oil or grease.
(e) The mounting blocks shall not be removed and shall be preserved with
arust inhibiting coating after completion of testing.
(f) If brazing or welding cannot be accomplished, the mounting blocks
shall be attached to the location with a thin layer of epoxy resin cement.
Blocks attached by cement shall be removed upon completion of test. The

210
transducers may be attached directly to the equipment being tested only
where there is insufficient space to accommodate the mounting block.
The pickup is calibrated in terms of the motion of the flat contacting
surface of the pickup. Because of the resilience of the fastener and the mass
of the pickup, this surface of the pickup will not move exactly as the surface
being measured moves. At low frequencies this difference is easily made
insignificant by the relatively simple techniques discussed. But at high fre-
quencies care must be used in fastening to keep this effect small.
The mass of the Type 1560-P52 Vibration Pickup is sufficiently small that
simple temporary fastenings are adequate even to frequencies beyond the
normal resonance at about 3200 Hz. This fact is illustrated by the response-
vs-frequency characteristics shown in Figure 8-19 and 8-20. In each instance,
the pickup was driven at a constant acceleration. The reference condition is
the response for the vibration pickup wrung to the smooth, flat surface of the
driver with petroleum jelly lubricant. The acceleration was . 002 g.

25

20

I5
RESPONSE IN dB

10
/ 3M Y9010
TRANSFER TAPE
5
..-- \I l
0 \ 3M Y400
DOUBLE FACED
5 s .,.."/ TAPE—

REFERENCE
10
RESPONSE

15
200 500 1000 2000 5000 10,000 20,000

FREQUENCY IN Hz

Figure 8-19. Frequency response of vibration pickup attached by means of Minnesota

'Mining Y9010 and Y400.

25

20

15
dB

10
IN
RESPONSE

5
PLAST ICI NE

REFERENCE
10

—15
200 500 1000 2000 5000 goo°
I 20,000
FREQUENCY IN Hz Veeonsi

Figure 8-20. Frequency response of vibration pickup attached by means of 1/16-in, thick
layer of plasticene.

211
 The effect of fastening by means of double-sided tape was generally less
han 10% deviation from the reference condition at all frequencies up to the
•esonance at 3200 Hz. In some instances, the deviation over the range to
1200 Hz was only about 2%. The variability was probably aresult of changes
ncontact adhesion obtained with different samples of the tape.
Plasticene as afastening means, even as thick as 1/16-in, showed very good
eproduction of the reference performance, being within 2to 5% up to 4000
.

U. In one instance, amarked departure from the reference performance was

'ound even at 500 Hz, and this was quickly traced to the fact that the pickup
iad come loose from the plasticene. This example illustrates the importance
)f careful inspection of the fastening during a test, particularly when one
:annot check the performance independently.
The response of the pickup when held to a smooth, flat, steel plate by
neans of the Type 1560-P35 Permanent Magnet Clamp is shown in Figure
l-21. Up to 5kHz, the response is very similar to the reference response. One
hould fasten the pickup carefully to the magnet so that no rocking motion is
)ossible, and the magnet itself should be placed on asmooth surface so that
t, too, will not rock; otherwise, serious errors may result.
25

20

15
dB

I
0
IN
RESPONSE

5
M60- P35
O PERMANENT _
MAGNET CLAMP

-5 REFERENCE

10

-I 5
200 500 1000 2000 5000 1Q000 20,000
FREQUENCY IN Hz Œffa
=igure 8-21. Frequency response of vibration pickup attached by Type 1560-P35
2ermanent-Magnet Clamp.

8.4.3 Stray Effects.

Effect of the Pickup on the Vibration. The mass added by the pickup to
he vibrating surface being measured changes the motion of that surface. If
he added mass is much smaller than that of the vibrating surface and is
:losely coupled to it, the effect is small except near resonant modes. Thus, it
simportant to have alightweight pickup.
One can often judge the effect of adding the mass of a pickup by noting
he difference in behavior with the pickup fastened and with another mass
:qual to that of the pickup, in addition to the pickup. If the difference is
iegligible for these two conditions, the effect of the pickup is usually
inimportant. Under certain conditions near the resonant vibration frequency
4* the device under test, even asmall mass can shift the resonance enough to
.ffect the motion at the original resonant frequency by alarge amount.
When it is possible to change the excitation rate or frequency so that
esonance with the pickup in place is re-established, the behavior at the new
esonant point will often be sufficiently similar to the resonance behavior

212
without the pickup that the resonant condition can be satisfactorily meas-
ured.
When stroboscopic observation of the motion is possible, the effect of the
mass of a pickup on the motion can often be judged by direct observation of
the behavior with and without the pickup present.
Mounting of the Device Under Test. The actual vibration that a device
experiences will depend on the way in which it is mounted. If it is rigidly
mounted to amassive concrete structure, the vibration may be much less than
if it is mounted with avery resilient mount. For many tests the very resilient
mounting is preferred in order to obtain the maximum motion. But often the
proper procedure is to mount the device for avibration test just as it will be
mounted in actual use.
Background Vibration. Some background vibration is always present. If a
motor is put on a factory floor for a vibration test, it will be possible to
measure motor vibration even when it isn't running. This background vibra-
tion must be considered as a lower limit to the vibration that can be
measured. But, of course, one can do something about this lower limit. Often
placing the device on athick felt or foam pad will isolate it sufficiently from
the background, but then the mounting is no longer rigid. Another approach
is to use aseparate, massive concrete block as atable on which to mount the
device in any way desired. The block is suspended by resilient mounts. The
natural vibration frequency of the block on its mounts should be made
significantly lower than any frequency of interest in the test.
Peak Versus rms or Average. Although a few applications of vibration
measurements require the use of the peak or peak-to-peak amplitude, most
experimenters specify these values only because they are traditional. When
vibration signals are analyzed to find the individual components, however, the
rms or rectified average values are more useful. This usefulness depends on
two facts. First, rms component values can be summed on an energy basis to
give the over-all rms value. For many wave shapes this result is also essentially
true for the rectified average values. But the result of combinations of peak
values of components can be misleading and confusing, particularly for
coherent periodic signals, which are relatively common in vibration work. The
second fact is that if the signal is random in amplitude distribution, there is
an additional inconsistency among peak values. As a result, if you measure a
peak value of a vibration signal, it is also wise to note the rms or average
value.

8.4.4 Calibration of Vibration Measurement Systems. In order to ensure

that one can make satisfactory vibration measurements, the instruments used
must be kept in proper operating condition. The vibration meter itself can be
checked electrically very simply by the built-in calibration system, and the
instrument should be checked at the start of atest, after the instrument has
been on and allowed to stabilize for afew minutes.
The vibration calibrator should also be used regularly to check the com-
plete measurement system. If the acceleration produced by the calibrator
reads between 340 and 430 rms (or 950 to 1220, peak-to-peak) in./s 2 on an
electrically calibrated vibration meter, there is reasonable assurance that the
pickup and the meter are operating correctly. If the agreement is not
satisfactory, one should first check that the correct pickup for the vibration
meter is being used, and that the internal reference dial in the meter is set to

213
the correct pickup sensitivity. lf these are all checked and agreement is still
unsatisfactory, another pickup should be tried (with the internal reference
dial set to its nominal sensitivity). The next step would be to have the pickup
and the calibrator checked at the General Radio Company.
Vibration pickups are rugged and stable, but they can be damaged.
Although a damaged pickup will ordinarily be detected by the check at 100
Hz provided by the vibration calibrator, it is possible, but most unusual, for
the sensitivity at other frequencies far from 100 Hz to be affected when that
at 100 Hz is not. Therefore, the frequency response of pickups should be
verified periodically by calibration at the National Bureau of Standards or at
the General Radio Company.

• 8.4.5 Simple Method To Determine Rotational Vibration.* When analysis

of rotational vibration must be made on an existing installation and ator-
sional pickup cannot readily be used, the following technique may be useful.
Two vibration pickups and a summing network are required in addition to
the vibration meter. One must assume that the engine behaves like arigid
mass and that its center of gravity is equidistant from all four mounting posts.

Figure 8-22. A simple resistive summing circuit.

I
- -

Figure 8-23. A convenient arrangement of the summing circuits of Figure 8-22.

A simple summing circuit is shown in Figure 8-22. The voltages el and e2

represent the output signals of two vibration pickups, and the voltage e,
represents the signal fed into the input of the vibration meter. If the three
resistors R are equal, e, will be 1/3 (el + e2 ). A practical arrangement of this
circuit is shown in Figure 8-23. Only two resistors are shown, since the third
resistor is in the input circuit of the vibration meter. One pickup is connected
to input No. 1and the other pickup is connected to input No. 2. The output
of the summing circuit is connected to the input of the vibration meter. When

'This method was suggested by Mr. George Karnpennan of ICamperman Associates,

Downers Grove, Ill. He has used this technique on numerous occasions with gratifying
results.

214
switch S-1 is at position 1, one third of the output of pickup No. 1is applied
to the input of the vibration meter. When S-1 is set at position 2, one third of
the output of pickup No. 2 is applied to the input of the vibration meter.
When S-1 is at position 3, one third of the sum of the outputs from the two
pickups is applied to the input of the vibration meter.

7,5
o'

8
b'
Figure 8-24. Outline of engine and mounting (top view).

The top view of a typical engine and its mounting is outlined in Figure
8-24. A and B represent the forward engine mounts while C and D represent
the rear engine mounts. With the two pickups (oriented for vertical-displace-
ment measurement) mounted on the two forward engine mounting brackets,
the translational or vertical amplitude is indicated on the vibration meter
when the output signals of the two pickups are summed. The rotational mode
of vibration is canceled out. When the outputs of the two pickups are
summed 180° out of phase or subtracted, the rotational amplitude of vibra-
tion is indicated on the vibration meter and the translational mode of
vibration is canceled out. The pickup can be mounted upside down to shift
the output 180° to perform the subtraction of outputs, or, the pickup can be
left in its normal mounting position and the 180° shift in the phase of its
output can be achieved with an electronic phase inverter. By making the set
of measurements outlined above on all pairs of mounting brackets (A,B-A,D-
C,D-B,C) the amplitude of any rotational or rocking motion about the axes
a-a' or b-b' can easily be sorted out from the direct vertical or translational
motion of the engine.

REFERENCES
Acoustical and Insulating Materials Association ( 1971), Performance Data: Architec-
tural Acoustical Materials, Acoustical and Insulating Materials Association, 205
West Touhy Avenue, Park Ridge, Illinois 60068.
P. K. Baade ( 1964), "Sound Radiation of Air-Conditioning Equipment; Measurement in
the Free-Field Above aReflecting Plane," ASHRAE Trans, Vol. 70, pp. 217-227.
L. L. Beranek ( 1954), Acoustics, McGraw-Hill, New York, Part XXIV.
A. J. Ellison, C. G. Moore, and S. J. Yang ( 1969), "Methods of Measurement of Acoustic
Noise Radiated by an Electric Machine," Proc Inst Electrical Eng., Vol. 116, #8,
August, pp. 1419-1430.
H. J. Gober and E. Lubcke ( 1966), "Sound Field in Very Wide and Long Rooms," J
Acoust Soc Am, Vol. 39, #6, June, p. 1266.
E. E. Gross, Jr. ( 1965), "Vibration Measurements with the Sound-Level Meter," General
Radio Experimenter, Vol. 39, #7, July, pp. 3-9.
H. C. Hardy ( 1959), " Standard Mechanical Noise Sources," Noise Control, Vol. 5, # 3,
May, pp. 22-25.

215
 ....:, .Ù C
,1A0 n •
5 eSi > eel

ese „"›.
::etl e
tob
V>, 1
14
oo
u 4-,
r.

_in tzt.
0) «.' .
F
-0,.,„
e {-+c o>
y ee
›, ,.
÷,è

gc
y

...
eL f) c
te 3
) \.°
cos qk

•,•:• ,..„ I. C uv
g E • cj
4C3 -4- â .'

.—.
• .... c4.
cts II) .1-'
Ir vs
O.
cos et'
e

• t e ._.

..5 5 ..e
o 4,.
.W511Ble:, .004111 à ...111:

(I) E , 'et, >

4,
Î Me , c•:e. IN i '
la
IMO. W. .•;: :Nrelledel.U.:eZ >;ctileliekge. 1%;,..:!.•.11.tiew».. .. eueme..
,Ale . ,. le, A.
: ISI ae
12 _aW
:••:.. -414.
.. tiag ...SX•f.noealtà0Bletz*A-e•lanour.e.::,:,K.
..70
tammedu‘zemieumo: ›m:equew..eetze›,•e
,6ffeev
etizieoeoto&.auuhoulamysettautatdegeeiaeleiefeue.
O 41.) 4,4
e.MgaMiesqreiriSiin 74 t:erelineeeNtieleeZ-NMgeileiglí'ee eSnIk:IP egOF
I.., 0 ..-•

y ••••• • 4,
-r Irlr
- rneeig>irr •,'" '..., '' r V e • i
= O. CU gOl
,k§M
e

, :NUUL.4
.•
:. • .. C
a.
V

.
eee 42;
Au.
Le r.
II
eliediU "R ••
ij euresser,

ce o
F»NâMK,

1: 1A0111r4 .
1eKeeidekM12 -
1 ,
,
e aral», ......
Nfekble«réeila ,AM MIIMIle'ee
,•Mi ..dOmesie.e.,eiro.aouip.kz,
eauliitZ* ›
eu;........ .eutiu:.;::-. .
7«ekesesi.2smeu».z.qUIZIQ*elalleZINCVMe.«
,.ine
be i
'elt:kerit
i DIM
II . 101i, Lo
c.1
.., çant:.:«. U l sgeeee.ww.d.eeew
eu o
o ur, 4»
W... mek*WICMItO',,Wigwey%r.kamee>
-• • - . .
wer r
. ele : egniwineer--1 e.e8e17,,,:4,
ve •et,..W..4
iF' '
le4 1.,

...e.,..0.„.-417.- . lrgo....ree-zo,--
01 •
CU
Il
..... arz.z.r.ir

o
zere, qz-ilmetilt.:,.-.L,24:.1-tiez -,,.,:,-ext2leil.e
:eze4ii
-pê,.. eurasce. ;:ellibentare:
.r., ,vt.
ir, tà

--
gl
CI
r
esi
teb

•mewfee:. . imid- 24t,eei.à - -

À0editeng Illtilte
-ct
eel> ‘keMegMbr
„.• .,,,„, 111e, ,
*t4 44gkk .
q
7
.fre
. e
rk • „ ono ,

Cd
ag..1....i.e•-• -
.,OrIlle.1,....4e.

>c ritalre 14g11ig« 'eel!" '1

-'
re. ''''':»>;:lo:ce ...... 4. CZieli. RAtemiustege;
-
*oleo , .desepalery*..14useep.40e en7e.SeQUiliKeetedtêlakileMMi.miF:Uplee:*21211enFetkq
me ,-
eml8MMirlie, .:w*-M->CMVVMWMi.-Vvemcm«uwÇVVM
The charts

102gaitallinelefeW edardiWIMWSEK4Wgelitiffe2iig§1116,12ILW.MOi.:
>
'eM,1
Q) .4i eN>, ee.UYWrgMeNUMbàn,"WbAkIreg>>.>Nbàwamewe
a §, .3 to g 00 .5 .5 00 cf,.5

8
o,
4.o
a o o b°
o
oar
4
55).
Appendix V

Definitions
This section on definitions includes most of the technical terms used in
this handbook. Most of the definitions are selected from the American
National Standard Acoustical Terminology ( S1.1-1960 (R1971)), and those
definitions are marked with an asterisk. They are printed with permission.
Many have been shortened by putting them in the usual dictionary form.
A number of these standard definitions are very technical in order to be
precise. Some readers may find it easier to refer to the discussion in the main
text of this handbook to obtain ageneral underitanding of those terms.
The nonstandard definitions have been adapted especially for this
handbook.
ACCELERATION*
vector that specifies the time-rate-of-change of velocity. Note 1: Various
self-explanatory modifiers such as peak, average, rms are often used. The time
interval must be indicated over which the average (for example) was taken.
Note 2: Acceleration may be ( 1) oscillatory, in which case it may be defined
by the acceleration amplitude (if simple harmonic) or the rms acceleration (if
random), or (2) nonoscillatory, in which case it is designated "sustained" or
"transient" acceleration.
ALIAS
in sampled, equally spaced data, two frequencies are aliases of one another
if sinusoids of those frequencies cannot be distinguished by the sampled
values.
AMPLITUDE DENSITY DISTRIBUTION ( PROBABILITY DENSITY DIS-
TRIBUTION) ( FREQUENCY DISTRIBUTION)
afunction giving the fraction of time that the pressure, voltage, or other
variable dwells in anarrow range.
AMPLITUDE DISTRIBUTION FUNCTION ( DISTRIBUTION FUNCTION)
(PROBABILITY FUNCTION) ( CUMULATIVE FREQUENCY FUNCTION)
a function giving the fraction of time that the instantaneous pressure,
voltage or other variable lies below agiven level.
ANALYZER
a combination of a filter system and asystem for indicating the relative
energy that is passed through the filter system. The filter is usually adjustable
so that the signal applied to the filter can be measured in terms of the relative
energy passed through the filter as afunction of the adjustment of the filter
response-vs-frequency characteristic. This measurement is usually interpreted
as giving the distribution of energy of the applied signal as a function of
frequency.
ANECHOIC ROOM ( FREE- FIELD ROOM)*
a room whose boundaries absorb effectively all the sound incident there-
on, thereby affording essentially free-field conditions.

256
are(ivalls, barriers, and total enclosures> Almost any degree of reduction of
airborne sound can be achieved by a total enclosure or a combination of
several enclosures. But as the required attenuation increases so does the
complexity, weight, and cost. In addition, great care must be taken that the
attenuation gained by the enclosure is not lost by sound transmission through
a ventilating duct or by solid-borne vibration. Because of this possible
flanking transmission in ventilating systems, total enclosures frequently re-
quire carefully designed ventilating systems with ducts lined with absorbing
material. These lined ducts are essentially mufflers for the air stream.
When a door is required in a total enclosure, it should be built with
air-tight seals at all joints. A refrigerator-type door is usually satisfactory
when it can be used. A total enclosure should also be lined at least on part of
the inside walls with absorbing material. This lining helps to keep the noise at
the walls of the enclosure at the lowest practical level.
A barrier is not as effective as a total enclosure, but it does help to shield
high-frequency sound. Little attenuation of low-frequency sound is obtained
unless the barriers are very large, and the attenuation of high-frequency sound
is usually only a few decibels unless the opening that remains is relatively
small. Here, too, absorbing material should cover the barrier to avoid exagger-
ating the level by reflections from the barrier.

9.3.4 Illustrative Example. In order to illustrate the possible noise reduc-

tion achieved by use of vibration isolation, barriers, enclosures, and acoustic
treatment, an example made up for the purpose is shown in aseries of figures,
Figures 9-1 and 9-2. We intend to show here only the general nature of the
noise reduction obtainable as given by changes in the octave-band spectrum
and the speech-interference level. Actual results will vary in detail, and
situations do occur where the results differ materially from those shown
because of factors not considered here. But, in general, the noise reduction
shown in the figures can be considered typical.
Figure 9- la shows the octave-band analysis of the noise from the assumed
machine. The speech-interference level is also shown. This machine is anoisy
one with aspectrum that shows appreciable noise energy all over the audible
range. All the noise measurements are assumed to be made in the relative
position shown for the microphone, designated M on the figures.
The use of vibration isolation mounts may be an important step in noise
control. As shown in Figure 9-1b, the initial result, however, is often only a
moderate reduction of the low-frequency noise. The machine itself usually
radiates most of the high-frequency noise directly to the air, and the amount
radiated by the floor is small. A reduction in the vibration level at the floor
only is then not important at high frequencies. At low frequencies, however,
the machine may be too small to be effective in radiating sound, and then the
floor may act as asounding board to contribute materially to low-frequency
sound radiation.
It is even possible to increase the noise as aresult of the use of vibration
mounts. This result is usually found when the stiffness of the mounting is of
such a value that some vibration mode is exaggerated by resonance, but
resonance can be avoided by proper design of the mounting. In the illustrative
example it is assumed that the mounting is sufficiently soft that the basic
vibration resonance of the machine on the mounting system is below 20 Hz.
In this particular example no significant change in the speech-interference
level is shown as aresult of the use of vibration isolation mounts alone.

219
 0

7.42122.66140 222226620 OCTAVE

0.
SP2tCH-t6T12ERENCE LEVEt
(6611 1

2000 4000 1000

3 soo eoo
6 ' 25 250
OCTAVE BAND - HI
W S,S

8
b.

VIBRATION moo:net
250 soo eoo 2000
OCTAVE BAND . 162

ACOUSTICAL ABSORBING MATERIAL

BAFFLE

C.

500 .00
125 100
OCTAVE BAND - H2

MATERIAL
ENCLOS URE OF ABSORBING

4000
2000
tiS 250 500 ‘000
OCTAVE BAND - HI

Fi
gure9-1. Examp les to illustrate the poss ibl e noise reduct ion effects

o fsome noise control measures; i0 is microphone position.

220
 RIGID, SEALED ENCLOSURE

PE 1
L

er
.400141 ED
a.

ORO MO D

.0

30
3S SS 125 250 500 000 e000 4000 POO
OCTAVE BAND - N,

100

/7//////////////////
50

00141 E0

b. 'PS I
L

OBI S MOD

7J// • 7, /
VIBRATION ISOLATION 30
3O 63 250 soo moo 2000 4o'
oo 0000
OCTAVE BAND - Hz

ACOUSTICAL ABSORBING MATERIAL

C.

Sl IS 63 ns 250 500 000

, 2000 4000 WOO
OCTAVE BAND - HT

5 250 soo
OCTAVE BONO- Hl

Figure 9-2. Examples to illustrate the noise reduction possible by the use of enclosures.

221
 The results shown in Figure 9-1c illustrate that abarrier is mainly effective
at high frequencies, and there it produces only amoderate reduction in noise
level.
The novice in this field sometimes assumes that the materials used for
sound absorption can also be used alone for sound isolation. If we build an
enclosure solely of these materials mounted on alight framework, we would
typically find the result shown in Figure 9-1d. Only at high frequencies do we
have anoticeable reduction in level, and even there it is asmall reduction.
A more satisfactory enclosure is built of more massive and rigid construc-
tional materials. Assume that we enclose the machine by awell-sealed, heavy,
plasterboard structure. Then we might observe the result shown in Figure
9-2a. Here an appreciable reduction is obtained over the middle- and high-fre-
quency range. The enclosure is not as effective as it might be, however,
because two important factors limit the reduction obtained. First, the vibra-
tion of the machine is carried by the supports to the floor and then to the
whole enclosure. This vibration then may result in appreciable noise radia-
tion. Second, the side walls of the enclosure absorb only asmall percentage of
the sound energy.
The addition of a suitable vibration isolation mounting will reduce the
noise transmitted by solid-borne vibration. This effect is illustrated in Figure
9-2b. Here we see anoticeable improvement over most of the audio spectrum.
When the sound absorption within an enclosure is small, the noise energy
from the machine produces a high level within the enclosure. Then the
attenuation of the enclosure operates from this initial high level. The level
within the enclosure can usually be reduced by the addition of some sound-
absorbing material within the enclosure, with the result that the level outside
the enclosure is also reduced. This effect is shown in Figure 9-2c, which
should be compared with Figure 9-2b.
If even more noise reduction is required than that obtained by the one
enclosure, asecond, lined, well-sealed enclosure can be built around the first.
The first enclosure is supported within the second on soft vibration mounts.
Then a noise reduction of the magnitude shown in Figure 9-2d can be
obtained.

9.4 SUMMARY OF NOISE REDUCTION PROCEDURES.

The approach to anoise reduction problem can be summed up as follows:
•1. Consider the source.
( Can aquieter machine by substituted?
Can the noise energy be reduced?
Can auseful change be made in the directivity pattern?
Are resilient mounts of any use here?
Can amuffler be used?
2. Consider the path from the source to the listener.
Can the source or the listener be readily moved?
Is acoustic treatment auseful solution?
Should barriers be erected?
Is atotal enclosure required?

9.5 VIBRATION REDUCTION.

The basic procedure for vibration reduction will be described briefly.
Many specialized techniques have been developed also, and a complete
summary of these is impractical here. More extensive information on vibra-

222
tion reduction will be found in the references.
The first step is usually a careful inspection to see if a common-sense,
simple, quick solution is available. A part may be loose or broken, and
fastening it or replacing it may cure the trouble. If asolution is not obvious, a
systematic approach to the problem is suggested.
,- The approach to reducing vibration is summarized as follows:
1. Change source or coupling to vibrational driving force.
Reduce its strength.
Eliminate it by substitution, or otherwise.
! Isolate it.
Change its character, frequency (speed).
/ 2. Reduce response to driving force.
K Insert isolating members.
Damp vibrating elements.
Detune resonant systems.
Change mass. Increase mass of stationary elements or reduce mass of
moving elements.
Change stiffness.
Add auxiliary mass damping or resonant absorbers.

9.5.1 Changing the Driving Force. In order to see how the driving force
can be changed, it is useful to review the many ways that avibratory force is
developed. Here there are two basic processes involved. Either mechanical
energy of some type is coupled into mechanical vibratory energy by one or
more methods, or energy in some other form is transformed into mechanical
vibratory energy, as outlined below.
1. Mechanical
Unbalanced rotating masses.
Reciprocating masses.
Fluctuating mechanical forces or torques.
Fluctuating loads.
Fluctuating mass or stiffness.
Poorly formed moving components.
Mechanical looseness.
Misalignment.

2. Transformation from another form of energy.

Varying electrical fields.
Varying hydraulic forces.
Aerodynamic forces.
Acoustic excitation.
Varying thermal conditions.
Sometimes the source of the vibratory force is readily apparent or well
known from experience. At other times use of some measuring instruments
can be invaluable in tracking down sources.
Here are some examples:
Stroboscopic observation of a cam and follower showed that above a
certain speed the follower did not remain in contact with the cam during
parts of the cycle. When the cam periodically came into contact with the
follower after the period of separation, a serious impact occurred, which
resulted in excessive vibration and noise. ( Figure 9-3).

223
PERIODIC QUANTITY *
oscillating quantity whose values recur for certain increments of the
independent variable.
PHON *
unit of loudness level. (See LOUDNESS LEVEL.)

PINK NOISE
noise whose noise-power-per-unit-frequency is inversely proportional to
frequency over aspecified range.

PITCH *
that attribute of auditory sensation in terms of which sounds may be
ordered on ascale extending from low to high. Pitch depends primarily upon
the frequency of the sound stimulus, but it also depends upon the sound
pressure and wave form of the stimulus. Note 1: The pitch of asound may be
described by the frequency or frequency level of that simple tone, having a
specified sound pressure level, which is judged by listeners to produce the
same pitch.
POINT SOURCE See " SIMPLE SOUND SOURCE."

POWER LEVEL
in decibels, is 10 times the logarithm to the base 10 of the ratio of agiven
power to a reference power. The reference power must be indicated. [The
reference power is taken as 1.0 x 10 -12 watt in this handbook.]

PRESBYCUSIS
the condition of hearing loss specifically ascribed to aging effects.
PRESSURE SPECTRUM LEVEL
of asound at aparticular frequency is the effective sound-pressure level of
that part of the signal contained within a band 1 cycle per second wide,
centered at the particular frequency. Ordinarily this has significance only for
sound having a continuous distribution of energy within the frequency range
under consideration. The reference pressure should be explicitly stated.

PRIMITIVE PERIOD ( PERIOD) *

of a periodic quantity is the smallest increment of the independent
variable for which the function repeats itself. Note: If no ambiguity is likely,
the primitive period is simply called the period of the function.
PROBABILITY DENSITY DISTRIBUTION
See AMPLITUDE DENSITY DISTRIBUTION.
PSIL (THREE-BAND PREFERRED-OCTAVE SPEECH- INTERFERENCE
LEVEL) ( SPEECH INTERFERENCE LEVEL)
average, in dB, of the sound-pressure levels of a noise in the three octave
bands of center frequency 500, 1000, and 2000 Hz. The speech interference
level, or SIL, without the qualifying " PREFERRED" is usually the arithmetic
average of the sound-pressure levels in the older series of three octave bands:
600 to 1200, 1200 to 2400, and 2400 to 4800 Hz.

PURE TONE See SIMPLE TONE.

QUANTIZATION
conversion of avalue into one of alimited set of values. The limited set is
usually a discrete series of total number equal to two raised to an integer
power, that is, abinary set.

263
 frequency analysis of a vibration often shows up strong components
whose frequencies can be related to certain shaft speeds or gear-tooth mesh-
ing frequencies, and in this way the source can be tracked down', Sometimes,
however, the relations are not simple. As pointed out by L. S. Wit ( 1962) the
gear-tooth meshing frequency may be modulated at rates determined by shaft
speeds, because of run out, and by the rate at which the torsional loading
varies ( Figure 9-4).
40 HOR12,1•IT al SC ALF It. 1
.t. !MY
RELATIVE ACCELERATION IN dB

400 500 600

FREQUENCY IN c/s =IQ

Figure 9-4. A part of the recorded frequency analysis of the vibration of a gear-belt
drive. The tooth-contact rate of 450 impacts per second determines the frequency of the
dominating component, and the gear belt with its speed of 5 rps introduces a host of
components spaced about the main component by multiples of 5 hertz. The torque
pulsations from the 1800 rpm synchronous motor and the 120-Hz magnetically driven
vibration in the motor also influence the spectrum.

4 or devices that are electrically driven, strong vibration components at

frequencies that are multiples of 120 Hz are good indications that these
vibration components are electro- magnetically excitec> Sometimes one can
check this deduction by monitoring the level of such components, first when
the device is operating normally, and then when the electric power is
suddenly disconnected. Usually the driven devices will coast long enough so
that the mechanical forces will not change rapidly even though the electrical
forces are changed abruptly.
hen a device can be driven at varying speeds, the effect of changed
speeds on the frequencies and amplitudes of the various important compo-
nents can be an important clue in tracking down the sources of those
componentsere the changes in shaft speeds and mesh frequencies can be
related to changing or steady frequencies.5e>This technique is particularly
helpful if the relative speeds of some parts n be changed or if aclutch can
be sed to deactivate some sections.
hen the indication on the meter of avibration meter or of abroad-band
analyzer fluctuates erratically over a range of 2 to 1 (6 dB) or more, the
vibration is usually random in character, and the source is then probably to
be found in some rattle, friction-induced vibration, turbulence, poor ball
bearings, gases or liquids in motion, or combustion processes. The relative
value of a peak and average reading also serves to differentia fe this type of
vibration from the simpler harmonic motion of rotating unbalanced masses.

225
For simple harmonic motion the peak value will be about 1.5 times the
average value (and the peak-to-peak, about 3times). For random signals the
ratio is usually much higher, that is 3to 4times (or 6to 8times for peak-to-
Pee)-
istening by means of a pair of earphones to the signal picked up by the
vibration pickup can be helpful in determining the cause of a vibration,
particularly if the source is defective ball or needle bearings or air leaks,
which give arough quality to the sound)The earphones should have agood
pair of ear cushions or muffs to keep out extraneous sound.
Once the mechanism producing the vibratory force is recognized, areview
of the possible means for reducing the force is in order. Thus, balancing
techniques can be applied, better gears or bearings can be substituted, proper
lubrication can be applied, the mechanical structure can be improved (for
example by lightening the moving members and increasing the weight of
stationary members), and gas or liquid velocities can be reduced.

9.5.2 Balancing Rotating Machinenenbalance in rotating devices is one

of the chief causes of excessive vibrationThis single cause is so important
that extensive discussions of it will be found in anumber of the books listed
at the end of this chapter. See, for example, Muster and Senger ( 1961) and
Wilcox ( 1967).
Balancing criteria and current practice have been reviewed by Muster and
Flores ( 1969). The degree of balance that is sought depends on the type of
device, its applications, and the rotor speed. The highest quality of balance is
required for gyroscopes and precision grinders, and the poorest quality is
tolerated for the crankshaft drives of slow, rigidly mounted engines. A
standard has been published by the Association of German Engineers (VDI
2060-October 15166) for recommended allowable degrees of unbalance, and
some indirect criteria have been published in terms of acceptable vibration at
abearing housing.

9.5.3 Reducing Response to Driving Force(A further important step in the

process of vibration reduction is to reduce the response to the driving fore..e'?•
(Morrow, 1963, pp 56 ff). Here, too, measurement techniques can be valuable
in guiding the approach to reducing the response. For example, exploring for
maxima in vibration level may show up resonance modes of vibrations of
plates and other structural members. It can show where damping may be
most effective or where resonant absorbers can be added. It may also show
where detuning can be used.
Resonance Effects.KThe phenomenon of resonant vibration occurs fre-
quently; for example, resonant vibration is essential to the operation of most
musical instruments. The undesired resonances in some automobiles at certain
speeds can be very annoying>
The effects of resonant vibration in rotating machinery can be so serious
that the design of these devicescludes the calculation of the critical speed-7
(resonance frequencies). These calculations are used to make certain that
whenever possible, the critical speeds are not included in the normal operat-
ing range of the device.
The resonance or natural modes of vibration for many types of simple
structures have been calculated. Some of these are beams, shafts, plates, and
stretched wires. The frequencies of resonance depend, for example, on the
shape, dimensions, stresses, mounting, and material characteristics. he fre-

226
quencies can also be affected by coupling to other structure
The nature of resonance is readily illustrated by vibration of atable on
which a mass is flexibly mounted with the table driven at a constant
amplitude but at different frequencies. At a certain critical frequency the
motion of the mass will be greater than for frequencies just slightly higher
and slightly lower. This frequency at which amaximum in vibration occurs is
a resonance frequency. If the structure being shaken is relatively complex,
many such maxima can be observed. ( It is often helpful to use astroboscopic
technique to make this motion visible at a slowed-down rate.) Minima of
motion may also be due to resonances.
In an actual operating device, resonant conditions may be obvious because
of excessive noise or observed vibration at certain speeds. Exploring by means
of avibration pickup, for the points at which vibration is much greater than
for other places on the device, will often locate the resonant elements. The
resonances may be of the simple type where a mass is mounted on aflexible
support, or they may be of the plate-mode type, where the mass and
flexibility of a plate or sheet are in resonance, so that different parts of the
plate are moving differently. In this latter instance very complicated motions
may, result.
<Unless there is some significant dissipation of energy (damping) as the
system vibrates, the resonance amplitude of motion may become very large,
even with a relatively/ small driving force- ) These large amplitudes must
ordinarily be avoided.\The two principal ways of reducing these amplitudes
are detuning and damping)If the driving force is at a relatively fixed
frequency, it may be relatively easy to move the frequency of resonance out
of the operating range by aCchange of the resonant-element mass or stiffness -)
or both. The use of damping devices or highly damped materials is the other
important possibility.
Many techniques for damping vibration have been developed. They include
dashpots and other viscous absorbing systems, mastic coatings, sandwich-type
dissipative materials, inherently dissipative plastics or metals, electromagnetic
damping, frictional rubbing devices, and dynamic absorbers.
Measurements of the vibration levels at various parts of the device under
study can help to show where damping devices can be applied most effective-
ly. Thus if aresonance condition is to be damped, an analyzer tuned to the
frequency of resonance should be used on the output of apickup. Then when
the measurements are made at different points on the vibrating device, only
the vibration component at the resonance frequency will be observed, so that
the actual resonance maxima can be obtained without being obscured by
high-amplitude low-frequency vibrations. When such measurements are made,
the vibration pickup must be light in weight compared with the mass of the
resonant element, so that it does not appreciably detune the resonant system.
Whenever possible stroboscopic observations should be made, since this can
be done without affecting the vibration.
As an example of the effectiveness of damping in reducing vibration,
Ruzicka ( 1964) reports on an aluminum chassis for electronic modules that
was giving trouble because of fatigue failures and incorrect operation be-
cause of the collision of modules during vibration of the chassis. The installa-
tion of stiffening plates made of visco- elastic-damped material reduced the
vibration amplification at the main resonance modes by factors of 3to 4,
and the vibration-caused problems were eliminated.

227
 Oscillating Condition n most instances resonance is exhibited when the
natural-mode frequency of avibrating member coincides with, or is very close
to, one of the component frequencies of the driving forl*KCertain unstable
systems, however, do not require this coincidence when the conditions make
the system self-oscillator))) They require what is essentially a reasonably
steady driving force Galloping transmission lines and some forms of ma-
chine-tool chatter, electrical brush squeal and automobile shimmy are exam-
ples of this type of excitation.
, The galloping and torsional oscillations in some suspension bridges are
aerodynamically induced forms of vibratn. Such torsional oscillations de-
stroyed the first Tacoma Narrows Bridge on November 7, 1940. The proper
aerodynamic design of such astructure can essentially eliminate this vibration
(Steinman, 1956).
The Mackinac Bridge as amatter of fact is aclassic example of the possible
tremendous value that can accrue from careful control of vibration. In its
design the vibratory driving force produced by wind was made essentially
zero by the particular open structure used. This new design also made
possible great savings in the structure. As an additional precaution the deck
openings and roadway are arranged to damp any vibration that may occur.
Variation of Parameters. ¿in any of these procedures for tracking down
vibration troubles, it is often helpful to change some element, for example
the mass, and observe how the change affects the vibration levels2. This
technique can be classed as the method of variation of parameters. In other
words, change things and see what happens. The way of "seeing" is, of
course, to use measurements that will give agood basis for judging what has
changed and by how much. In general, one follows alogical guessing proce-
dure. The results of the experiments help one to eliminate or confirm the
various possible sources of vibration effects.

9.5.4. Vibration Isolation. he reduction of the effects of vibration by

isolation is widely usec4(Crede, 1951; Den Hartog, 1956; Vigness, 1965; SAE
Committee G-S, 1962)(This isolation technique is usually illustrated with a
vibrating device mounted on afoundation by means of soft springs or other
resilient device if the isolation system is properly designed, the vibratory
force transmitted to the foundation will be less when the springs are used
than when the device is clamped directly to afoundation. The device itself,
however, will ordinarily vibrate with agreater amplitude when mounted on a
soft mount,lhus it is essential to realize that the isolation is working in only
one direction, that is, the original source of vibratory force is not reduced by
this isolation Of course, if the foundation is vibrating as a result of some
other driving force, one can reduce the effects of the vibration on adevice by
suspending it on asuitable soft mount. Some scientific instruments must be
isolated in this way from building vibrations in order to operate satisfactorily.
Many commercial vibration isolators, or shock mounts, are available, and
the manufact,urers of these mounts usually supply information for their
proper use.lt
- is most important in applying isolators to avoid having the
natural frequency of the mass of the device and the resilient suspension be
nearly the same as the frequency of the driving force?When such acondition
occurs, the transmitted vibration may be greater with the use of isolators than
without. A frequency analysis of the vibration, which gives the component
frequencies of the driving force, and aknowledge of the mechanical constants
should make it possible to avoid this simple resonance effect.

228
 Supports should be located to avoid cross coupling from one mode of
vibration to another. Such a requirement ordinarily means that the line of
action of the support should pass through the center of gravity of the device
being supported,'
The foundation, the isolating suspension system, and the supported struc-
ture will have, individually and in combination, resonant modes at frequen-
cies higher than the first natural resonance. Sometimes these higher modes
cause trouble, because the isolation is reduced from that normally expected
(Plunkett, 1958).
l'he usual commercial vibration isolators include sufficient damping so
that effects of the higher-order resonances in the isolator are not seriou-e But
the isolation is usually significantly less at high frequencies than one would
expect on the basis of the simple idea of aweight supported on aspring.
Torsional vibration is isolated by the use of flexible couplings, flexible
shafts, and belts. These, too, include some damping, and they also introduce
resonant modes of torsional vibration in conjunction with the rotational
inertia of the coupled system.
Multiple isolators need careful design in order to be effective. When two
isolator units are used in cascade, serious effects that interfere with satisfac-
tory isolation may occur ( Skudrzyk, 1959).

9.5.5 Maintenance. When maintenance of proper performance or accept-

able noise and vibration levels is the goal, symptoms are used as aguide to
discover the source of any trouble that may develop and to decide on the
remedy. Before these symptoms are reviewed, it is also helpful to keep in
mind the many ways that machine performance is affected by changes that
occur with time. A systematic classification of the sources of these changes
should serve to point up the many possibilities that exist. They are:
1. Wear
2. Erosion

3. Corrosion
4. Aging
Curing
Crystalization and fatigue
Solidifying of grease or packing
Loss of adhesion or bonding
5. Inelastic behavior
Parts stressed out of shape
Bent parts
Increased tolerances
6. Loosening of fastenings

7. Broken or damaged parts

8. Incorrect or inadequate lubrication
9. Foreign matter
Dirt chips, dust, grit
Contaminants
Humidity
Ice accumulation
Paint and other finishes

229
 The following are recommendations of the International Organization for
Standardization: ( available from ANSI)

ISO/R16-1965 Standard Tuning Frequency

ISO/R31/Part VII-
1965 Quantities and Units of Acoustics
ISO/R131-1959 Expression of the Physical and Subjective Magnitudes of Sound
or Noise
ISO/R140-1960 Field and Laboratory Measurements of Airborne and Impact
Sound Transmission
ISO/R226-1961 Normal Equal- Loudness Contours for Pure Tones
ISO/R266-1962 Preferred Frequencies for Acoustical Measurements
1SO/R354-1963 Measurement of Absorption Coefficients in a Reverberation
Room
ISO/R357-1963 Power and Intensity Levels of Sound or Noise
ISO/R 362-1964 Measurement of Noise Emitted by Vehicles
ISO/R389-1964 Reference Zero for Pure-Tone Audiometers
ISO/R454-1965 Relation Between Sound Pressure Levels of Narrow Bands of
Noise in aDiffuse Field and in aFrontally-Incident Free
Field for Equal Loudness
ISO/R495-1966 Preparation of Test Codes for Measuring the Noise Emitted by
Machines
ISO/R507-1970 Describing Aircraft Noise Around an Airport
ISO/R532-1966 Calculating Loudness Level
ISO/R717-1968 Rating of Sound Insulation for Dwellings
ISO/R1761-1970 Monitoring Aircraft Noise Around an Airport

The following are st andards of the International Electrotechnical Commission:

(available from ANSI)

IEC/50-08(1960) International Electrotechnical Vocabulary Group 08:Electro-

Acoustics
IEC/118(1959) Measurements of the Electro -
Acoustical Characteristics of
Hearing Aids
IEC/123(1961) Sound Level Meters
IEC/124(1960) Rated Impedances and Dimensions of Loudspeakers
IEC/126(1961) IEC Reference Coupler for the Measurement of Hearing Aids
Using Earphones Coupled to the Ear by Means of Ear
Inserts
IEC/177(1965) Pure Tone Audiometers for General Diagnostic Purposes
IEC/178(1965) Pure Tone Screening Audiometers
IEC/179(1965) Precision Sound Level Meters
IEC/184(1965) Specifying the Characteristics of Electromechanical Transducers
for Shock and Vibration Measurements
IEC/200(1966) Measurement of Loudspeakers
IEC/222(1966) Specifying the Characteristics of Auxiliary Equipment for
Shock and Vibration Measurement
IEC/225(1966) Octave, Half-Octave and Third-Octave Band Filters Intended for
the Analysis of Sounds and Vibrations
IEC/263(1968) Scales and Sizes for Plotting Frequency Characteristics
IEC/303(1970) IEC Provisional Reference Coupler for the Calibration of Ear-
phones Used in Audiometry
IEC/327(1971) Precision Method for Pressure Calibration of One-inch Standard
Condenser Microphones by the Reciprocity Technique

American Society for Testing and Materials, 1916 Race St., Philadelphia, PA 19103

Methods of Test for.

C367-57 Strength Properties of Prefabricated Architectural Acoustical

Materials
C384-58 Impedance and Absorption of Acoustical Materials by the Tube
Method

271
 Air leaks
Hydraulic leaks
3. Impact vibration and rattles
Parts colliding
Broken or loose pieces
Electromagnetically driven loose pieces
Water hammer
Surge
In addition to position, frequency, and character of the vibration, timing
may also furnish an important clue to the nature of the difficultyçpere,
stroboscopic observation with a photoelectric pickoff to trigger the strobo-
scope can be helpfuPas illustrated by the cam and follower study previously
mentioned.
When stroboscopic observation is not possible, the vibration signal may be
observed on an oscilloscope with timing supplied by the photoelectric pick-
off.

9.5.6 Conclusion. If a simple solution is not obvious, the quantitative

results of measurements are often essential elements in the efficient analysis
and solution of the problem. As various control procedures are used, vibra-
tion measurements can show the progress being made and when the attack on
the vibration problem must be shifted from one form or place to another.

REFERENCES
AIMA ( 1971-72), Performance Data, Architectural Acoustical Materials, Bull. # 31,
Acoustical and Insulating Materials Association, 206 West Touhy Avenue, Park
Ridge, Illinois, 60068.
American Foundrymen's Society ( 1966), Foundry Noise Manual, American Foundry-
men's Society, Des Plaines, III., 2nd edition.
AIHA ( 1966), Industrial Noise Manual, American Industrial Hygiene Association, 14125
Prevost, Detroit, Michigan 48227, 2nd edition.
ASHRAE ( 1967), ASHRAE Guide and Data Book, Systems and Equipment, American
Society of Heating, Refrigerating and Air-conditioning Engineers, Inc., 345 East 47th
Street, New York, N. Y. 10017. •
L. L. Beranek, ed. ( 1971), Noise and Vibration Control, McGraw-Hill., New York.
L. L. Beranek, cd ( 1960), Noise Reduction, McGraw-Hill, New York.
British Internal Combustion Engine Research Association ( 1958), A Handbook on
Torsional Vibration, University Press, Cambridge, England.
C. E. Crede ( 1951), Vibration and Shock Isolation, John Wiley 8c Sons, Inc., New York.
J. P. Den Hartog ( 1956), Mechanical Vibrations, McGraw-Hill, New York.
C. Duerden ( 1970), Noise Abatement, Butterworths, London.
W. Furrer ( 1964), Room and Building Acoustics and Noise Abatement, Butterworth,
London.
C. M. Harris, cd ( 1957), Handbook of Noise Control, McGraw-Hill, New York.
C. M. Harris and C. E. Crede ( 1961), Shock and Vibration Handbook, McGraw-Hill, New
York.
E. M. Kerwin, Jr., ( 1959), " Damping of Flexural Waves by a Constrained Viscoelastic
Layer," Acoust Soc Am, Vol 31, #5, July, pp 952-962.
V. O. Knudsen and C. M. Harris ( 1950), Acoustical Designing in Architecture, John
Wiley & Sons, Inc., New York.
G. Kurtze ( 1964), Physik und Technik der Larmbekampfung, G. Braun, Karlsruhe.
B. J. Lazan ( 1968), Damping of Materials and Members in Structural Mechanics,
Pergamon Press, New York.
P. Lord and F. L. Thomas, ed. ( 1963), Noise Measurement and Control, Heywood &
Company Ltd, London.
J. N. MacDuff and J. R. Curreri ( 1958), Vibration Control, McGraw-Hill, New York.
R. T. McGoldrick ( 1957), A Vibration Manual for Engineers, US Navy, PB131785,
December, 2nd edition.

231
 SAE Sound Level Committee
J6a Ride and Vibration Data Manua
J336 SAE Recommended Practice, Sound Level for Truck Cab
Interior
J366 SAE Recommended Practice, Exterior Sound Level for Heavy
Trucks and Buses
J377 SAE Standard, Performance of Vehicle Traffic Horns
1672a SAE Standard, Exterior Loudness Evaluation of Heavy Trucks
and Buses
J919 SAE Recommended Practice, Measurement of Sound Level at
Operator Station
J952a SAE Standard, Sound Levels for Engine Powered Equipment
J986a SAE Standard, Sound Level for Passenger Cars and Light Trucks
J994 SAE Recommended Practice, Criteria for Backup Alarm Devices

Industry Groups
Air Conditioning and Refrigeration Institute (
ARI), 1815 North Fort Meyer Drive,
Arlington, Virginia 22209
ARI 443-66 Rooms Fan-Cool Air Conditioner
ARI 270-67 Sound Rating of Outdoor unitary Equipment

Air Diffusion Council (

ADC), 435 North Michigan Ave., Chicago, III 60611
AD-63 Measurement of Room-to- Room Sound Transmissions Through
Plenum Air Systems
1062-R2 • Equipment Test Code
Air Moving and Conditioning Association (
AMCA), 205 West Touhy Avenue, Park
Ridge, Ill 60068
Bulletin 300-67 Test Code for Sound Rating
Bulletin 301-65 Method of Publishing Sound Ratings for Air Moving Devices
Bulletin 302-65 Application of Sone Loudness Ratings for Non- Ducted Air
Moving Devices
Bulletin 303-65 Application of Sound Power Level Ratings for Ducted Air
Moving Devices
Publication 311-67 AMCA Certified Sound Ratings Program for Air Moving Devices

American Gear Manufacturers Association (

AGMA), One Thomas Circle, Washington,
D.C. 20005
295.02 AGMA Standard Specification for Measurement of Sound in
High Speed Helical and Herringbone Gear Units, November
1965.
The Anti-Friction Bearing Manufacturers Association, Inc., 60 East 42nd Street, New
York, N.Y. 10017
AFBMA Standard, Section 13, Roller Bearing Vibration and
Noise
Association of Home Appliance Manufacturers, 20 North Wacker Drive, Chicago, Ill
60606
Standard RAC- 2-SR Room Air Conditioner Sound Rating, January 1971

Compressed Air and Gas Institute (

CAGI), 122 East 42nd Street, New York, N.Y. 10017

(see ANSI S5.1-1971)

National Electrical Manufacturers Association (
NEMA), 155 East 44th St., New York,
N.Y. 10017
SM33-1964 Standards Publication, Gas Turbine Sound and Its Reduction
MG1 Motors and Generators
Section 4.3.2 Method of Measuring Machine Noise
TR1-1963 Transformers, Regulators and Reactors ( Sections 9-04 and
9-05)

1
273
 Chapter 10

Some Case Histories

In order to illustrate some of the procedures given in this book, we shall
describe in this chapter how some industrial noise problems might be hand-
led. They are taken from actual experience, and where instruments are
mentioned, the latest equipment is named, although some of the instruments
actually used were earlier models. The principles and techniques illustrated
remain unchanged, and the slight departure from authenticity is made up for
bv the greater usefulness of reference to acurrent instrument.

10.1 NOISE FROM AN AIR COMPRESSOR.

Engineers in agroup of offices were annoyed by an intense low-frequency
noise whenever an air compressor in another part of the same building was
running. The noise was most intense when the office windows were open; the
air intake for the compressor was about 50 ft away on the near side of the
building. Furthermore, the noise level varied markedly in the office ;that is, in
the middle of the office the noise was hardly noticeable, but near the
windows or the door on the opposite wall the noise was loudest. This
standing-wave pattern was confirmed by a quick check on a sound-level
meter, with weighting control in the flat position, to show that the maxima
were about 162 inches apart. When the windows were closed, the maxima
were not obvious because of other background noises. With one or more
windows open, the pattern was relatively unaffected by opening or closing
the door.
The obvious explanation of this behavior was that the pulses produced at
the air intake were propagated through the windows and excited aresonant
mode of the office. But one of the engineers suggested that the result might
be produced in a different fashion. He suggested that the driving force could
be a vibration transmitted through the building and that the windows needed
to be open in order for the room resonance to coincide with the frequency of
the driving force. ( He admitted this explanation was far fetched.)
In order to decide what to do about the problem, some simple measure-
ments were made. Since the annoying noise was low in frequency, it was
decided to use the combination of a Type 1560-9531 1-inch Ceramic Micro-
phone Set driving a Type 1564 Sound and Vibration Analyzer, which could
measure noise components at frequencies as low as 2.5 Hz. A Type 1562
Sound-Level Calibrator was used to set the level controls. The microphone
was set up at one of the maximum level points near the door, and the
third-octave analysis shown in Figure 10- la was obtained. The strong compo-
nents near 16 and 40 Hz were remeasured with the 1/10-octave bandwidth to
be actually at 14 and 37 Hz. The strong component at 37 Hz appeared to be
the major offending noise.
A Type 1560-P52 Vibration Pickup was connected to the Type 1564. This
combination was calibrated by means of a Type 1557 Vibration Calibrator.

233
a.

1
00 110

90 100

5 90

:<> 60

50 '

-T— ,*. ---- r -

lb 'do •
7,9*;
lO Kio idoo 000
FREQUENCY IN Hz FREQUENCY IN Hz

C.

WTT_TIU; "MT

e
î

5 80

'
0 70

Figure 10-1. Noise reduction possible

through careful acoustical analysis
followed by corrective action.0

.= 0

H i 11 I
•7- -errr .... -
io loo
FREQUENCY IN Hz

Then some exploration of the office floor for a 37-Hz component indicated
nothing significant. An analysis of the vibration of the pump structure
produced the result shown in Figure 10-1b. These measurements satisfied the
one engineer that his explanation based on vibration was probably incorrect.
Since the 37-Hz component was so dominant, it could be assessed as a
pure-tone. The equal-loudness contours of Figure 3-3 indicated that its
loudness level was about 70 phons. Although the rating procedures of
Chapter 3 apply generally to broad-band noise, it is obvious from those
ratings that a significant drop in level for this very loud tone would be
necessary in order to make it acceptable. If it could be lowered to 40 phons,
it would probably be unobjectionable. This drop would require adecrease in
sound-pressure level to 70 dB.
Many possibilities for correcting the annoying condition were considered,
for example, changes in the offices so that they would not resonate at the
troublesome frequency, a change in compressor speed, and rerouting of the
air intake. The solution adopted was to add a pneumatic filter to the air

234
 WALL WITH MANY WINDOWS

1+
BRAID,.
ING MACHINE BENCHES

ANY WINDOWS
WALL WITH

BENCHES

CONCRETE FLOOR AND CEILING ( HEIGHT-

CONCRETE BLOCK I
WALL

()Cave Band- Center Frequency - Hz

Measure-
ment
Position 63 125 250 500 1000 2000 • 4000 8000
Band Levels in dB re 2001 /1n2
1 65 68 73 86 91 93 94 90
2 64 65 68 82 86.5 90 92 88
3 63 63 68 81 87 89 91 87
4 66 76 69 81 86 90 91 87
5 65 67 69 85 86 89 90 87
6 65 72 69 83 86 89 91 88
8 65 69 67 80 84 88 89 86
9 60 62 66 80 84 88 89 86
10 65 69 68 80 85 88 89 87
12 62 60 64 78 83 86 87 84
13 61 60 65 78 81 85 86 83
14 65 70 69 81 85 89 90 87

Microphone on analyzer - 42" above floor.

Figure 10-2. An octave-band analysis of

braiding- machine noise (paragraph 10-21.

intake. With the data on the significant components at hand, it was easy to
design a proper filter. A thirty-foot pipe was already being used from the air
infake to the compressor, and calculations showed that it could form part of
the filter. An air tank was added at the center of this pipe, and the noise was
so reduced that it was no longer troublesome. The noise analysis in the office
after this change is shown in Figure 10-1c.

10.2 BRAIDING-MACHINE NOISE.

When a manufacturer introduced a process of putting a braided nylon
sheath around acable of wires, the employees in the vicinity of the braiding
machine complained about the noise it made. In common with other ma-
chines of this type, this was a broad-band noise source, and the obvious
instrument to use was an octave-band analyzer.
A 1933 Precision Sound-Level Meter and Analyzer was used to analyze the
noise at a number of places around the machine. Some of the results are
shown in Figure 10-2.
A study of the machine indicated that it would be very difficult to modify
it to reduce the level at the operators' ears without reducing the production
rate. The machine was run intermittently for only afraction of the day, and
this intermittency provided some margin to protect the operator. But it was
decided to supply protective earmuffs for the operator as an added margin of
safety. To protect the other workers, the machine was put in aseparate room

235
heavily treated with acoustical tile. In addition, to avoid troubles from open
windows in the summer bypassing the wall barriers, the room was air-condi-
tioned. The air conditioning made it unnecessary to open the windows, and it
made it possible to wear the earmuffs with comfort even in the summer.

10.3 REDUCTION OF FLUTTER IN A TAPE RECORDER.

Measurements of the flutter (variations in tape speed) of atape recorder
showed strong components at 15, 30, and 75 Hz and other minor compo-
nents. Although it was expected that eccentricities in the drive produced
some of this flutter, acheck on the effects of vibration was made.
A Type 1560-P54 Vibration Pickup was mounted on the tape deck and
connected through a Type 1560-P40 Preamplifier to aType 1564 Sound and
Vibration Analyzer. The analyzed vibration showed strong components at 30
and 75 Hz, but very little at 15 Hz. The 30 Hz component corresponded to
the motor speed. Although the motor and attached flywheel had been
balanced before being mounted, the vibration was easily reduced by rebalanc-
ing of these in place. After balancing, the flutter component at 30 Hz was
negligible, and the 75 Hz component was markedly reduced also.
The 15-Hz component corresponded to the capstan speed, and it was the
largest remaining component. The fact that there was no appreciable vibra-
tion at this frequency seemed surprising at first. The flutter could be reduced
significantly by the placing of eccentric weights on the capstan flywheel; but
then the vibration of the tape deck increased markedly at 15 Hz. Obviously,
this flutter component was caused by eccentricity in the capstan, and the
vibration introduced was canceling the effects of eccentricity.
Measurements at the capstan bearing showed that the vibration at 75 Hz
was a maximum there. What was happening here, apparently, was that the
capstan flywheel and shaft structure had avibration resonance at 75 Hz that
was excited either by the fifth harmonic of the capstan rotation frequency or,
more likely, by a combination of amultiple of the motor rotation frequency
and the capstan rotation frequency. Because of the resonance, eery little
energy was required to produce asignificant vibration. This mode could be
reduced significantly by a change in the resonant frequency, but the balanc-
ing of the motor and flywheel had already reduced it so that the 15-Hz
component was the only significant one remaining.

10.4 AN OIL- PUMP PROBLEM.

An oil pump, used in aproduction setup to supply oil at high pressure to a
number of hydraulic presses, was so noisy that the workmen objected to
using it. This pump had been installed to speed up production with new
presses, but the men preferred to use an earlier production method because it
was not then necessary to use the noisy pump. The problem was to find out
what should be done to make the noise less objectionable.
In this example, it was assumed that the pump itself could not be modified
to reduce the noise, since correcting basic design faults would be a major
problem. Errors in alignment or looseness of mounting, as the source of the
high noise levels, however, should be taken into consideration. On that basis,
the apparent procedure was to investigate these possibilities, to measure the
noise produced by the machine, to measure the background noise level, and
then to decide what recommendations should be made.

236
Community Noise Measurement

GR COMMUNITY NOISE MEASUREMENT SETS

SOUND- LEVEL VEHICLE- NOISE

MEASUREMENT SET MEASUREMENT SET

Whether or not your community has This GR set features simplicity of opera- 1567 Sound-Level Calibrator, included
enacted noise ordinances, GR's 1565- tion and calibration so that reliable meas- in the set. The sound-level calibrator and
9905 Sound-Level Measurement Set will urements can be made by law enforce- sound-level meter are protected during
be avaluable asset. The set consists of a ment officers who have had no prior ex- storage and transportation from site to
1565-C Sound-Level Meter for making perience in acoustic measurements. The site by an impact-resistant, padded car-
noise measurements from 30 to 130 dB, set includes aGR 1565-9015 Sound-Level rying case.
a 1567 Sound- Level Calibrator for rou- Meter that measures in the range of 50 to
tine calibration of the meter, awind- 140 dB. Since most vehicle-noise meas-
Set Specifications
screen, and arugged carrying/storage case. urement regulations specify microphone
location during measurement, with the 1565-9906 Vehicle-Noise Measurement
The 1565-C Sound- Level Meter has been observer out of the sound field, the set Set: Consists of 1565-9015 Sound-Level
specifically designed for community noise includes aremote microphone, extension Meter, 1567 Sound-Level Calibrator, mi-
measurements. It meets the American cable, and tripod. crophone, adaptor for microphone, tripod,
extension cable ( 100 ft), windscreen, carry-
National Standards Institute ( ANSI) Type
2Specifications, which detail the required Calibration of the sound-level meter for ing case, batteries, screwdriver for calibration
instrument performance. The 1565-C is optimum measurement accuracy is easily adjust, and minature phone plug that con-
performed in just afew seconds with the nects to sound-level-meter output.
easy to operate and requires no prior
knowledge of acoustics. Its rugged ceram-
ic microphone, high-impact case, and all
solid-state construction make it extremely
durable, even under the severe operating
conditions encountered during field meas-

f
urements.

Set Specifications
1565-9905 Sound- Level Measurement
Set: Consists of 1565-C Sound- Level
Meter, 1567 Sound- Level Calibrator,
windscreen, carrying case, batteries,
screwdriver for calibration adjust, and
miniature phone plug that connects to
sound-level-meter output.

ser

Sound Level Measurement Set Vehicle-Ncuse Measurement Set

279
 100
I
MICROPHONE -16" FROM PUMP SHAFT

OCTAVE - BAND
"k 90
— LEVELS

o
C•1

• 80

DISCRETE COMPONENTS
o-w----?"- MEASURED ON TYPE
tio 70 1568 WAVE ANALYZER
o 1
4 „BACKGROUND NOISE
/---

60
/

20 50 100 200 500 1,000 2Poo 5,000 10,000

FREQUENCY IN HERTZ

Figure 10-3. Frequency analysis of the noise produced by apump. Levels measured
with the octave-band analyzer are shown together with components measured on the
Type 1568 Wave Analyzer. Background band levels are shown by horizontal dashed
lines; solid horizontal lines represent pump noise plus background.

A preliminary survey around and over the structure but some 5 ft away
was made in octave-bands. As expected, there was no obvious directional
pattern.
The first measurement was made close to the pump. The microphone, only
16 in. from the pump shaft was on the 1933, which in turn was set on an
empty cardboard packing case on the concrete floor. This first position was
selected at this point to make certain that the background noise from other
machines would not obscure any significant components.
With the pump turned on, the output from the analyzer was monitored by
the pair of earphones. Listening to the output of the various bands showed
that the noise in the 1000 and 2000-Hz bands was the dominating part of the
annoying, loud noise heard from the machine.
The complete analysis was made at this point as shown on the graph of
Figure 10-3. Then the pump was turned off, and the background noise was
analyzed. In all frequency bands but the lowest, this background noise was so
low that it could be neglected. It was obvious from this analysis that most of
the noise was in the range from 200 to 2800 Hz.
There were no apparent characteristic, pitched sounds in the noise heard
from the machine, but it could be expected that some would be present. Just
to make sure that nothing important would be overlooked, an analysis of the
noise was also made with the Type 1568-A wave analyzer on the output of
the 1933. The only discrete components ( definite peaks in response as the
analyzer was tuned) that were observed are shown on the graph. Of these com-
ponents, the one at 205 Hz was the basic pumping rate of seven times the rota-
tional speed. A comparison of the levels from this analysis with that in octave-
bands showed that most of the energy in the range from 200 to 700 Hz was
from discrete components, but above that the noise was generally unpitched.
The next step was to use avibration test to find out if the mounting was
satisfactory. The vibration pickup and control box were connected to the
analyzer. Exploration with the pickup and the analyzer showed the following
behavior. The pump itself was vibrating most strongly; the high-frequency

238
and low-frequency components were all present. The driving motor was not
vibrating seriously. The storage tank vibrated most strongly at low frequen-
cies. As the probe was moved about the mounting base toward the concrete
floor the amplitude of motion decreased. At the floor the motion was
insignificant. This vibration test confirmed that the mounting was not faulty.
The final measurements were octave-band analyses at anumber of points 5
ft from the pump and one point 12 ft away. The results of these analyses are
shown in the data sheet of Figure 10-4.

WALL

PUMP

A
€0- 5' 5 — gel°
ELEV."2' ELEV " 2'

5'
. 5; ELEV . 5'6'
Dj
ELEV.'3 . DISTANCE MEASURED
DIRECTLY TO PUMP
ELEVATIONS WITH
12 TO PUMP RESPECT TO FLOOR
E
--> ELEV 3'

MICROPHONE OVER- OCTAVE BANI)CENTER FREQUENCY

SOURCE
LOCATION ALL 63 125 250 500 1000 2000 4000 8000
1—
A Pump + Bkgd* 86 72 76 78 76 80 78 81 76 74

A Bkgd 76-78 72 7E 72 65 81 5E 58 62 63

P Pump + Bkgd 89 72-76 74 71 81 85-88 8/-93 '32-83 76 75

B Bkgd 76 70-74 SC 62 6. 5?. 64 70 S6

C Pump + Bkgd 88 74-78 7 C-71 80 84 82 82-83 77-78 75

C Bkgd 76-78 72-74 ^2- 74 60 57 54 56 62 62

D. Pump + Bkgd 87-88 72- 76 7 5-77 82 32 .76 22 74 72

E Pump + Bkgd 84-85 70 74 74 72 80 75 76 72 71

'Bkgd Background

Figure 10-4. A diagram of the several positions used in making octave- band analyses of
pump noise. Results obtained at the various locations are given in the table.

The nearest workmen were about 7 feet from the pump, so that the levels
at 5 feet were nearly representative of the conditions they encountered. A
comparison of the levels from the pump with the background data and with
the speech-interference criteria given in Chapter 3 indicated that a 20-dB
reduction in noise level in the bands from 250 to 2000 Hz would have been
desirable.
Therefore, as a solution to the problem, the following suggestions were
made:
One possible solution is to use a different pump based on a principle of
operation that produces less noise as abyproduct.
Another possible solution is to enclose the whole pump in atight housing
with lined ducts for air ventilation. The housing should be treated on the
inside with acoustic absorbing materials.

239
 A third solution is to move the pump to another location outside the
working area, and this solution was adopted. The pump was moved to a
nearby boiler room.
The use of earplugs, sometimes a solution to noise problems, was not
adopted here because of the need for communications and the reluctance of
personnel to wear such devices except as alast resort.
What had been accomplished by these measurements? First, they had ruled
out the possibility of asimple solution, such as isolating the whole structure
by vibration mounts, putting flexible couplings in the pipe lines, or using
acoustic baffles. Second, they provided the data needed for a preliminary
design of a housing, so that its probable cost could be weighed against other
possible solutions. In short, these measurements provided the necessary data
for adecision by management.

10.5 SUMMARIES OF SOME EXAMPLES FROM THE LITERATURE.

As an example of vibration isolation, Miller and Dyer ( 1958) report tests
on the vibration levels induced in a concrete floor by a printing press. When
used with 1/4-in, cork mounting, the press produced such high vibration
levels in a concrete floor that the noise in the engineering offices below was
excessive. Replacement of these mounts by a much thicker felt mount
reduced the vibration amplitudes in the floor to about 1/10 the previous
values in the important frequency range of 75 to 600 Hz.
As an example of detection of resonant modes of vibration, Austen and
Priede ( 1958) found excessive vibration of the valve cover and the timing
cover of a diesel engine. They found, by exploring with a vibration pickup
and by making a narrow-band analysis of the vibration signal, that the valve
cover was vibrating strongly at 1150 Hz, while the timing cover had strong
components of vibration at 760, 1350, and 2800 Hz. Replacing the cast
aluminum covers by "deadened" covers eliminated this excessive vibration.
Feinberg ( 1965) analyzes an interesting example of agyrocompass mount-
ed in a conventional vibration-isolation system. The performance of the
gyrocompass was unsatisfactory when subjected to vibration in the frequency
range from 220 to 350 Hz. The criterion for acceptable vibration levels was
determined by vibrating the gyrocompass as a function of frequency and
observing the level at which the performance became unacceptable. This
vibration tolerance showed a minimum of 2 cm/s 2 rms ( 66 dB re 10 -5 m/s 2
rms) in the frequency range from 240 to 350 Hz. In order to achieve the
required low vibration level in this frequency range atuned two-degrees-of-
freedom filter was designed. Compared to the original simple isolation sys-
tem, the resulting vibration levels at frequencies below 150 Hz were generally
higher, but the criterion level was not exceeded, and the levels in the sensitive
region from 240 to 350 Hz were, with the new system, sufficiently low that
the criterion was satisfied everywhere.
Judd ( 1969) describes methods of controlling furnace noise. He shows the
result of modifying burner tips to reduce the effects of firebox resonance as
observed in an octave-band analysis of the noise. The very high levels in the
octave bands from 2000 to 8000 Hz were reduced by the use of amixer with
a larger spud to permit a lower fuel gas pressure. These high-frequency levels
were also reduced by the use of enclosures and intake silencers.
Mills ( 1969) shows how small but significant reductions in noise in a
textile weaving mill were obtained by the acoustical treatment of the ceiling,

240
walls and floor, by replacement of metal with nylon pinion gears in the drive
assembly, and by other minor modifications of the looms. The important
reductions in level were obtained in the octave bands from 250 to 8000 Hz.
Judd and Spence ( 1969) show how significant reductions in noise from
electric motors can be obtained by selection of the proper type, by the use of
a treated fan shroud, or by full enclosures. The differences in level were
mainly in the octave bands above 250 Hz, except for the full enclosure where
20- to 25-dB reductions in level were obtained over the full audible range.
Torpey ( 1969) describes how power-generating equipment can be made
quieter by avariety of techniques. Among these are sound-attenuating ducts
for air intake and exhaust and covering surfaces with sound-absorbing mater-
ials. His experiments were based mainly on measurements of the noise with
octave or 1/3-octave-band analyzers. He used an analyzer with an even
narrower band to find that avery high 250-Hz octave-band level was caused
by the tone produced by agenerator cooling fan with afundamental blade
pulsing frequency in that band.
Hudson and Crocker ( 1970) show the effects of abarrier in reducing noise
from an air-cooled refrigerant compressor. Noise reductions of from 7.5 to 16
dB(A) were achieved, and they proved adequate in satisfying complaints.
Excessive vibration of a cooling tower ( Buscarello, 1968) was traced to
tower resonance excited by impulses from rotating fan blades. Stroboscopic
observation showed that the tip of each fan blade rose about 3in. as it passed
over a nearby drive shaft. The resulting impulses occurred 900 times per min,
which was the resonant frequency of the tower. The excessive vibration was
avoided by stiffening the tower to raise its resonant frequency.
Many examples of noise or vibration reduction are given in Chapter 11 of
the All-IA Industrial Noise Manual (1966), in Chapter 7cf the Foundry Noise
Manual ( American Foundrymen's Society, 1966), in Harris ( 1957) and in
Harris and Crede ( 1961).
REFERENCES
American Foundrymen's Society ( 1966), Foundry Noise Manual, American Foundry-
men's Society, Des Plaines, Ill, 2nd edition.
AIHA ( 1966), Industrial Noise Manual, American Industrial Hygiene Association, 14125
Prevost, Detroit, Mich 48227, 2nd edition.
A.E.W. Austen and T. Priede ( 1958), " Origins of Diesel Engine Noise," Proceedings of
the Symposium on Engine Noise and Noise Suppression, London, Institution of
Mechanical Engineers, Oct, pp 19-32.
L. L. Beranek, ed. ( 1971), Noise and Vibration Control, McGraw-Hill, New York.
R. T. Buscarello ( 1968), " Practical Solutions for Vibration Problems," Chem Eng, Vol
75, Aug 12, pp 157-166.
M. Feinberg ( 1965), " New Methods Simplify Analysis of Vibration Isolation Systems,"
Part I, Mach Des, Vol 37, #18, Aug 5, pp 142-149.
C.M. Harris, cd ( 1957), Handbook of Noise Control, McGraw-Hill, New York.
C. M. Harris and C. E. Crede ( 1961), Shock and Vibration Handbook, McGraw-Hill, New
York.
R. R. Hudson and M. J. Crocker ( 1970), " The Practical Problems Associated with
Quieting a Refrigerant Compressor," Applied Acoustics Vol. 3 #4, Oct, pp 251-258.
S. H. Judd ( 1969), " Engineering Control of Furnace Noise," Amer Indust Hyg Assoc J,
Vol 30, #1, Jan-Feb, pp 35-40.
S. H. Judd and J. A. Spence ( 1969), " Noise Control for Electric Motors," Amer Indust
Hyg Assoc J, Vol 30, #6, Nov-Dec, pp 588-595.
L. N. Miller and I. Dyer ( 1958), " Printing Machine Isolation," Noise Control, Vol 4, #4,
July, pp 21-23.
R. O. Mills ( 1969), " Noise Reduction in aTextile Weaving Mill," Amer Indust Hyg Assoc
J, Vol 30, #1, Jan-Feb, pp 71-76.
P. J. Torpey ( 1969), " Noise Control of Emergency Power Generating Equipment," Amer
Indust Hyg Assoc J, Vol 30, #6, Nov-Dec 1969, pp 588-595.

241
Appendixes

I Decibel Conversion Tables 244

II Chart for Combining Levels of Uncorrelated Noise Signals . 252

Ill Table for Converting Loudness to Loudness Level 253

IV Vibration Conversion Charts 254

V Definitions 256

VI Words Commonly Used to Describe Sounds 269

VII Standards and Journals 270

VIII Hearing- Loss Statutes 275

IX Catalog Section 278

243
 GR 1562-A and 1562-Z MICROPHONE
and AUDIOMETER CALIBRATORS

• 125 top 2000 Hz

• !A:1.3-dB accuracy
• fits many microphones
• approved by Bureau of Mines

GR 1562-2
Audiometer GR 1562-A
Calibration Set
,i00)Sound-Level Calibrator

specifications

1565-2 AUDIOMETER CALIBRATION SET

1562-A SOUND- LEVEL CALIBRATOR
Supplied: 1565 Sound Level Meter, 1562 Sound- Level Celrbrator, earphone
oupler , spare batteries st orage case Acoustic Output: FREQUENCIES 125, 250.500, 1000, and 2000 Hr. 12%.
Mechanical: DIMENSIONS ( wshxd) 11.25s4.25.10 in ( 286x108x254 SOUND PRESSURE LEVEL, 114 dB re 20 pN/rnT, accuracy at 23 °C and
um,L WEIGHT 51512 3kg/ net. 12 11,16 kg) shipping. 760 elm Hg is, tor WE 640AA or equivalent microphone, 7:0.3 dB at 500 Hz
and 710.5 dB at other frequencies and, for other microphones. /10.5 dB at 500
EARPHONE COUPLERS Hr and ±0.7 dB at other frequencies Temperature coefficient of output
sound level is 0to - 0012 dBP°C, pressure correction chart supplied.
Electrical Output: 1 V ±2% behind 6 kit. flat 2
-2%with < 0.5% distortion.
available at phooe tack.
Environmental: TEMPERATURE 0 to • 40°C operating. - 24 to 055°C
non operating HUMIDITY 95% RH and + 30°C. VIBRATION 0.03 in.
from 10 to 55 Hz. SHOCK 30G, 11 ms. DROP. 30m.
Supplied: Carrying case, adaptors for 5/8. and 1- in. microphones ( lits 1.1/8-
in. microphones:without adaptor), battery .
Available: 1560.9561 Coupler Adapt•
or Set. for coupling 1562 to 1/8. 1/4,
and 1/2- in. microphones.
Power: Battery operated 19 V. Burgess PM6 or equal). 120 huse.
GR 1560-P131 and -P82 GR 1560 P83
Mechanical: DIMENSIONS. 5 in. 1128 mm) long s 2.25 in. ( 57 mm) dia.
WEIGHT 11610.5 kg) net.4 lb 11.9 kg) shipping.
Typa: ANSI Type 1. VOLUME GR 9A I modified version of
6 crn 3 including NBS type 9-A). VOLUME. One r
Ipt lon
equivalent volume of 5.642 cm 3 including volume
15629701 1562-A Sound- Leval Calibrator
microphone AXIAL added by microphone. AXIAL
HOLDING FORCE HOLDING FORCE 450 1560.9561 Coupler Adaptor Set, adapts 1562 A to 1/8.
500 grams, grams nominal 1/4. and 1/2 in microphone
Frequency: 125 Hz to 8 kHz audiometric frequencies, re- 8410 3000 Battery. for 1562-A lutes IT replacement
sponse is equal to NBS9A Coupler within 1dB to 11565-2 AudionteMr Calibration Sat
4 kHz and 1.5 dB to 8 kHz when used with TDH- 1565-9900 with 1560-P82 Earphone Coupler
39 or TDH.49 earphone in MX-41/AR earcushion. 1565-9901 with 1560-P83 Earphone Coupler
Mechanical: 1560-P81 and -P82: DIMENSIONS Coupler. 2.25 in dia is 1.06 1560.9681 1560-P81 Earphone Coupler. ANSI type 1.
M. high ( 57s27 nms), over-all lwxhiecl), 2.2503,13 in. 1571(76)(76 rnm). I1/8 in
WEIGHT. 0.5 lb ( 0.3 kg) net, 2lb ( 1kg) shipping. 16804,83: DIMENSIONS: 1560-9682 1560-P82 Earphone Coupler. ANSI type 1,
Coupler. 2.94 in. dia u 1.25 in. high ( 75)(32 trim), oser - all ( wkhxd).
2.94k3.5x3.5 in. 175090c90 mrm IG iIT 0.5 1010.3 kg) net. 2 ' bit kg) 1560-9683
Shipping. 1560 P83 Earphone Coupler. GR type 9A
84 10-9591 Battery Set ror 1565.Z ( uses 1). replacement

GR 1567 SOUND- LEVEL CALIBRATOR

Economical Calibration The 1567 Sound- Level Calibrate ,

is specially designed for you who wish to check instrument sen
sitivity only ) not frequency response). The 1567 has a single-
level 6utput and is well suited for calibrating the 1563 and
1565-8 Sound- Level Meters at 1 kHz.

SPECIFICATIONS

Acoustic Output: FREQUENCY: 1000 Hz. ± 3%. SOUND-PRES-

SURE LEVEL: 114 dB re 20 pN/rTe; accuracy ( at 23° C and
760 mrn Hg) is 00.5 dB for 1565-B, i1 dB for 1563.

Environment: TEMPERATURE: 0 to 55° C operating. Tempera- Mechanical: Cylindrical housing. DIMENSIONS ( die n h):
ture coefficient of sound-pressure level is zero ± 0.01 dB/° C. 2.38x4.44 in. ( 61x113 mm). WEIGHT: 1 lb (0.5 kg) net, 4 lb
(1.9 kg) shipping.
0 to 23"C; - 0.017 o 0.008 clEIrC, 23 to 50° C. Pressure-correc-
lion chart supplied. HUMIDITY: 0 to 95% RH,
Catalog
Suppiied: Carrying case, adaptor for 1565-B, battery. Number DmmrlptIon
Power: Battery operated, using 9-V Burgess 2U6 or equiva- 1567-9701
lent; 10011 of use. 1567 Sound- Level Calibrator
13410-3200 Battery, spare required)

284
by 10 for each time you added 20 dB. For the power ratio, divide the value
from the left-hand power-ratio column by 100 for each time you added 20 dB.
Example — Given: —49.2 dB
—49.2 dB + 20 dB + 20 dB = — 9.2 dB
Pressure ratio: — 9.2 dB —>
.3467 x 1/10 x 1/10 = . 003467
Power ratio: — 9.2 dB —>
.1202 x 1/100 x 1/100 — .00001202

Table I
l: Pressure Ratios to Decibels
For ratios smaller than those in table — Multiply the given ratio by 10
successively until the product can be found in the table. From the number of
decibels thus found, subtract +20 decibels for each time you multiplied by 10.
Example — Given: Pressure ratio = . 0131
.0131 x 10 x 10 = 1.31
From Table II, 1.31 —>
2.34 dB — 20 dB — 20 dB = — 37.66 dB
For ratios greater than those in table — Divide the given ratio by 10
successively until the remainder can be found in the table. To the number of
decibels thus found, add + 20 dB for each time you divided by 10.
Example — Given: Pressure ratio = 712
712 x 1/10 x 1/10 = 7.12
From Table II, 7.12
17.05 dB + 20 dB + 20 dB = 57.05 dB

Use of Tables to Convert Vibration Readings

These decibel tables offer a convenient means of converting decibel vibra-
tion readings obtained with the sound-level meter and vibration pickup into
displacement in inches, velocity in inches per second, and acceleration in
inches per second per second.
Each control box nameplate is inscribed with a conversion table, which
applies when that control box is used with the pickup and sound-level meter
indicated on the nameplate. The conversion figures for the Type 1560-P21B
Control Box are:
English Units Met ic Units
SLM* Vibration SLM + Vibration
Parameter Reading ( dB) (rms) Reading ( dB) (nns)

Acceleration 50 1 in./s/s 80 1 iii/s/s

Velocity 90 1 in./s 120 1m/s
Displacement 120 1in. 90 1mm
*SLM sensitivity set for —40 dB re IV/NMI' (- 60 dB re 1V/µ bar)
+SLM sensitivity set for — 38.1 dB re 1V/N/m 2 (-58.1 dB re 1Vim bar)

N.B. For Types 759-P36 and 1560-P21 Control Boxes, the conversion figures
are different from the above. When these control boxes are used, substitute
values given on the nameplate for those used below to obtain correct
conversion.
NOTE: In Tables Iand II, the term " pressure ratio" is equivalent to the
term "voltage ratio" as used in the following instructions.

245
 To Convert dB SLM Readings
Into RMS Amplitude in In.
1. Note decibel readings of sound-level meter when vibration pickup is in
contact with vibrating surface and control box switch is set at DISPlace-
ment.
2. If reading for Step 1 is below 120 dB: Subtract +20 dB successively
from 120 minus dB reading until the remainder falls within the range of
Table Iof decibel tables. To determine rms amplitude in inches, multiply
the voltage ratio (left-hand column) corresponding to the dB remainder by
0.1 for each time you subtracted 20 dB. Figures obtained are expressed
directly in inches rms amplitude.
If reading for Step 1 is above 120 dB: Subtract +20 dB successively from
dB reading minus 120 dB until the remainder falls within the range of Table I.
To determine amplitude in inches, multiply the voltage ratio (right-hand
voltage ratio column) corresponding to the dB remainder by 10 for each time
you subtracted 20 dB. Figures obtained are expressed directly in inches rms
amplitude.

To Convert dB SLM Readings

Into RMS Velocity in in./s
1. Note dB reading of sound-level meter with vibration pickup in contact
with vibrating surface and control box switch set at VELocity.
2. If reading for Step 1is below 90 dB: Subtract + 20 dB successively from
90 minus dB reading until the remainder falls within the range of Table
I of decibel tables. To determine rms velocity in inches per second,
multiply the voltage ratio (left-hand voltage ratio column) correspond-
ing to the dB remainder by 0.1 for each time you subtracted 20 dB.
The value obtained is velocity expressed directly in inches per second
TMS.

If reading for Step 1is above 90 dB: Subtract +20 dB successively from dB
reading minus 90 until the remainder falls within the range of Table I. To
determine rms velocity in inches per second, multiply the voltage ratio
(right-hand voltage ratio column) corresponding to the dB remainder by 10
for each time you subtracted 20 dB. The value obtained is velocity expressed
in inches per second rms.

To Convert dB SLM Readings

Into RMS Acceleration in in./s/s
1. Note dB reading of sound-level meter with vibration pickup in contact
with vibrating surface and control box switch set at ACCeleration.
2. If reading of Step 1 is below 50 dB: The value obtained from the
left-hand ratio column corresponding to 50 minus dB reading is acceler-
ation expressed directly in inches per second per second rms.
If reading for Step 1is above 50 dB ( maximum 132 dB): Subtract +20 dB
successively from dB reading minus 50 until the remainder falls within the
range of Table I. To determine rms acceleration in inches per second per
second, multiply the voltage ratio (right-hand voltage ratio column) corres-
ponding to the dB remainder by 10 for each time you subtracted 20 dB. The
value obtained is acceleration expressed directly in inches per second per
second rms.

246
Example:
With the vibration pickup placed in contact with some vibrating surface
and the control box switch, let us say, on DISPlacement, areading of 54 dB is
obtained. Then, following outlined procedure:
1. DB reading = 54 dB.
2. 120 — 54 = 66 dB.
66 — (+ 20) — (+20) — (+ 20) = 6 dB remainder.
Voltage ratios corresponding to 6 dB ( left-hand column) equal 0.5012; 20
dB was subtracted from 66 dB three times; therefore 0.5012 should be
multiplied by 0.1 three times.
Result = 0.0005012 or (to 2 significant figures) 0.00050 inch rms ampli-
tude.
Like procedure should be followed for the calculation of velocity or
acceleration.

Acceleration and Velocity Level

in order to convert the readings obtained with the sound-level meter and
vibration pickup system into acceleration level re 10' 3 cmisec 2 ( often called
adB) or velocity level re 10 -6 cm/sec (often called vdB), proceed as follows:
When the conversion figures on the nameplate are:
Velocity 90 dB = 1in./sec
Acceleration 50 dB = 1in./sec 2

add 38.1 dB to sound-level meter reading to get velocity level when the
control box is set to velocity, and add 18.1 dB to sound-level meter reading
to get acceleration level when the control box is set to acceleration.

247
 TABLE I

GIVEN: Decibels TO FIND: Power and Pressure Ratios

TO ACCOUNT FOR THE SIGN OF THE DECIBEL

For positive (+) values of the decibel - Roth For negative (-) values of the decibel- Both
pressure and power ratios are greater than unity, pressure and power ratios are less than unity. Use
Use the two right-hand columns, the two left-hand columns.

Power Pressure
Example-Given: ± 9.1 dB. Find. Ratio Ratio
+0.1 dB 8.128 2.851
-9.1 dB o.leso 0.3508
- dB+
- dB+
4•• •.>
Pressure Power Pressure Power
dB Pressure Power Pressure Power
Ratio Ratio Ratio Ratio Ratio Ratio d' Ratio Ratio
1.0000 1.0000 0 1.000 1.000
.9886 .5623 .3162 5.0 1.778 3.162
.9772 .1 1.012 1.023
.9772 .5559 .3090 5.1 1.799 3.236
.9550 • .2 1.023 1.047
.9661 .9333 .5495 .3020 5.2 1.820 3.311
.3 1.035 1.072 .5433
.9550 .9120 .4 .2951 5.3 1.841 3.388
1.047 1.096 .5370 .2884 5.4 1.862 3.467
.9441 .8913 .5 1.059 1.122
.9333 .8710 .5309 .2818 5.5 1.884 3.548
.6 1.072 1.148 .5248
.9226 .8511 .2754 5.6 1.905 3.631
.7 1.084 1.175 .5188
.9140 .8318 .8 .2892 .5.7 1.928 3.715
1.096 1.202 .5129 .2630
.9016 .8128 .9 1.109 5.8 1.950 3.802
1.230 .5070 .2570 5.9 1.972 3.890
.8913 .7943 1.0 1.122 1.259
.8810 .7762 .5012 .2512 6.0 1.995 3.981
1.1 1.135 1.288 .4955
.8710 .7586 .2455 6.1 2.018 4.074
1.2 1.148 1.318 .4898
.8810 .7413 1.3 .2399 6.2 2.042 4.169
1.161 1.349 .4842 .2344
.8511 .7244 1.4 6.3 2.065 4.266
1.175 1.380 .4786 .2291 6.4 2.089 4.365
.8414 .7079 1.5 1.189 1.413
.8318 .6918 .4732 .4239 6.5 2.113 4.467
1.6 1.202 1.445 .4677
.8222 .6761 1.7 .2188 6.6 2.138 4.571
1.216 1.479 .4624 .2138
.8128 .6607 1.8 1.230 6.7 2.163 4.677
1.514 .4571 .2089 6.8
.8035 .6457 1.9 1.245 2.188 4.786
1.549 .4519 .2042 6.9 2.213 4.898
.7943 .6310 2.0 1.259 1.585 .4467 .1995 7.0 2.239
.7852 .6166 2.1 1.274 5.012
1.622 .4416 .1950 7.1 2.265
.7762 .6046 2.2 1.288 1.660 5.129
.7874 .5888 .4365 .1905 7.2 2.291 5.248
2.3 1.303 1.698
.7586 .4315 .1860 7.3 2.317 5.370
.5754 2.4 1.318 1.738 .4266 .1820 7.4 2.344 5.495
.7499 .5623 2.5 1.334 1.778
.7413 .4217 .1778 7.5 2.371 5.623
.5495 2.6 1.349 1.820
.7328 .5370 .4169 .1738 7.6 2.399 5.754
2.7 1.365 1.862 .4141
.7244 .5248 .1698 7.7 2.427 5.888
2.8 1.380 1.905 .4074 .1660
.7161 .5129 2.9 7.8 2.455 6.026
1.396 1.950 .4027 .1622 7.9 2.483 6.166
.7079 .5012 3.0 1.413 1.995
.6998 .3981 .1585 8.0 2.512 6.310
.4898 3.1 1.429 2.042
.6918 .3936 .1549 8.1 2.541 6.457
.4786 3.2 1.445 2.089
.6839 .3890 .1514 8.2 0.570 6.607
.4677 3.3 1.462 2.138
.6761 .3846 .1479 8.3 2.600 6.761
.4571 3.4 1.479 2.188 .3802 .1445 8.4 2.630 6.918
.6683 .4467 3.5 1.496 2.239
.6607 .3758 .1413 8.5 2.661 7.079
.4365 3.6 1.514 2.291
.6531 .3715 .1380 8.6 2.892 7.244
.4266 3.7 1.531 2.344 .3673
.8457 .4169 3.8 .1349 8.7 2.723 7.413
1.549 2.399 .3631 .1318
.6383 .4074 3.9 8.8 2.754 7.586
1.567 2.455 .3589 .1288 8.9 2.786 7.762
.6310 .3981 4.0 1.585 2.512
.6237 .3548 .1259 9.0 2.818 7.943
.3890 4.1 1.603 2.570 .3508
.6166 .3802 4.2 .1230 9.1 4.851 8.128
1.622 2.630 .3467 .1202
.6095 .3715 4.8 1.641 9.2 2.884 8.318
2.692 .3428 .1175 9.3
.6026 .3631 4.4 1.660 2.754 2.917 8.511
.3388 .1148 9.4 2.951 8.710
.5957 .3548 4.5 1.679 2.818
.5888 .3350 .1122 9.5 2.985 8.913
.3467 4.6 1.698 2.884 .3311 .1096 9.6 3.020
.5821 .3388 4.7 1.718 2.951 9.100
.5754 .3278 .1072 9.7 3.055 9.383
.3311 4.8 1.738 3.020
.5689 .3236 .1047 9.8 3.090 9.550
.3236 4.9 1.758 3.090 .3199 .1023 9.9 3.126 9.772

248
 TABLE I ( continued)

-dB+ - dB+
«41.•

Pressure Power Pressure Power Pressure Power ” Pressure Power

Ratio Ratio dB Ratio Ratio Ratio Ratio du Ratio Ratio

.3162 .1000 10.0 3.162 10.000 .1585 .02512 16.0 6.310 39.81
.3126 .09772 10.1 3.199 10.23 .1567 .02455 16.1 6.383 40.74
.3090 .09550 10.2 3.236 10.47 .154e .02399 16.2 6.457 41.69
.3055 .09333 10.3 3.273 10.72 .1531 .02344 16.3 6.531 42.66
.3020 .09120 10.4 3.311 10.96 .1514 .02291 16.4 6.607 43.65

.2985 .08913 10.5 3.350 11.22 .1496 .02239 16.5 6.683 44.67
.2951 .08710 10.6 3.388 11.48 .1479 .02188 16.6 6.761 45.71
.2917 .08511 10.7 3.428 11.75 .1462 .02138 16.7 6.839 46.77
.2884 .08318 10.8 3.467 12.02 .1445 .02089 16.8 6.918 47.86
.2851 .08128 10.9 3.508 12.30 .1429 .02042 16.9 6.998 48.98

.2818 .07943 11.0 3.548 12.59 .1413 .01995 17.0 7.079 50.12
.2786 .07762 11.1 3.589 12.88 .1396 .01950 17.1 7.161 51.29
.2754 .07586 11.2 3.631 13.18 .1380 .01905 17.2 7.244 52.48
.2723 .07413 11.3 3.673 13.49 .1365 .01862 17.3 7.328 53.70
.2692 .07244 11.4 3.715 13.80 .1349 .01820 17.4 7.413 54.95

.2661 .07079 11.5 3.758 14.13 .1334 .01778 17.5 7.499 56.23
.2630 .06918 11.6 3.802 14.45 .1318 .01738 17.6 7.586 57.54
.2600 .06761 11.7 3.846 14.79 .1303 .01698 17.7 7.674 58.88
.2570 .06607 11.8 3.890 15.14 .1188 .01660 17.8 7.762 60.26
.2541 .06457 11.9 3.936 15.49 .1274 .01622 17.9 7.852 61.66

.2512 .06310 12.0 3.981 15.85 .1259 .01585 18.0 7.943 63.10
.2483 .06166 12.1 4.027 16.22 .1245 .01549 18.1 8.035 64.57
.2455 .06026 12.2 4.074 16.60 .1230 .01514 18.2 8.128 66.07
.2427 .05888 12.3 4.121 16.98 .1216 .01479 18.3 8.222 67.61
.2399 .05754 12.4 4.169 17.38 .1202 .01445 18.4 8.318 69.18

.1371 .05623 12.5 4.217 17.78 .1189 .01410 18.5 8.414 70.79
.2344 .05495 12.6 4.266 18.20 .1175 .01380 18.6 8.511 72.44
.2317 .05370 12.7 4.315 18.62 .1161 .01349 18.7 8.610 74.13
.2291 .05248 12.8 4.365 19.05 .1148 .01318 18.8 8.710 75.86
.2265 .05129 12.9 4.416 19.50 .1135 .01288 18.9 8.811 77.62

.2239 .05012 13.0 4.467 19.95 .1122 .01259 19.0 8.913 79.43
.2213 .04898 13.1 4.519 20.42 .1109 .01230 19.1 9.010 81.28
.2188 .04786 13.2 4.571 20.89 .1096 .01202 19.2 9.120 83.18
.1163 .04677 13.3 4.624 21.38 .1084 .01175 19.3 9.226 85.11
.2138 .04571 13.4 4.677 21.88 .1072 .01148 19.4 9.333 87.10

.2113 .04467 13.5 4.732 22.39 .1059 .01122 19.5 9.441 89.13
.2089 .04365 13.6 4.786 22.91 .1047 .01096 19.6 9.550 91.20
.2065 .04266 13.7 4.842 23.44 1035 .01072 19.7 9.661 93.33
.2042 .04169 13.8 4.898 23.99 .1023 .01047 19.8 9.772 95.50
.2018 .04074 13.9 4.955 24.55 .1012 .01023 19.9 9.886 97.72

.1995 .03981 14.0 5.012 25.12 .1000 .01000 20.0 10.000 100.00
.1972 .03890 14.1 5.070 25.70
.1950 .03802 14.2 5.129 26.30
.1928 .03715 14.3 5.188 26.92 -dB+
.1905 .03631 14.4 5.248 27.54 41,

.1884 .03548 14.5 5.309 28.18 Pressure Power ,” Pressure Power

.1862 .03467 14.6 5.370 28.84 Ratio Ratio «u Ratio Ratio
.1841 .03388 14.7 5.433 29.51
.1820 .03311 14.8 5.495 30.20 3.162 X 10-1 lel 10 3.162 10
.1799 .03236 14.9 5.559 30.90 10 1 10 -2 20 10 10'
3.162)(10 -1 10 -1I 30 3.162X 10 10 5
.1778 .03162 15.0 5.623 31.62 10-i 40 10' 10"
10-2
.1758 .03090 15.1 5.689 32.36
.1738 .03020 15.2 5.754 33.11
.1718 .02951 15.3 5.821 33.88 3.162 X10-5 10-5 50 3.162X 10' 10 5
.1698 .02884 15.4 5.888 34.67 10 -5 10 -6 60 10 5 104
3.162X 10 -4 10 -' 70 3.162X 10 5 10'
.1679 .02818 15.5 5.957 35.48 10-4 10 -5 80 10" 108
.1660 .02754 15.6 6.026 36.31 3.162X10' 10 -9 90 3.162X 10' 109
.1641 .02692 15.7 6.095 37.15
.1622 .02630 15.8 6.166 38.02
.1603 .02570 15.9 6.237 38.90 10-5 10 -1 ° 100 10 5 10 10

249
 TABLE II

GIVEN: { Pressure } Ratio TO FIND: Decibels

POWER RATIOS
To find the number of decibels one-half of the number of decibels thus
corresponding to a given power found.
ratio- Assume the given power ratio
to be a pressure ratio and find the Example- Given:a power ratio of 3.41.
corresponding number of decibels from Find: 3.41 in the table:
the table. The desired result is exactly
3.41-.10.655 dR X 34 5.328 d
Pressure
Ratio .00 .01 .02 .03 .04 .05 . 06 1 . 07 .08 .09
1.0 .000 .086 .172 .257 .341 .424 .506 .588
1.1 .828 .668 .749
.906 .984 1.062 1.138 1.214 1.289
1.2 1.364 1.438 1.511
1.584 1.656 1.727 1.798 1.868 1.938
1.3 2.007 2.076 2.144 2.212
2.279 2.345 2.411 2.477 2.542 2.607 2.671 2.734 2.798 2.860
1.4 2.923 2.984 3.046 3.107 3.167 3.227 3.287 3.346 3.405 3.464
1.5 3.522 3.580 3.637 3.694 3.750 3.807 3.862 3.918
1.6 4.082 3.973 4.028
4.197 4.190 4.244 4.297 4.350
1.7 4.402 4.454 4.506 4.558
4.609 4.660 4.711 4.761 4.811 4.861 4.910 4.959 5.008 5.057
1.8 5.105 5.154 5.201 5.249 5.296 5.343 5.390 5.437
1.9 5.575 5.483 5.529
5.621 5.666 5.711 5.756 5.801 5.845 5.889 5.933 5.977
2.0 6.021 6.064 6.107 6.150 6.193 6.235 6.277 6.319
2.1 6.444 6.361 6.403
6.486 6.517 6.569 6.608 6.649
2.2 6.689 6.729 6.769 6.809
6.848 6.888 6.927 6.966 7.008 7.044 7.082 7.121 7.159 7.197
2.3 7.235 7.272 7.310 7.347 7.384 7.421 7.458 7.495
2.4 7.604 7.632 7.568
7.640 7.676 7.712 7.748 7.783 7.819 7.854 7.889 7.924
2.5 7.959 7.993 8.028 8.062 8.097 8.131 9.185 8.199
2.6 8.299 8.232 8.266
8.833 8.366 9.399 8.432 8.465
2.7 8.498 8.530 8.563 8.595
8.627 8.659 8.691 8.723 8.755 8.787 8.818 8.850 8.881 8.912
2.8 8.943 8.974 9.005 9.036 9.066 9.097 9.127 9.158 9.188
2.9 9.248 9.218
9.278 9.308 9.337 9.367 9.396 9.426 9.455 9.484 9.51s
3.0 9.542 9.571 9.600 9.629 9.657 9.686 9.714 9.743 9.771
3.1 9.827 9.855 9.799
9.883 9.911 9.939 9.966 9.994
3.2 10.021 10.049 10.076
10.108 10.130 10.157 10.184 10.211 10.288 10.264 10.291 10.317 10.344
3.3 10.370 10.397 10.423 10.449 10.475 10.501 10.527 10.553 10.578
3.4 10.630 10.655 10.604
10.681 10.706 10.731 10.756 10.782 10.807 10.832 10.857
3.5 10.881 10.906 10.931 10.955 10.980 11.005 11.029 11.053 11.078
3.6 11.126 11.150 11.102
11.174 11.198 11.222 11.246 11.270 11.293
3.7 11.364 11.317 11.341
11.387 11.411 11.434 11.457 11.481 11.504 11.527 11.550 11.573
3.8 11.596 11.618 11.641 11.664 11.687 11.709 11.732 11.754 11.777
3.9 11.821 11.844 I1.799
11.866 11.888 11.910 11.932 11.954 11.976 11.998 12.019
4.0 12.041 12.063 12.085 12.106 12.128 12.149 12.171 12.192 12.213
4.1 12.156 12.277 12.234
12.298 11.819 12.340 12.361 12.882 12.403
4.2 12.465 12.424 12.444
I2.486 11.506 12.527 12.547 12.568 12.588 12.609 12.629 12.649
4.3 12.669 12.690 11.710 12.730 11.750 12.770 12.790 12.810 12.829 12.849
4.4 12.869 12.889 12.908 12.928 12.948 12.967 12.987 13.006 13.026 13.045
4.5 13.064 13.084 13.103 13.122 13.141 13.160 13.179 13.198 13.217
4.6 13.255 13.274 13.236
13.293 13.312 13.330 13.349 13.368 13.986 18.405 13.423
4.7 13.442 13.460 13.479 13.497 18.516 13.534 13.552 13.570 13.589 13.607
4.8 13.625 13.643 13.661 13.679 13.697 13.715 13.733 13.751 13.768 13.786
4.9 13.804 13.822 13.899 13.857 13.875 13.892 13.910 13.927 19.945 13.962
5.0 13.979 13.997 14.014 14.031 14.049 14.066 14.083 14.100 14.117 14.134
5.1 14.151 14.168 14.185 14.202 14.219 14.236 14.253 14.270 14.287 14.303
5.2 14.320 14.337 14.353 14.970 14.387 14.403 14.420 14.436 14.453 14.469
5.3 14.486 14.502 14.518 14.535 14.551 14.567 14.583 14.599 14.616 14.632
5.4 14.648 14.664 14.680 14.696 14.712 14.728 14.744 14.760 14.776 14.791
5.5 14.807 14.823 14.839 14.855 14.870 14.886 14.902 14.917 14.933 14.948
5.6 14.964 14.979 14.995 15.010 15.026 15.041 15.056 15.072 15.087 15.102
6.7 1.5.117 1.5.133 1.5.148 15.163 15.178 15.193 15.208 15.224 15.239 15.254
5.8 15.269 15.284 1.5.298 15.319 15.328 15.349 15.358 15.373 15.988 15.402
5.9 15.417 15.432 15.446 15.461 15.476 15.490 15.505 15.619 15.534 15.549

250
 TABLE II ( continued)

Pressure
.00 .01 .02 .0S .04 .05 .06 .07 .08 .09
Ratio

6.0 15.563 15.577 15.592 15.606 15.621 15.635 15.649 15.664 15.678 15.692
6.1 15.707 15.741 15.735 15.749 15.763 15.778 15.792 15.806 15.820 15.834
6.4 15.848 15.862 15.876 15.890 15.904 15.918 15.931 15.945 15.959 15.973
6.3 15.987 16.001 16.014 16.028 16.042 16.055 16.069 16.083 16.096 16.110
6.4 16.144 16.137 16.151 16.164 16.178 16.191 16.205 16.218 16.432 16.245

6.5 16.458 16.272 16.285 16.298 16.312 16.325 16.338 16.351 16.365 16.378
6.6 16.391 16.404 16.417 16.430 16.443 16.456 16.469 16.488 16.496 16.509
6.7 16.521 16.534 16.547 16.560 16.573 16.586 16.599 16.612 16.625 16.637
6.8 16.650 16.663 16.676 16.688 16.701 16.714 16.726 16.739 16.752 16.764
8.9 16.777 16.790 16.802 16.815 16.827 16.840 16.852 16.865 16.877 16.890

7.0 16.902 16.914 16.927 16.939 16.951 16.964 16.976 16.988 17.001 17.013
7.1 17.025 17.037 17.050 17.062 17.074 17.086 17.098 17.110 17.122 17.135
7.2 17.147 17.159 17.171 17.183 17.195 17.207 17.219 17.231 17.243 17.255
7.3 17.266 17.278 17.290 17.302 17.314 17.326 17.338 17.349 17.361 17.373
7.4 17.385 17.396 17.408 17.420 17.431 17.443 17.455 17.466 17.478 17.490

7.5 17.501 17.513 17.524 17.536 17.547 17.559 17.570 17.582 17.593 17.605
7.6 17.616 17.628 17.639 17.650 17.662 17.673 17.685 17.696 17.707 17.719
7.7 17.730 17.741 17.752 17.764 17.775 17.786 17.797 17.808 17.820 17.831
7.8 17.844 17.853 17.864 17.875 17.886 17.897 17.908 17.919 17.931 17.942
7.9 17.953 17.964 17.975 17.985 17.996 18.007 18.018 18.029 18.040 18.051

0.0 18.062 18.073 16.083 18.094 18.105 18.116 18.127 18.137 18.148 18.159
8.1 18.170 18.180 18.191 18.204 18.212 18.223 18.234 18.244 18.255 18.266
8.4 18.276 18.287 18.297 18.308 18.319 18.349 18.340 18.350 18.361 18.371
8.3 18.382 18.392 18.402 18.413 18.423 18.434 18.444 18.455 18.465 18.475
8.4 18.486 18.496 18.506 18.517 18.527 18.537 18.547 18.558 18.568 18.578

8.5 18.588 18.599 18.609 18.619 18.629 18.639 18.649 18.660 18.670 18.680
8.6 18.690 18.700 18.710 18.720 18.730 18.740 18.750 18.760 18.770 18.780
8.7 18.790 18.860 18.810 18.820 18.830 18.840 18.850 18.860 18.870 18.880
8.8 18.890 18.900 18.909 18.919 18.929 18.939 18.949 18.958 18.968 18.978
8.9 18.988 18.998 19.007 19.017 19.027 19.036 19.046 19.056 19.066 19.075

9.0 19.085 19.094 1Ç . 104 19.114 19.123 19.133 19.143 19.152 19.162 19.171
9.1 19.181 19.190 19.200 19.209 19.219 19.228 19.438 19.447 19.257 19.266
9.2 19.276 19.285 19.295 19.304 19.313 19.323 19.332 19.342 19.351 19.360
9.3 19.370 19.379 19.388 19.398 19.4e 19.416 19.426 19.435 19.444 19.453
9.4 19.463 19.474 19.481 19.490 19.499 19.509 19.518 19.547 19.536 19.545

9.5 19.554 19.564 19.573 19.582 19.591 19.600 19.609 19.618 19.647 19036
9.6 19.645 19.654 19.664 19.673 19.682 19.691 19.700 19.709 19.718 19.726
9.7 19.735 19.744 19.753 19.762 19.771 19.780 19.789 19.798 19.807 19.816
9.8 19.825 19.833 19.842 19.851 19.860 19.869 19.878 19.886 19.895 19.904
9.9 19.913 19.921 19.930 19.939 19.948 19.956 19.965 19.974 19.983 19.991

Pressure
Ratio 2 1 2 3 4 5 6 7 8 9

10 20.000 20.828 21.584 22.279 22.923 23.522 24.082 24.609 25.105 25.575
20 26.021 26.444 26.848 27.235 27.604 27.959 28.299 28.627 28.943 29.248
30 29.542 29.827 30.103 30.870 30.630 30.881 31.126 31.364 31.596 31.821
40 32.041 32.256 32.465 32.669 32.869 33.064 33.455 33.442 33.625 33.804

50 33.979 34.151 34.320 34.486 34.648 34.807 34.964 35.117 35.269 35.417
60 35.563 35.707 35.848 35.987 36.124 36.258 36.391 36.521 36.650 36.777
70 36.902 37.025 37.147 37.266 37.385 37.501 37.616 37.730 37.844 37.953
80 38.062 38.170 38.276 38.382 38.486 38.588 38.690 38.790 38.890 38.988
90 39.085 39.181 39.276 39.370 39.463 39.554 39.645 39.735 39.825 39.913

100 40.000 -- -- -- -- -- -- -- -- --

251
 Appendix II

Chart for Combining

Levels of Uncorrelated
Noise Signals*
TO ADD LEVELS
Enter the chart with the NUMERICAL DIFFERENCE BETWEEN TWO
LEVELS BEING ADDED. Follow the line corresponding to this value to its
intersection with the curved line, then left to read the NUMERICAL DIF-
FERENCE BETWEEN TOTAL AND LARGER LEVEL. Add this value to
the larger level to determine the total.

Example: Combine 75 dB and 80 dB. The difference is 5dB. The 5-dB line
intersects the curved line at 1.2 dB on the vertical scale. Thus the total value
is 80 + 1.2 or 81.2 dB.

TO SUBTRACT LEVELS
Enter the chart with the NUMERICAL DIFFERENCE BETWEEN TOTAL
AND LARGER LEVELS if this value is less than 3 dB. Enter the chart with
the NUMERICAL DIFFERENCE BETWEEN TOTAL AND SMALLER LEV-
ELS if this value is between 3and 14 dB. Follow the line corresponding to
this value to its intersection with the curved line, then either left or down to
read the NUMERICAL DIFFERENCE BETWEEN TOTAL AND LARGER
(SMA LLER) LEVELS. Sub tract this value from the total level to determine
the unknown l evel.

Example: Subtract 81 dB from 90 dB. The difference is 9 dB. The 9-dB

vertical line intersects the curved line at 0.6 dB on the vertical scale. Thus the
unknown level is 90 — 0.6 or 89.4 dB.

o 4txte,

é-_1 / e 4,<
•ri e/o

o e ‘,4,
z i oe.
La 49e
* le Cks, Fe.
ww
w cr w
2 0
e <el,
la œ Z3
la
cr a 0
La _J l
Ia. /
0 &e,
Là.
0 +6,
a z //
a 4a,
-J
a ,,,, ,to
o
i
La

D
z 3 4
6 78 9 10
II 12 13 14
DECIBELS
NUMERICAL DIFFERENCE BETWEEN TOTAL AND SMALLER LEVELS
• Thls chart Is based on one developed by R.A4usa.

252
Appendix III

Table for Converting

Loudness to
Loudness Level
A simplified relation between the loudness in sones and the loudness level
in phons has been standardized internationally ( ISO/R131-1959). This rela-
tion is a good approximation to the psychoacoustical data and is useful for
engineering purposes, but it should not be expected to be accurate enough for
research on the subjective aspects of hearing.
The relation is
2 ( P-40)/10

where S is the loudness in sones and P is the loudness level in phons.

A table of loudness in sones for loudness levels ranging from 20 to 130
phons in increments of 1 phon, calculated from the above relation, is given
below.
Examples:
Given - loudness level of 72 phons.
Find - in table under "+ 2" in the " 70" row - 9.2 sones.

LOUDNESS IN SONES

Phons 0 ,1 +2 +3 +4 +5 +6 -7 +8 +9

20 .25 27 .29 .31 .33 .35 .38 .41 .44 .47

30 .50 .54 .57 .62 .66 .71 .76 .81 .87 .93

40 1 1.07 1.15 1.23 1.32 1.41 1.52 1.62 1.74 1.87

50 2 2.14 2.30 2.46 2.64 2.83 3.03 3.25 3.48 3.73

60 4 4.29 4.59 4.92 5.28 5.66 6.06 6.50 6.96 7.46

70 8 8.6 9.2 9.8 10.6 11.3 12.1 13.0 13.9 14.9

80 16 17.1 18.4 19.7 21.1 22.6 24.3 26.0 27.9 29.9

90 32 34.3 36.8 39.4 42.2 45.3 48.5 52.0 55.7 59.7

100 64 68.6 73.5 78.8 84.4 90.5 97 104 Ill 119

110 128 137 147 158 169 181 194 208 223 239

120 256 274 294 315 338 362 388 416 446 478

253
 ....:, .Ù C
,1A0 n •
5 eSi > eel

ese „"›.
::etl e
tob
V>, 1
14
oo
u 4-,
r.

_in tzt.
0) «.' .
F
-0,.,„
e {-+c o>
y ee
›, ,.
÷,è

gc
y

...
eL f) c
te 3
) \.°
cos qk

•,•:• ,..„ I. C uv
g E • cj
4C3 -4- â .'

.—.
• .... c4.
cts II) .1-'
Ir vs
O.
cos et'
e

• t e ._.

..5 5 ..e
o 4,.
.W511Ble:, .004111 à ...111:

(I) E , 'et, >

4,
Î Me , c•:e. IN i '
la
IMO. W. .•;: :Nrelledel.U.:eZ >;ctileliekge. 1%;,..:!.•.11.tiew».. .. eueme..
,Ale . ,. le, A.
: ISI ae
12 _aW
:••:.. -414.
.. tiag ...SX•f.noealtà0Bletz*A-e•lanour.e.::,:,K.
..70
tammedu‘zemieumo: ›m:equew..eetze›,•e
,6ffeev
etizieoeoto&.auuhoulamysettautatdegeeiaeleiefeue.
O 41.) 4,4
e.MgaMiesqreiriSiin 74 t:erelineeeNtieleeZ-NMgeileiglí'ee eSnIk:IP egOF
I.., 0 ..-•

y ••••• • 4,
-r Irlr
- rneeig>irr •,'" '..., '' r V e • i
= O. CU gOl
,k§M
e

, :NUUL.4
.•
:. • .. C
a.
V

.
eee 42;
Au.
Le r.
II
eliediU "R ••
ij euresser,

ce o
F»NâMK,

1: 1A0111r4 .
1eKeeidekM12 -
1 ,
,
e aral», ......
Nfekble«réeila ,AM MIIMIle'ee
,•Mi ..dOmesie.e.,eiro.aouip.kz,
eauliitZ* ›
eu;........ .eutiu:.;::-. .
7«ekesesi.2smeu».z.qUIZIQ*elalleZINCVMe.«
,.ine
be i
'elt:kerit
i DIM
II . 101i, Lo
c.1
.., çant:.:«. U l sgeeee.ww.d.eeew
eu o
o ur, 4»
W... mek*WICMItO',,Wigwey%r.kamee>
-• • - . .
wer r
. ele : egniwineer--1 e.e8e17,,,:4,
ve •et,..W..4
iF' '
le4 1.,

...e.,..0.„.-417.- . lrgo....ree-zo,--
01 •
CU
Il
..... arz.z.r.ir

o
zere, qz-ilmetilt.:,.-.L,24:.1-tiez -,,.,:,-ext2leil.e
:eze4ii
-pê,.. eurasce. ;:ellibentare:
.r., ,vt.
ir, tà

--
gl
CI
r
esi
teb

•mewfee:. . imid- 24t,eei.à - -

À0editeng Illtilte
-ct
eel> ‘keMegMbr
„.• .,,,„, 111e, ,
*t4 44gkk .
q
7
.fre
. e
rk • „ ono ,

Cd
ag..1....i.e•-• -
.,OrIlle.1,....4e.

>c ritalre 14g11ig« 'eel!" '1

-'
re. ''''':»>;:lo:ce ...... 4. CZieli. RAtemiustege;
-
*oleo , .desepalery*..14useep.40e en7e.SeQUiliKeetedtêlakileMMi.miF:Uplee:*21211enFetkq
me ,-
eml8MMirlie, .:w*-M->CMVVMWMi.-Vvemcm«uwÇVVM
The charts

102gaitallinelefeW edardiWIMWSEK4Wgelitiffe2iig§1116,12ILW.MOi.:
>
'eM,1
Q) .4i eN>, ee.UYWrgMeNUMbàn,"WbAkIreg>>.>Nbàwamewe
a §, .3 to g 00 .5 .5 00 cf,.5

8
o,
4.o
a o o b°
o
oar
4
55).
Appendix V

Definitions
This section on definitions includes most of the technical terms used in
this handbook. Most of the definitions are selected from the American
National Standard Acoustical Terminology ( S1.1-1960 (R1971)), and those
definitions are marked with an asterisk. They are printed with permission.
Many have been shortened by putting them in the usual dictionary form.
A number of these standard definitions are very technical in order to be
precise. Some readers may find it easier to refer to the discussion in the main
text of this handbook to obtain ageneral underitanding of those terms.
The nonstandard definitions have been adapted especially for this
handbook.
ACCELERATION*
vector that specifies the time-rate-of-change of velocity. Note 1: Various
self-explanatory modifiers such as peak, average, rms are often used. The time
interval must be indicated over which the average (for example) was taken.
Note 2: Acceleration may be ( 1) oscillatory, in which case it may be defined
by the acceleration amplitude (if simple harmonic) or the rms acceleration (if
random), or (2) nonoscillatory, in which case it is designated "sustained" or
"transient" acceleration.
ALIAS
in sampled, equally spaced data, two frequencies are aliases of one another
if sinusoids of those frequencies cannot be distinguished by the sampled
values.
AMPLITUDE DENSITY DISTRIBUTION ( PROBABILITY DENSITY DIS-
TRIBUTION) ( FREQUENCY DISTRIBUTION)
afunction giving the fraction of time that the pressure, voltage, or other
variable dwells in anarrow range.
AMPLITUDE DISTRIBUTION FUNCTION ( DISTRIBUTION FUNCTION)
(PROBABILITY FUNCTION) ( CUMULATIVE FREQUENCY FUNCTION)
a function giving the fraction of time that the instantaneous pressure,
voltage or other variable lies below agiven level.
ANALYZER
a combination of a filter system and asystem for indicating the relative
energy that is passed through the filter system. The filter is usually adjustable
so that the signal applied to the filter can be measured in terms of the relative
energy passed through the filter as afunction of the adjustment of the filter
response-vs-frequency characteristic. This measurement is usually interpreted
as giving the distribution of energy of the applied signal as a function of
frequency.
ANECHOIC ROOM ( FREE- FIELD ROOM)*
a room whose boundaries absorb effectively all the sound incident there-
on, thereby affording essentially free-field conditions.

256
AUDIOGRAM ( THRESHOLD AUDIOGRAM)
agraph showing hearing-threshold level ( HTL) as afunction of frequency.
AUDIOMETER
an instrument for measuring hearing threshold level.
AUTOCOR RELATION
a measure of the similarity of afunction with adisplaced version of itself
as a function of the displacement. The displacement is usually in terms of
time and, when the displacement is zero, the value of the autocorrelation is
equal to the mean square value of the function.
AUTOSPECTRUM ( POWER SPECTRUM)
a spectrum with the coefficients of the components expressed as the
square of the magnitudes.
BAFFLE *
a shielding structure or partition used to increase the effective length of
the external transmission path between two points in an acoustic system as,
for example, between the front and back of an electroacoustic transducer.
COHERENCE.
a measure of the reliability of atransfer function estimate. It is zero when
the transfer function has no statistical validity and unity when the estimate is
not contaminated by interfering noise.
CONFIDENCE LIMITS
the upper and lower values of the range over which a given percent
probability applies. For instance, if the chances are 99 out of 100 that a
sample lies between 10 and 12, the 99% confidence limits are said to be 10
and 12.
CRITICAL SPEED*
aspeed of arotating system that corresponds to aresonance frequency of
the system.
CR OSSCO RRE LAT I ON.
a measure of the similarity of two functions with the displacement
between the two used as an independent variable. The displacement is usually
in terms of time. When the two functions are alike, acrosscorrelation is an
autocorrelation.
CROSS-SPECTRUM
ameasure in the frequency domain of the similarity of two functions.
DATA WINDOW
the interval that includes all the sampled values in a calculation, also the
form of a weighting function that is regarded as multiplying the data that
enters into acalculation.
DEAD ROOM* ( See also ANECHOIC ROOM)
a room that is characterized by an unusually large amount of sound
absorption.
DECAY RATE ( See RATE OF DECAY)
DECIBEL *
one-tenth of a bel. Thus, the decibel is aunit of level when the base of the
logarithm is the tenth root of ten, and the quantities concerned are propor-

257
tional to power. Note 1: Examples of quantities that qualify are power (any
form), sound pressure squared, particle velocity squared, sound intensity,
sound energy density, voltage squared. Thus the decibel is aunit of sound-
pressure-squared level; it is common practice, however, to shorten this to
sound pressure level because ordinarily no ambiguity results from so doing.
Note 2: The logarithm to the base the tenth root of 10 is the same as ten
times the logarithm to the base 10: e.g., for anumber x2 ,logi o° *1 x2 = 10

log i0x2 = 20 log io x. This last relationship is the one ordinarily used to
simplify the language in definitions of sound pressure level, etc.
DEGREES OF FREEDOM ( STATISTICAL)
a measure of stability relating to the number of independent equivalent
terms entering into adistribution.

DIRECTIVITY FACTOR *
1. of a transducer used for sound emission is the ratio of the sound
pressure squared, at some fixed distance and specified direction, to the
mean-square sound pressure at the same distance averaged over all
directions from the transducer. The distance must be great enough so
that the sound appears to diverge spherically from the effective acoustic
center of the sources. Unless otherwise specified, the reference direc-
tion is understood to be that of maximum response.
2. of a transducer used for sound reception is the ratio of the square of
the open-circuit voltage produced in response to sound waves arriving in
a specified direction to the mean-square voltage that would be pro-
duced in a perfectly diffused sound field of the same frequency and
mean-square sound pressure.
Note 1: This definition may be extended to cover the case of finite frequency
bands whose spectrum may be specified. Note 2: The average free-field
response may be obtained, for example,
1. By the use of aspherical integrator
2. By numerical integration of asufficient number of directivity patterns
corresponding to different planes, or
3. By integration of one or two directional patterns whenever the pattern
of the transducer is known to possess adequate symmetry.

DIRECTIONAL GAIN ( DIRECTIVITY INDEX)*

of atransducer, in decibels, is 10 times the logarithm to the base 10 of the
directivity factor.

DISPLACEMENT*
avector quantity that specifies the change of position of abody or particle
and is usually measured from the mean position or position of rest. In
general, it can be represented by a rotation vector or translation vector oi
both.
DISTRIBUTION
See AMPLITUDE DISTRIBUTION FUNCTION.

EARPHONE ( RECEIVER) *
an electroacoustic transducer intended to be closely coupled acoustically
to the ear. Note: The term "receiver" should be avoided when there is risk of
ambiguity.

258
EFFECTIVE SOUND- PRESSURE
(ROOT- MEAN-SQUARE SOUND PRESSURE) *
at a point is the root-mean-square value of the instantaneous sound
pressures, over atime interval at the point under consideration. In the case of
periodic sound pressures, the interval must be an integral number of periods
or an interval long compared to a period. In the case of non-periodic sound
pressures, the interval should be long enough to make the value obtained
essentially independent of small changes in the length of the interval. Note:
The term " effective sound pressure" is frequently shortened to "sound

pressure."
FFT ( FAST-FOURIER TRANSFORM)
any of a number of calculation procedures that yields a set of Fourier
coefficients ( component amplitudes) from atime-series frame with much less
computational effort for large frame sizes than is possible by the classical
approach of successive calculation of each coefficient.

FILTER
a device for separat in g com ponents of a signal on the basis of their
frequency. It allows components in one or more frequency bands to pass
relatively unattenuated, and it attenuates components in other frequency

bands.
FOLDING FREQUENCY
reciprocal of twice the time interval between sampled values. The folding

frequency is equal to its own alias.

FRAME
aset of points or values that are processed as agroup.

FRAME SIZE
the number of sampled values in aframe.
FREE SOUND FIELD ( FREE FIELD) *
a field in an homogeneous, isotropic medium, free from boundaries. In
practice, it is afield in which the effects of the boundaries are negligible over
the region of interest. Note: The actual pressure impinging on an object (e.g.,
electro-acoustic transducer) placed in an otherwise free sound field will differ
from the pressure that would exist at that point with the object removed,
unless the acoustic impedance of the object matches the acoustic impedance

of the medium.
FREQUENCY ( IN CYCLES PER SECOND OR HERTZ)
the time rate of repetition of aperiodic phenomenon. The frequency is the

reciprocal of the period.

FREQUENCY DISTRIBUTION ( STATISTICAL)
see AMPLITUDE DENSITY DISTRIBUTION.

9 quantity that is the acceleration produced by the force of gravity, which

varies with the latitude and elevation of the point of observation. By interna-
tional agreement, the value 980.665 cm/s 2 = 386.087 in./s 2 = 32.1739 ft /s 2

has been chosen as the standard acceleration of gravity.

GAUSSIAN DISTRIBUTION ( NORMAL DISTRIBUTION)

a particular amplitude distribution of fundamental importance in the
theory of probability. Its histogram is the familiar " bell-shaped curve." It

259
 FREQUENCY- RESPONSE 81 SPECTRUM RECORDER ASSEMBLIES

Several GR instruments can be used with the 1521-8 be ordered under a single catalog number and include all
Graphic Level Recorder for automatic plotting of the fre- accessories normally needed. Or the component items
quency response of devices or the frequency spectrum of can be ordered individually to convert existing equipment
acoustic or electrical noise or of a complex electrical into fully automatic recording assemblies.
waveform. Automatic plotting with these instruments re- Custom assemblies of GR analysis equipment and sound
places tedious point-by-point manual methods and pro- and vibration instruments can be built to order to meet a
vides much more information in the form of finer-resolution variety of special requirements.
curves. Listed below are several such assemblies that can

RECORDING ANALYZERS
Catalog

1910-A RECORDING WAVE ANALYZER

The 1910-A is particularly useful in analyzing and re-

cording the frequency components present in mechanical
vibrations, acoustic signals, and in complex electrical sig-
nals including random noise. Its linear frequency scale,
20- Hz to 54-kHz range, three bandwidths (3, 10, and 50
Hz), and 80-d13 dynamic range permit higher-order, closely
spaced and weak components to be found with ease.
The complete assembly Includes the following:
1901 Wave Analyzer, including 15110- PIS Adaptor cable and
ot er accessories
1521 Be(opre-teisantliefmoret5er H
12
151100210r
a
p p
p dhi
cLevelaced
medium-speord er
matar
w i th
1521-P3 80-dB Potentiometer ( 1521-9603)
1521-P1013 Drive Unit ( 1521.9467) ( installed)
1900-P1 Link Unit ( 1900-9601) ( installed)
1900-P3 Link Unit ( 1900-96031
Chart ,,,,,, 10 rolls ( 1521-9464), scale 0-10 kHz
Chart Paper, 10 rolls ( 1521-9465), scale 0-50 kHz
Accessories Available: 1560-P40 and 1560-P42 Preamplifiers; choice
of vibration pickups or microphones.
Mounting: Rack- Bench Cabinets; Includes end frames for bench
use and supports for rack mounting. 1910-A Recording Ware Analyzer,
Dimensions (width x height 4 depth): Bench. 19 x 25 19 4 15V4 in. 60- Hz 115-V Model 1910-9701
(485 x 645 4 390 mm); rack, 19 x 24:9 x 1319 in. ( 485 4 625 x 60- Hz 230-V Model 1910-9711
340 mm). 50- Hz 115-V Model 1910-9493
Weight: Net, 116 lb ( 53 kg); shipping. 227 lb ( 104 kg). 50- Hz 230-V Model 1910.9494

1911-A
RECORDING SOUND AND VIBRATION ANALYZER
This assembly will generate continuous frequency plots
of the 1/3- or 1/10-octave spectrum of sound and vibra-
tion signals over the range of 4.5 Hz to 25 kHz. Thus
1/3-octave measurements can be made in accordance with
several common military and industrial noise-control speci-
fications. While the third-octave bandwidth is convenient
for testing compliance to a specification for maximum al-
lowable noise or vibration level, the 1/10-octave bandwidth
permits identification of individual frequency components,
leading to their reduction or elimination. The analyzer
will accept signals from a sound- level meter, vibration
meter, or bther stable amplifier, or directly from a micro-
phone or vibration pickup. It includes a storage drawer
and oower system control.
The 11111.9 consists of the fellewIng:
1584-A Sound and Vibration Analyzer, rack model
152141 ( or - B(21 for 50-114 supply) Graphic Level Recorder with
40-dB Potentiometer ( 1521-9602) and medium- speed motor
1521-P109 Drina Unit ( 1521-9467)
1521-P15 Link Unit ( 1521-9615), with 16- tooth sprocket in-
stalled ( standard 24-tooth sprocket al. included)
Chart Paper, 10 rolls ( 1521-9469). calibrated 2.5-25 normalized.
logarithmic
Adaptor Cable, double plug to offset phone plug.
Accessories Available: 1560.P40 and 1560-P42 Preamplifiers; 80- da
potentiometer; choice of vibration pickups.
1911-A Recording Sound and Vibration Analyzer,
Mounting: Completely assembled with end frames for bench use. 1911-9701
Hardware for rack mounting is supplied. 60- Hz 115-V Model
60- Hz 230-V Model 1911-9711
Dimensions ( width o height o depth): 1944 n 3144 0 1544 in. ( 500 50- Hz 115-V Model 1911-9493
o 8004 400 mm). 50-Hz 230-V Model 1911-9494
Weight: Net, 101 lb ( 47 kg); shipping, 158 lb ( 72 kg).

299
LIVE ROOM *
a room that is characterized by an unusually small amount of sound
absorption.
LOUDNESS *
the intensive attribute of an auditory sensation, in terms of which sounds
may be ordered on a scale extending from soft to loud. Note: Loudness
depends primarily upon the sound pressure of the stimulus, but it also
depends upon the frequency and wave form of the stimulus.
LOUDNESS CONTOUR*
acurve that shows the related values of sound pressure levels and frequen-
cy required to produce agiven loudness sensation for the typical listener.
LOUDNESS LEVEL*
of a sound, in phons, is numerically equal to the median sound pressure
level, in decibels, relative to 0.0002 microbar, of afree progressive wave of
frequency 1000 Hz presented to listeners facing the source, which in a
number of trials is judged by the listeners to be equally loud. Note: The
manner of listening to the unknown sound, which must be stated, may be
considered one of the characteristics of that sound.
LOUDSPEAKER ( SPEAKER) *
an electroacoustic transducer intended to radiate acoustic power into the
air, the acoustic waveform being essentially equivalent to that of the electrical
input.
MASK I NG *
1. the process by which the threshold of audibility for one sound is raised
by the presence of another (masking) sound.
2. the amount by which the threshold of audibility of asound is raised by
the presence of another (masking) sound. The unit customarily used is
the decibel.
MECHANICAL IMPEDANCE *
the impedance obtained from the ratio of force to velocity during simple
harmonic motion.
MECHANICAL SHOCK *
occurs when the position of asystem is significantly changed in arelatively
short time in a nonperiodic manner. It is characterized by suddenness and
large displacement, and develops significant inertial forces in the system.
ME L*
a unit of pitch. By definition, a simple tone of frequency 1000 Hz, 40
decibels above alistener's threshold, produces apitch of 1000 mels. The pitch
of any sound that is judged by the listener to be ntimes that of a1-mel tone
is nmels.
MICROBAR, DYNE PER SQUARE CENTIMETER *
aunit of pressure commonly used in acoustics. One microbar is equal to 1
dyne per square centimeter. Note: The term "bar" properly denotes a
pressure of 106 dynes per square centimeter. Unfortunately, the bar was once
used to mean 1dyne per square centimeter, but this is no longer correct.
MICROPHONE *
an electroacoustic transducer that responds to sound waves and delivers
essentially equivalent electric waves.

261
NN I
the noise and number index based on perceived noise level. It is used for
rating airplane flyby noise.
NOISE *
1. any undesired sound. By extension, noise is any unwanted disturbance
within a useful frequency band, such as undesired electric waves in a
transmission channel or device.
2. an erratic, intermittent, or statistically random oscillation.
Note 1: If ambiguity exists as to the nature of the noise, aphrase such as
"acoustic noise" or "electric noise" should be used. Note 2: Since the above
definitions are not mutually exclusive, it is usually necessary to depend upon
context for the distinction.
NOISE LEVEL*
1. the level of noise, the type of which must be indicated by further
modifier or context.
Note: The physical quantity measured (e.g. voltage), the reference quantity,
the instrument used, and the bandwidth or other weighting characteristic
must be indicated.
2. For airborne sound unless specified to the contrary, noise level is the
weighted sound pressure level called sound level; the weighting must be
indicated.
NORMAL DISTRIBUTION
See GAUSSIAN DISTRIBUTION.
NOYS
aunit used in the calculation of perceived noise level.
NYOUIST INTERVAL
period equal to the reciprocal of twice the frequency of that component
of the signal having the highest frequency. It is the maximum sampling-time
interval that permits reconstruction of aband-limited signal.
OCTAVE *
1. the interval between two sounds having abasic frequency ratio of two.
2. the pitch interval between two tones such that one tone may be
regarded as duplicating the basic musical import of the other tone at
the nearest possible higher pitch.

Note 1: The interval, in octaves, between any two frequencies is the loga-
rithm to the base 2 (or 3.322 times the logarithm to the base 10) of the
frequency ratio. Note 2: The frequency ratio corresponding to an octave
pitch interval is approximately, but not always exactly, 2:1.
OSCILLATION *
the variation, usually with time, of the magnitude of a quantity with
respect to aspecified reference when the magnitude is alternately greater and
smaller than the reference.
PEAK-TO-PEAK VALUE*
of an oscillating quantity is the algebraic difference between the extremes
of the quantity.
PERCEIVED NOISE LEVEL
the level in dB assigned to anoise by means of acalculation procedure that
is based on an approximation to subjective evaluations of "noisiness."

262
PERIODIC QUANTITY *
oscillating quantity whose values recur for certain increments of the
independent variable.
PHON *
unit of loudness level. (See LOUDNESS LEVEL.)

PINK NOISE
noise whose noise-power-per-unit-frequency is inversely proportional to
frequency over aspecified range.

PITCH *
that attribute of auditory sensation in terms of which sounds may be
ordered on ascale extending from low to high. Pitch depends primarily upon
the frequency of the sound stimulus, but it also depends upon the sound
pressure and wave form of the stimulus. Note 1: The pitch of asound may be
described by the frequency or frequency level of that simple tone, having a
specified sound pressure level, which is judged by listeners to produce the
same pitch.
POINT SOURCE See " SIMPLE SOUND SOURCE."

POWER LEVEL
in decibels, is 10 times the logarithm to the base 10 of the ratio of agiven
power to a reference power. The reference power must be indicated. [The
reference power is taken as 1.0 x 10 -12 watt in this handbook.]

PRESBYCUSIS
the condition of hearing loss specifically ascribed to aging effects.
PRESSURE SPECTRUM LEVEL
of asound at aparticular frequency is the effective sound-pressure level of
that part of the signal contained within a band 1 cycle per second wide,
centered at the particular frequency. Ordinarily this has significance only for
sound having a continuous distribution of energy within the frequency range
under consideration. The reference pressure should be explicitly stated.

PRIMITIVE PERIOD ( PERIOD) *

of a periodic quantity is the smallest increment of the independent
variable for which the function repeats itself. Note: If no ambiguity is likely,
the primitive period is simply called the period of the function.
PROBABILITY DENSITY DISTRIBUTION
See AMPLITUDE DENSITY DISTRIBUTION.
PSIL (THREE-BAND PREFERRED-OCTAVE SPEECH- INTERFERENCE
LEVEL) ( SPEECH INTERFERENCE LEVEL)
average, in dB, of the sound-pressure levels of a noise in the three octave
bands of center frequency 500, 1000, and 2000 Hz. The speech interference
level, or SIL, without the qualifying " PREFERRED" is usually the arithmetic
average of the sound-pressure levels in the older series of three octave bands:
600 to 1200, 1200 to 2400, and 2400 to 4800 Hz.

PURE TONE See SIMPLE TONE.

QUANTIZATION
conversion of avalue into one of alimited set of values. The limited set is
usually a discrete series of total number equal to two raised to an integer
power, that is, abinary set.

263
RANDOM NOISE*
an oscillation whose instantaneous magnitude is not specified for any given
instant of time. The instantaneous magnitudes of arandom noise are specified
anly by probability distribution functions giving the fraction of the total tithe
that the magnitude, or some sequence of magnitudes, lies within aspecified
range. Note: A random noise whose instantaneous magnitudes occur accord-
ing to Gaussian distribution is called Gaussian random noise.
RATE OF DECAY*
the time rate at which the sound pressure level (or other stated characteris-
tic) decreases at agiven point and at agiven time. A commonly used unit is
the decibel per second.
RESONANCE*
of asystem in forced oscillation exists when any change however small in
the frequency of excitation causes adecrease in the response of the system.
Vote: Velocity resonance, for example, may occur at afrequency different
from that of displacement resonance.
RESONANCE FREQUENCY
(RESONANT FREQUENCY)*
afrequency at which resonance exists. Note: In case of possible confusion
:he type of resonance must be indicated: e.g., velocity resonance frequency.
RESPONSE*
of a device or system is the motion (or other output) resulting from an
excitation (stimulus) under specified conditions. Note 1: Modifying phrases
must be prefixed to the term response to indicate kinds of input and output
rhat are being utilized. Note 2: The response characteristic, often presented
graphically, gives the response as a function of some independent variable
;uch as frequency or direction. For such purposes it is customary to assume
:hat other characteristics of the input (for example, voltage) are held con-
;tant.
REVERBERATION*
1. the persistence of sound in an enclosed space, as aresult of multiple
reflections after the sound source has stopped.
2. the sound that persists in an enclosed space, as a result of repeated
reflection or scattering, after the source of sound has stopped.
Vote: The repeated reflections of residual sound in an enclosure can alterna-
:ively be described in terms of the transient behavior of the modes of
/ibration of the medium bounded by the enclosure.
REVERBERATION TIME*
of aroom is the time that would be required for the mean squared sound
aressure level therein, originally in asteady state, to decrease 60 dB after the
iource is stopped.
ROOT- MEAN SQUARE ( rms)
square root of the arithmetical mean of the squares of aset of instantane-
mis amplitudes, or of aset of values of afunction of time or other variable.
3AMPLING
transformation of acontinuous function into adiscrete series of values in
ippropriate order.
3IG MA ( a)
See STANDARD DEVIATION.

264
Type 1921 REAL- TIME ANALYZER

• 3.15 Hz to 80 kHz
• 30 to 45 1/10-, 1/3-,
or 1/1-octave bands
• 70-dB dynamic range,
60 dB displayed
• 100-mV sensitivity
• calibrated attenuators
• digital detection provides
true rms answers
• 9known integration periods
• corrected spectrum displayed
directly

HmIIIIII

1111111111111111111111119/1/1e

Long aneed, now areality Since the inception of the first control and interpretation of the measurements with a
sound and vibration instruments, the need for rapid, on- minimum of operator attention. Even less work and faster
line measurements was apparent. But it took some thirty results are possible when auxiliary instrumentation is
years to fulfill the need, adequately and economically, and wedded to the 1921, instrumentation such as computer
the fulfillment came with the introduction of real-time control, scanner input, dc recorders, or data printers — all
analysis embodied in the 1921. of which interface with ease.
The 1921 is at least 30 times faster than its contemporary A unique solution The input signal is applied to a set of
serial analyzers, which means it can analyze more data analog filters (from 30 to 45 depending on your require-
faster and can measure non-stationary signals without tape ments) that cover a frequency range from 3.15 Hz to 80
loops. Its 70-dB dynamic range readily accepts the ran- kHz. These filters include, as an option, individual attenu-
dom or totally unspecified signals routinely encountered ators to permit pre-whitening or other signal conditioning;
in acoustic work. The answers are not only fast, they are they are housed in aunit that can be purchased separately
repeatable and statistically reliable because of the ana- if desired.
lyzer's -±0.5-dB accuracy and known integration times. The outputs of the filters are processed in another unit,
Optional attenuators provide ±- 0.25-d8 accuracy for dy- the rms detector. This detector is unique in that it proc-
namic-range extension, transducer flattening, subjective esses the signals from the filters digitally; i.e., the outputs
correlation, etc., and the filters incorporate A-, B-, or are sampled, the sampled data are converted to digital
C-weighting networks. The entire capability has been spe- binary form, and the binary numbers are fed to a digital
cifically designed to permit simple, rapid, and precise processor that computes root-mean-square levels.

303
 of this sound to the reference pressure. The reference pressure shall be
explicitly stated. Note 1: The following reference pressures are in common
use;
(a) 2x 10 -4 microbar [ 20 pN/m 2
(b) 1microbar

Reference pressure (a) is in general use for measurements concerned with

hearing and with sound in air and liquids, while (b) has gained widespread
acceptance for calibrations of transducers and various kinds of sound meas-
urements in liquids.

[The reference pressure used in this handbook is 20 micronewtons per square

meter (20 pN/m2).J Note 2: Unless otherwise explicitly stated, it is to be
understood that the sound pressure is the effective (rrns) sound pressure.
Note 3: It is to be noted that in many sound fields the sound-pressure ratios
are not the square roots of the corresponding power ratios.
SPECTRUM*

I. of afunction of time is adescription of its resolution into components,

each of different frequency and (usually) different amplitude and
phase.

2. also used to signify a continuous range of components, usually wide in

extent, within which waves have some specified common characteristic;
e.g., "audio-frequency spectrum."

Note 1. The term is also applied to functions of variables other than time,
such as distance.

SPECTRUM LEVEL ( ACOUSTICS)

ten times the common logarithm of the ratio of the squared sound-pres-
sure-per-unit-bandwidth to the corresponding reference quantity. The unit
bandwidth
imi rn 2 ) 2 ¡Hz.is the hertz and the corresponding reference quantity is (20

SPEECH INTERFERENCE LEVEL

See PSIL.

STANDARD DEVIATION ( SIGMA, a)

linear measure of variability equal to the square root of the variance.
STANDING WAVE*

periodic wave having a fixed distribution in space which is the result of

interference of progressive waves of the same frequency and kind ;character-
ized
in by the existence of nodes or partial nodes and anti-nodes that are fixed
space.

STATIONARY ( STATISTICAL)

notterm that
change describes
with time. a noise whose spectrum and amplitude distribution do

THRESHOLD OF AUDIBILITY
(THRESHOLD OF DETECTABILITY)*
for a specified signal is the minimum effective -sound-pressure level of the
signal that is capable of evoking an auditory sensation in aspecified fraction
of the trials. The characteristics of the signal, the manner in which it is
presented to the listener, and the point at which the sound pressure level is
measured must be specified. Note 1: Unless otherwise indicated, the ambient
noise reaching the ears is assumed to be negligible. Note 2: The threshold is
usually given as a sound-pressur e level in decibels, relative to 0.0002 micro-

266
bar. Note 3: Instead of the method of constant stimuli, which is implied by
the phrase " aspecified fraction of the trials," another psychophysical method
(which should be specified) may be employed.
THRESHOLD OF FEELING ( OR TICKLE) *
for a specified signal is the minimum sound-pressure level at the entrance
to the external auditory canal which, in aspecified fraction of the trials, will
stimulate the ear to a point at which there is asensation of feeling that is
different from the sensation of hearing.
TIF ( TELEPHONE INFLUENCE FACTOR)
an index of the potential interfering effect of aparticular power circuit on
atelephone circuit. ( See AIEE Trans. Vol. 79, Part I, 1960, pp. 659-664.)
TIME SERIES
values ordered in time, asuccession of discrete observations made at points
in time or covering discrete intervals of time. The spacing of observations is
ordinarily uniform on the time scale.
TONE *
(a) asound wave capable of exciting an auditory sensation having pitch.
(b) asound sensation having apitch.
TRANSDUCER *
a device capable of being actuated by waves from one or more transmis-
sion systems or media and of supplying related waves to one or more other
transmission systems or media. Note: The waves in either input or output
may be of the same or different types (e.g., electric, mechanical, or acoustic).
TRANSFER FUNCTION
measure of the relation between the output signal and the input signal of a
system or device, ordiparily the ratio of the output signal to the input signal.
TRANSIENT VIBRATION *
temporarily sustained vibration of a mechanical system. It may consist of
forced or free vibration or both.
ULTRASONICS *
the technology of sound at frequencies above the audio range. Note:
Supersonics is the general subject covering phenomena associated with speed
higher than the speed of sound ( as in the case of aircraft and projectiles
traveling faster than sound). This term was once used in acoustics synono-
mously with " ultrasonics;" such usage is now deprecated.
VARIANCE
quadratic measure of variability, the average of the mean squares of the
deviations from the arithmetic mean of aset of values of avariable.
VELOCITY *
avector that specifies the time-rate-of-change of displacement with respect
to areference frame. Note: If the reference frame is not inertial, the velocity
is often designated relative velocity.
VIBRATION *
an oscillation wherein the quantity is aparameter that defines the motion
of amechanical system.
VIBRATION ISOLATOR *
aresilient support that tends to isolate a system from steady-state excita-
tion.

267
VIBRATION METER ( VIBROMETER)*
an apparatus for the measurement of displacement, velocity, or accelera-
tion of avibrating body.
NAVEFORM
instantaneous amplitude as afunction of time.
NAVEFORM AVERAGING ( SUMMATION ANALYSIS)
summing of corresponding ordinates of selected frames of a wave. The
;ummed values may be divided by the number of frames summed to convert
:o an average.
NE IG HTING**
prescribed frequency response provided in asound-level meter.
NHITE NOISE
power per-unit-frequency is substantially independent of frequency over a
;pecified range. Note: White noise need not be random.

This material is reproduced from the American National Standard Acoustical Terminol-
gy, S1.1-1960, copyrighted by ANSI, copies of which may be purchased from the
,merican National Standards Institute at 1430 Broadway, New York, N. Y. 10018.
*ANSI S1.4-1971.
**ANSI S3.6- I969

268
Appendix VI

Words Commonly Used

to Describe Sounds
The words listed below are commonly used to describe sounds of various
types. Such words are often helpful in conveying information on the general
nature of asound.

BANG CLUCK HUM RING SWOOSH

BARK CLUNK JINGLE RIPPLING TAP
BEEP CRACK JANGLE ROAR TATTOO
BELLOW CRACKLE KACHUNK RUMBLE TEARING

BLARE CRASH KNOCK RUSHING THROB

BLAST CREAK MEW RUSTLE THUD

BLAT DINGDONG MOAN SCREAM THUMP

BLEAT DRIP MOO SCREECH THUNDER

BONG DRUMMING MURMUR SCRUNCH TICK

BOOM FIZZ NEIGH SHRIEK TI CK-TOCK

BRAY GLUG PATTER SIZZLE TINKLE
BUZZ GNASHING PEAL SLAM TOOT

CACKLE GOBBLE PEEP SNAP TRILL

CHEEP GRATING PING SNARL TWANG

CHIME GRINDING POP SNORT TWITTER

CHIRP GROAN POW SPLASH WAIL

CLACK GROWL POUNDING SPUTTER WHEEZE

CLANG GRUMBLE PULSING SQUAWK WHINE

CLANK GRUNT PURR SQUEAK WHIR

CLAP GURGLE PUT-PUT SQUEAL WHISPER
CLATTER HISS RAP SQUISH WHISTLE

CLICK HOOT RAT-A-TAT STAMP YAP

CLINK HOWL RATTLE SWISH YELP
ZAP

269
Type 1921-P2 STORAGE DISPLAY UNIT

lo especially useful with real-time analyzers

• clear, bright spectrum display
• large 8x10cm viewing area
• bench or rack models

Spectrum at aglance As an accessory for the 1921 Real- inch bistab e storage CRT, built-in vertical and horizontal
Time Analyzer, the 1921-P2 displays the results of the deflection amplifiers, and Z-axis modulation capability. It
spectrum analysis— amplitude vs frequency — in asingle, is supplied with inter-connecting cable for the 1921 ana-
easily interpreted format. The display functions of the lyzer and graticules marked with 1 2 -
/ octave standard band
unit are programmable and necessary control and deflec- numbers, center frequencies, and a decibel scale.
tion voltages are provided by the analyzer.
The 1921-P2 is a slightly modified Tektronix* Type 603 •Registered trademark of Tektronix Inc.
Storage Display Unit. It has all solid-state circuits, a 5-

specifications
Frequency: Useful to 100 kHz for waveform display of Y or T plot Power: 90 to 136 end 180 to 272 V, 48-440 Hz, 57 W.
with 1. phase shift between X and Y. Mechanical: Bench or rack models. Dimensions ( w x h x0): Bench,
Display: 8 cm vertical X 10 cm horizontal display on 5- in, flat- 8.5 X 6 X 17.5 in. (216 x 152 x 445 mm); rack, 19 x 5.25 x 17.5
faced bistable storage tube phosphor similar to Pl. Linearity, in. (483 X 133 x 445 mm). Weight: 18 lb ( 8 kg) net, 25 lb ( 12 kg)
2% vertical and 5% horizontal difference between any 2 cm. shipping.
Speed, > 5 cm/ ms. stored line-writing speed. Storage. Z-axis on-
time should be be 9 es to ensure good storage. Viewing time up to Catalog
15 m recommended; erasure becomes more difficult if information Number I Description
is stored longer. Erasure, 200 ms erase time.
1921-P2 Storage Display Unit
Supplied: Five graticule grids graduated vertically every 5 dB from
Bench Model
0 to 60 dB and horizontally in 30 bands, 1 each for bands 5 to 34,
Rack Model
11 to 40, 14 to 43, and 20 to 49, and one with bands unmarked.

Type 1566 MULTICHANNEL AMPLIFIER

• 16 channels
• manual or remote channel selection
• 2- Hz to 100-kHz response
• 55-dB gain, manually or remotely adjusted
• calibration noise source built in

Many inputs — one output Many sound and vibration vides a built-in pink-noise calibration source that speeds
measurements can be simplified by use of a scanner that not only the check out of the scanner but also that of any
connects, in sequence or in any arbitrary order, the outputs analyzer connected to it. The 1566 is particularly useful
from a number of transducers to a single analyzer. A scan- with the 1921 Real-Time Analyzer. This combination can
ner system can be set up to measure signals individually automatically analyze the spectrum from each transducer
scanned or it can measure the space-averaged spectrum
or to average all signals.
using 2, 4, 6, 8, 10, 12, or 16 microphones. This feature
The 1566 scans up to 16 channels ( up to 99 with aspecial makes possible automatic real-time sound-power meas-
additional unit), amplifies each by up to + 55 dB, and pro- urements.

specifications
Channels: 16 plus 1 for calibration, expandable to 99 ( additional is — 3 dB per octave. Noise signal applied to internal calibration
channels housed in a special unit). Control: Active channel is se- channel is adjustable from 30 to 100 mV rms. Rear- panel noise
lected manually or by external 1-2-4-8 BCD signal, or automati- output is fixed at 100 mV rms and can be loaded by 0.05 µF
cally scanned in sequence with range of channels to be scanned without affecting spectrum up to 100 kHz.
selected by thumbwheel switches; dwell time adjustable from Supplied: Power cord, two 24-pin data plugs.
100 ms to 10 s or infinity (channel advance initiated by external
signals); scan set to occur once or repetitively and started, stopped Available: 1560-P40 and - P42 Preamplifiers ( 1566 provides power
on active channel. or reset to lowest channel by pushbuttons or for up to 99 of either). 1566-9500 Cable Set for connection to 1921
external closures to ground. Display: Two, high- intensity neon Real-Time Analyzer, microphones, vibration pickups.
readout tubes display active channel. Power: 100 to 125 or 200 to 250 V, 50 to 60 Hz, 30 W.
Frequency: 2 Hz to 100 kHz, flat within i-0.5 dB.
Mechanical: Bench or rack mount. Dimensions ( w x h x d): Bench.
Sensitivity: 1.8 mV to 1.6 V for 1-V output; gain set in 1-dB in- 19 1
/ x 5 x 20 in. ( 495 x 130 x 510 mm); rack, 19 x 31
2 /
2 x 18 1
/
2 in.
crements by panel control or 1.2-4-8 BCD signal at standard OIL (485 x 89 x 470 mm). Weight: Bench. 32 lb ( 15 kg) net, 47 lb
levels ( logic 0 c ground, logic 1 • 3 V). Rear-panel adjust-
(213 kg/ shipping; rack 26 lb ( 12 kg) net, 41 lb ( 19 kg) shipping.
ment provides 10-de continuous control of gain for all channels
for calibration. Each channel includes a 6.dB gain adjustment Ca talog
for transducer sensitivity equalization. Description
Impedance: Input, 100 ka. Output, 600 0.
Noise: < 10 µV equivalent input noise in each channel when gain 1588 Multi:funnel Amplifier
is maximum and source impedance is ,`" 100 D. 1566-9700 Bench Model
Calibration: Built-in pink- noise (± 0.5 dB) source with symmetrical 15669701 Rack Model
Gaussian distribution from 2 Hz to 100 kHz. Spectrum- level slope 1566-9500 Coble Set

307
 The following are recommendations of the International Organization for
Standardization: ( available from ANSI)

ISO/R16-1965 Standard Tuning Frequency

ISO/R31/Part VII-
1965 Quantities and Units of Acoustics
ISO/R131-1959 Expression of the Physical and Subjective Magnitudes of Sound
or Noise
ISO/R140-1960 Field and Laboratory Measurements of Airborne and Impact
Sound Transmission
ISO/R226-1961 Normal Equal- Loudness Contours for Pure Tones
ISO/R266-1962 Preferred Frequencies for Acoustical Measurements
1SO/R354-1963 Measurement of Absorption Coefficients in a Reverberation
Room
ISO/R357-1963 Power and Intensity Levels of Sound or Noise
ISO/R 362-1964 Measurement of Noise Emitted by Vehicles
ISO/R389-1964 Reference Zero for Pure-Tone Audiometers
ISO/R454-1965 Relation Between Sound Pressure Levels of Narrow Bands of
Noise in aDiffuse Field and in aFrontally-Incident Free
Field for Equal Loudness
ISO/R495-1966 Preparation of Test Codes for Measuring the Noise Emitted by
Machines
ISO/R507-1970 Describing Aircraft Noise Around an Airport
ISO/R532-1966 Calculating Loudness Level
ISO/R717-1968 Rating of Sound Insulation for Dwellings
ISO/R1761-1970 Monitoring Aircraft Noise Around an Airport

The following are st andards of the International Electrotechnical Commission:

(available from ANSI)

IEC/50-08(1960) International Electrotechnical Vocabulary Group 08:Electro-

Acoustics
IEC/118(1959) Measurements of the Electro -
Acoustical Characteristics of
Hearing Aids
IEC/123(1961) Sound Level Meters
IEC/124(1960) Rated Impedances and Dimensions of Loudspeakers
IEC/126(1961) IEC Reference Coupler for the Measurement of Hearing Aids
Using Earphones Coupled to the Ear by Means of Ear
Inserts
IEC/177(1965) Pure Tone Audiometers for General Diagnostic Purposes
IEC/178(1965) Pure Tone Screening Audiometers
IEC/179(1965) Precision Sound Level Meters
IEC/184(1965) Specifying the Characteristics of Electromechanical Transducers
for Shock and Vibration Measurements
IEC/200(1966) Measurement of Loudspeakers
IEC/222(1966) Specifying the Characteristics of Auxiliary Equipment for
Shock and Vibration Measurement
IEC/225(1966) Octave, Half-Octave and Third-Octave Band Filters Intended for
the Analysis of Sounds and Vibrations
IEC/263(1968) Scales and Sizes for Plotting Frequency Characteristics
IEC/303(1970) IEC Provisional Reference Coupler for the Calibration of Ear-
phones Used in Audiometry
IEC/327(1971) Precision Method for Pressure Calibration of One-inch Standard
Condenser Microphones by the Reciprocity Technique

American Society for Testing and Materials, 1916 Race St., Philadelphia, PA 19103

Methods of Test for.

C367-57 Strength Properties of Prefabricated Architectural Acoustical

Materials
C384-58 Impedance and Absorption of Acoustical Materials by the Tube
Method

271
 C423-66 Sound Absorption of Acoustical Materials in Reverberation
Rooms (S1.7-1970)
C522-69 Airflow Resistance of Acoustical Materials
C643-69 Painting Ceiling Materials for Acoustical Absorption Tests
Recommended Practices for:

E90-70 Laboratory Measurement of Airborne Sound Transmission Loss

of Building Partitions
E336-67T Measurement of Airborne Sound Insulation in Buildings
C636-69 Installation of Metal Ceiling Suspension Systems for Acoustical
Tile and Lay-In Panels
Definition of Terms Relating to:

C634-69 Acoustical Tests of Building Constructions and Materials

Standards Prepared by Professional Societies

American Society of Heating, Refrigerating and Air-Conditioning Engineers

(ASHRAE), 345 East 47th Street, New York, N.Y. 10017
36-62
Measurement of Sound Power Radiated from Heating, Refrig-
erating and Air Conditioning Equipment
36A-63 Method of Determining Sound Power Levels of Room Air-
Conditioners and Other Ductless, Through-the-Wall Equip-
ment
36B-63 Method of Testing for Rating the Acoustic Performance of Air
Control and Terminal Devices and Similar Equipment
36R-Feb 1970
(Draft) Method of Testing for Sound Rating Heating,
Refrigerating and Air Conditioning Equipment
Instrument Society of America, 400 Stanwix St., Pittsburgh, PA 15222.

RP37.2 Guide for Specifications and Tests for Piezoelectric Acceleration

S37.10 Transducers for Aerospace Testing ( 1964).
Specification and Tests for Piezoelectric Pressure and Sound.
Pressure Transducers ( 1969)

Institute
10017 of Electrical and Electronic Engineers, 345 East 47th Street, New York, N.Y.

IEEE 85 Airborne Noise Measurements on Rotating Electric Machinery

(1965)
IEEE 219 Loudspeaker Measurements (ANSI S1.5-1963)
IEEE 258 Methods of Measurement for Close-Talking Pressure Type
Microphones (January 1965)
IEEE 269 Method for Measuring Transmission Performance of Telephone
Sets (May 1966)

Society of Automotive Engineers, 345 East 47th Street, New York, N.Y. 10017
SAE Committee A-21, Aircraft Noise Measurement

ARP 796 Measurements of Aircraft Exterior Noise in the Field

AIR 817 A Technique for Narrow Band Analysis of aTransient
AIR 852 Methods of Comparing Aircraft Takeoff and Approach Noises
ARP 865a Definitions and Procedures for Computing the Perceived Noise
Level of Aircraft Noise
ARP 866 Standard Values of Atmospheric Absorption as aFunction of
Temperature and Humidity for Use in Evaluating Aircraft
ARP 876 Jet Noise
Flyover
Prediction
Noise

AIR 902 Determination of Minimum Distance from Ground Observer

to Aircraft for Acoustic Tests
AIR 923 Method for Calculating the Attenuation of Aircraft Ground to
Ground Noise Propagation During Takeoff and Landing
ARP 1080 Frequency Weighting Network for Approximation of Perceived
Noise Level for Aircraft Noise
AIR 1115 Evaluation of Headphones for Demonstration of Aircraft Noise

272
 SAE Sound Level Committee
J6a Ride and Vibration Data Manua
J336 SAE Recommended Practice, Sound Level for Truck Cab
Interior
J366 SAE Recommended Practice, Exterior Sound Level for Heavy
Trucks and Buses
J377 SAE Standard, Performance of Vehicle Traffic Horns
1672a SAE Standard, Exterior Loudness Evaluation of Heavy Trucks
and Buses
J919 SAE Recommended Practice, Measurement of Sound Level at
Operator Station
J952a SAE Standard, Sound Levels for Engine Powered Equipment
J986a SAE Standard, Sound Level for Passenger Cars and Light Trucks
J994 SAE Recommended Practice, Criteria for Backup Alarm Devices

Industry Groups
Air Conditioning and Refrigeration Institute (
ARI), 1815 North Fort Meyer Drive,
Arlington, Virginia 22209
ARI 443-66 Rooms Fan-Cool Air Conditioner
ARI 270-67 Sound Rating of Outdoor unitary Equipment

Air Diffusion Council (

ADC), 435 North Michigan Ave., Chicago, III 60611
AD-63 Measurement of Room-to- Room Sound Transmissions Through
Plenum Air Systems
1062-R2 • Equipment Test Code
Air Moving and Conditioning Association (
AMCA), 205 West Touhy Avenue, Park
Ridge, Ill 60068
Bulletin 300-67 Test Code for Sound Rating
Bulletin 301-65 Method of Publishing Sound Ratings for Air Moving Devices
Bulletin 302-65 Application of Sone Loudness Ratings for Non- Ducted Air
Moving Devices
Bulletin 303-65 Application of Sound Power Level Ratings for Ducted Air
Moving Devices
Publication 311-67 AMCA Certified Sound Ratings Program for Air Moving Devices

American Gear Manufacturers Association (

AGMA), One Thomas Circle, Washington,
D.C. 20005
295.02 AGMA Standard Specification for Measurement of Sound in
High Speed Helical and Herringbone Gear Units, November
1965.
The Anti-Friction Bearing Manufacturers Association, Inc., 60 East 42nd Street, New
York, N.Y. 10017
AFBMA Standard, Section 13, Roller Bearing Vibration and
Noise
Association of Home Appliance Manufacturers, 20 North Wacker Drive, Chicago, Ill
60606
Standard RAC- 2-SR Room Air Conditioner Sound Rating, January 1971

Compressed Air and Gas Institute (

CAGI), 122 East 42nd Street, New York, N.Y. 10017

(see ANSI S5.1-1971)

National Electrical Manufacturers Association (
NEMA), 155 East 44th St., New York,
N.Y. 10017
SM33-1964 Standards Publication, Gas Turbine Sound and Its Reduction
MG1 Motors and Generators
Section 4.3.2 Method of Measuring Machine Noise
TR1-1963 Transformers, Regulators and Reactors ( Sections 9-04 and
9-05)

1
273
National Fluid Power Association, Thiensville, Wisconsin 53092

NFPA T3.9-70.12 Method of Measuring Sound Generated by Hydraulic Fluid

Power Pumps

National Machine Tool Builders Association (

NMTBA), 2139 Wisconsin Ave., Washington,
D.C. 20007

Noise Measurement Techniques, June 1970

Journals in the Field of Sound and Vibration.

Acustica, S. Hirzel, Stuttgart 1, Birkenwaldstr. 44, Postfach 347, Germany. ( An inter-
national journal on acoustics.)
Akusticheskü Zhurnal, ( in Russian), Academy of Sciences of USSR, Moscow.
Applied Acoustics, Elsevier Publishing Co. Ltd., Ripple Road, Barking, Essex, England.
(An international journal.)
IEEE Transactions on Audio and Electroacoustics, The Institute of Electrical and
Electronics Engineers, Inc., 345 East 47 Street, New York, N.Y. 10017.
The Journal of the Acoustical Society of America, Acoustical Society of America, 335
East 45 Street, New York, N.Y. 10017. ( The most comprehensive scientific journal in
acoustics with occasional papers on vibration. The Society also publishes cumulative
indexes to the Journal and to other acoustical literature.)
The Journal of the Acoustical Society of Japan, ( in Japanese), Acoustical Society of
Japan, Ikeda Building/2-7-7, Yoyogi, Shibuya-Ku, Tokyo, Japan.
Journal of the Audio Engineering Society, Audio Engineering Society, Room 929,
Lincoln Building, 60 East 42nd Street, New York, N.Y. 10017.
Journal of Auditory Research, The C. W. SkiDing Auditory Research Center, Inc., Box
N, Groton, Connecticut 06340.
Journal of Sound and Vibration, Academic Press, Inc., Limited, Berkeley Square House,
Berkeley Square, London W1X 6BA and Academic Press Inc., 111 Fifth Avenue, New
York, N.Y. 10003. (An official medium of publication for the British Acoustical
Society.)
Kampf Dem firm, ( in German), J. F. Lehmanns Verlag, Agnes-Bemauerplatz 8, 8000
Miinchen 21, Germany.
Liirmbekiimpfung, ( in German), Verlag Fiir Angewandte Wissenschaften GmbH, Hard-
strasse 1, 757 Baden-Baden, Germany.
NOISE/NEWS, Institute of Noise Control Engineering, P. 0. Box 1758, Poughkeepsie,
NY 12601. ( Published in cooperation with the Acoustical Society of America).
Revue D' Acoustique ( in French), G.A.L.F. ( Groupement des Acousticiens de Langue
Francaise), Secrétariat, Dipartment Acoustique du CNET, 22-Lannon, France.
Sound and Vibration, Acoustical Publications, Inc., 27101 E. Oviatt Road, Bay Village,
Ohio 44140.

Numerous other societies in the United States are interested in sound and vibration
measurements, for example:

American Hearing Society

American Industrial Hygiene Association
American Medical Association
American Psychological Association
American Society of Heating, Refrigerating and Air-Conditioning Engineers
American Society for Testing and Materials
American Speech and Hearing Association
American Society of Mechanical Engineers
Institute of Electrical and Electronic Engineers
Institute of Environmental Sciences
Instrument Society of America
Society of Automotive Engineers
Society of Experimental Psychologists
Society of Experimental Stress Analysis

Many others in other countries, and many trade journals, publish occasional papers
on acoustics, noise control and vibration.

274
 Type 1569 AUTOMATIC LEVEL REGULATOR

• 2 Hz to 100 kHz
• 50-dB control range
• acoustic-system accessory

VOLTAGE
The 1569 Automatic Level Regulator is intended as an '"V°L. CONTROLLED •IIPLIPIER
INPUT ATTENUATOR
accessory for an oscillator or for a source of narrow-
band noise. Its primary functirn is to control the signal
level in swept-frequency sound and vibration tests.
The regulator senses a control voltage from a micro- LP PA-TER WIT., CONTROL
AVERAGE
phone, accelerometer, or other transducer monitoring the ADJUSTAGLE
CUT- 0, DETECTOR
SIGNAL
•MPLIPIE
5
,41
On
sound or vibration to be controlled and adjusts its output ARTOUENCT

to maintain constant level (see diagram). Output level is 018V3111 of 1569 Automatic Level Regulator
indicated by a panel meter with a linear-dB scale, show-
operating frequency and magnitude-phase conditions in
ing the operator where in its 50-dB control range the
regulator is operating. Regulation is such that a level the control loop.
The 1569 can also be used to regulate voltage from
variation (without the regulator) of 25 dB, for instance, an oscillator or other signal source. In this mode, the
is compressed to a variation of 1 dB. The control rate
is adjustable by means of a panel control to suit the control range is limited to 15 dB.

— See GR Experimenter for April 1968.

specifications
Noise: Typically better than 65 dB below 3 V in 100-kflz band.
Frequency Range: 2 Hz to 100 kHz. Harmonic Distortion: < 1% total for < 1-V output level.
Control Range: 50 dB. Automatic " Shot- Damn": A loss of drive ( input) voltage from sig-
Compression Ratio: 25 ( 0.04 dB per dB). nal source causes the output voltage to drop to zero to protect
DRIVE ( INPUT) equipment connected to output.
Voltage Required ( for normal operation): 1 V. CONTROL-SIGNAL INPUT
Impedance: 100 Id/. Voltage: 5 rnV to 4 V required.
OUTPUT Impedance: 25 MU.
Voltage: 3 V max to 10 rnV min. Control Rates and Corresponding Min Operating Frequencies:
Impedance: 600 0 Any load impedance can be connected with- 1000 dB/s 1200 dB/s 1100 dB/5 130 dEl/s 1 10 dB/s 1 3 dB/s
out affecting linear operation of output circuit.
600 Hz F20060 1 60 Hz 1 20 Hz 1 6 Hz 1 2 Hz

Power Required: 100 to 125 or 200 to 250 V ( switch selected),

50 to 60 Hz, 4 W.
Accessor ies Supplied: Power cord, mounting hardware with rack
or bench models.
Accessories Available: GR 1560-P40 Preamplifier ( power for pre-
amplifier available at rear- panel input connector): 1304 Beat-
Frequency Audio Generator, 1521 Graphic Level Recorder; micro-
phones and vibration pickups.
Mounting: Rack- Bench Cabinet.
Dimensions ( width e height x depth): Bench model, 19 x 3,/a x
13 in. ( 485 x 99 x 330 mm); rack model, 19 x 31 / x 10 1
2 / in. ( 485
2
x89 x275 mm).
Weight: Net, 13 lb ( 6 kg): shipping, 30 lb ( 14 kg).

Catalog 1
Number I Description

l 1569 Automatic Lem Regulator

1569-9700 Bench Model
1569-9701 Rack Model
Typical Measurement System Using 1569

311
 0101 1
4 01- 116 0444*04 991$0M4 $ 00110900 00 940 9 4
1 101 1
0 01 ....e-wre 1904 / 490449). ap

.11 . . uà
• • • • us 11 133111001.11111001113
•

111 0 14*00a 40004 11 1910 0 006 1 mu 0 ... coi 0800001 sail 0 0311 mend
_
a .4
0 11 096 191 11 0490 • 0 111
MO t1 541110411
pi 1« 10 005911
saà 84 10112 S .0 1 Ott ll 1 SS i
t,
111 111 10011 9 0
MI MI 011 0/ mg i1 NI 1941 911 IN 9f
006 uà
111 0 isomm MI 0 1/ 111 060 / I 190/1 * I 0900M 09 11
1 11 900611 / SIN

111 11 11 411 MI 111111 008 LI

0 001 /It 009/ 1 1101,1116108

.81 8.8 148 9a• 11 0M 0 001 90 001( 0 1901 001 104 OS 0 0 0111111

0 su 006 0
06 0 711 94 511 ( IS 211 E1 0.6 511 1140 a loI 111 0801019

0..4 16 ou * *I .1
,00 09049 01/ 0 141 WI saà .4 min
89 • E
CI 1.4 100 004 11111 4111 111 OSE / 0 spaN ou m 0 171131
_

1 1 sei
s 009 0 m m 096 not 1 US t1 0904M 001 00 006 11 0040400
-
us tat
01 . 11 111 - M I0 0900.9001 8000M OS 1 014
01 1t 910110 1111103

111 0 00 4 .14
1 1
1 IN 0 tOS I0 0* ( 0 00.9 60 01•10 00 1 114 09019009 00001

ow tat IN pi 0 4 1 *V0 1
1110, 1411 1 0311 0 091 t 10* 001 11.0 11 a 01 0
o.)....fi .4 owns. mom

-000 aN pi 0 11 090 001010 0900M 01 saà 0 111011111111

0 000
41 0 1110 0 0111 1 01010 01 tt 03 0 0
1 • /1 NOD

11 9.4 us 14 096 al «I CPI 0 KW 9 $ 110 00/ 090011 WI 11 .11 mown»

01 01 NI 0W 4110 IN 111 »I 0 006 10 090000111 114

00 501 01 , MOO

.II Um NI 1 0110
1 11 WM 01 00/1 t 15 I0 090011 04 10111 OS .
11 11 00004011M

111 006 011 04 ei mr »or 40 8

919 1 001 1 00S/ 1 OKI 1 0000M OSI O900MOO 11 11 1101000 ¡MI
1111 fit 4
096 su up 104 0,11 NW "" ' Milt 00r0 0 11.1 11 01 MO
0 à
1•11 MM

- - - - - IN - 0/ 11 1901 11
Ott I1 11161 09 11 0 0)11111 ION

090 sey, up soi, 1 XI 00 0001 1 001/ 1 09001104 140 09 «I .8 /*if 401

0 0 - - IWO 0900M 110 0110000 1

41 111 611 11/ 1 .
0 4 .1 310204 41

NI so8 ow iri ow 9011 WI 001110 (MS 0 0400090 Of 11» 0/ .0 us 0011N

04 0 uà 108 090
MI 0110 011S 0 OW 00 1 011 49000000 01 000 1911111111

• • • • • . • 01111101

- 009 004 uà 09 10/0 41

MI. 11
190 111105111
"9. ''')/ :1
' 000

006 001 - - 11 IN No 009 0 OM I1 1101 011 411 OP INI 0 40111111

0 saà 84009 14 0 4 000 M 0 mein SU 91 1190 WI 11141 St 11 110 010534111

- 0 - 04 11 /11 - • • • 11 .0 1013111

011 su 8411004 001 II 02.0 01

MOW 04 010 011 0» 1191à 11 111 .0 0111110111111

à - 41041
1 - 1111 - 01(6 S 11 It 0900M ( Ft 1101 1 11 0 01198/111

99/ 01 sm 0119 1100 9 Wit 119( 1 +NM. (101 100 OS 000

1001
fl/ 1 11 saà IN»

- - - 111 0 - - - - - 90 111101101

- 840144 04 0 1
1 IN 0 0* i0 WIE 0 01»ii (
61 19016 10 04 11 MOM»

_ _ _ - 94 *1St gal s aao 011 1941 0( 1 100 MIMI

I .088 - - tià III 008 11 6I Ott 11 ,080 00 401 000000 00 01 006 Vie

NI 0 «11 in wi Dent Oa t0 1011 01 1 0 09

1 111111.4

_ au 840004 04 m 111 000 ON 11 1 031 t0 19.11 S01 1901 (It 0.4 uà Ill

11 11 8 41 0 41 .411 10110.1
. NI /1/ X/ 91 0% I1 1111 501 101 SI 01 oe, 11901

94 0 004 ut 11 110 000 01 In MIS 1 001 04008005 .8

1 uà 1190901

_ 0 010 006 440 m 1 05S t0 000 11 00006 05i 110 09 00 uà

1 1151/019

- pi 8490104 11 0 111 111 »I 0 01 ls .... 051 1.11 Oe .1% 0 011011

11
1 ta4 * 410 410 IOW 0 00011 099 t0 1101 00/ 111 15 ON 148 11111103 10 00

11 ul 009 - - 111 04 501(10 19 5 0 191 1/1 S/1 110.1 S0 01/ 0011 3110111

NI 000 pi 000 la 000 101 US IX 60 141 11 1061 1 ON 11 1080310403

_ 0 _ - - 10 0 0 210 00 - - 0 pi 0010010
à 0
104 0 41004 us 401011.1 009 21S 11 t0 101 01/ 1 09 11 0
00X 0 11 101 004 111/01110

sat 0 so. 41 1 41 161 0111 0 OS( 10 096 11 001 01 0909010 111 0 51510111

999900504 0 40004 41110004 *my rwo 90 morn 04 110 0400011 09 11041 IX 111 001*. 0001119

911 004 1406 449004 0 00 1111 oei moon omit 09

00M W 090000 01 .4 1113119
.49001
. . • 111 08000091 0000605
g

OSI 1 0 059 00 111 al 0111110

14 101
_ -

r
r '''' _.

--ef .s..:, foe

-'
IS Cadre Gnen le

_Jr
...ewer. ell Neanag

e
If wr
ts f

-
f!
Y8

e 1
' a
.....,_ it ,t,i 11 fie:! i
f
« _ 1
_ î It
rf 11
.....
 Jrisdiction Comments

ALABAMA • Workm•n's Compensation laws administered by courts. May use any Criteria appropriate to determine issu•.

' ALASKA Awards for tinnitus and non- organic loss if it effects hearing and is due to work. ( Plan to adopt OSHA noise guidelines./

I ARIZONA Noise-induced hearing loss recognized under gradual theory — series of traumatic incidents. Compensation for tinnitus; non.
organic lost based on medical evidence of impairment of function and inability to work.

ARKANSAS Might provide hearing aid and credit for improvement as done with g

5 CALIFORNIA
'nte_ne,:elostrses noonn:eghe.du:d o.c,:aLtond:r,fe:dast,. ::re'r'inennix1.'ndards f
or time" and

I COLORADO At present time only traumatic hearing loss is compensable. Consideration being given to include noise- induced hearing loss.
Have adopted Walsh Healey regulations.

7 CONNECTICUT

8 DELAWARE All Workmen's Compensation acts are being reviewed by Itse assembly.

9 DIST. OF COLUMBIA Award for tinnitus and non-organic hearing loss if associated with hearing loss.

0 FLORIDA Compensation for non-organic hearing loss possible on medical evidence if related to employment.

1 GEORGIA Comp aaaaa ion for noise- induced hearing loss decided by court decision. Treated as accidental ( traumatic) loss. Provision for
temporary as well as scheduled loss depends upon medical opinion for various criteria used.

12 HAWAII Have regulation on noise exposure above 90 dlIA. Compensation for tinnitus and non-organic loss on medical, psychiatric basis.

13 IDAHO Consideration of presbycusis, tinnitus, and non-organic loss based on medical evidence.

14 ILLINOIS Occupational safety bill passed Apr. 29. 1971 ( like Walsh Healey). Hearing loss must be total in one or both ears.

IS INDIANA

16 IOWA Credit for improvement with hearing aid. Mean between improved and unimproved hearing.

17 KANSAS Uses modified weighted AMA formula of 1942.

18 KENTUCKY Provision for hearing aid if medically indicated.

19 LOUISIANA Compensation based upon 65 per cent of wages during disability not beyond 300 weeks if temporary, or 400 weeks if total, and
must leave work to prove disability.

20 MAINE Deduction for presbycusis — l/x dB each year after age 40.

21 MARYLAND

n MASSACHUSETTS • Must be total loss of hearing for all practical purposes to be compensable.

23 MICHIGAN Hearing loss must be total in one or both ears to be compensable. Compensation based on wage loss and inability to work.

24 MINNESOTA Award for tinnitus and non-organic hearing loss on basis of wage loss. Award on one ear up to 85 weeks— 90 dBA used.

25 MISSISSIPPI Award for tinnitus and non- organic hearing loss if lob- related.

26 MISSOURI Deduction for presbycusis — /

2 dB for each year after 40. Separate schedule tor loss resulting from noise and trauma.
1

27 MONTANA Legislature passed law covering occupational hearing loss to become effective Jan. 1, 1972.

28 NEBRASKA Compensate for tinnitus if associated with hearing loss. Non-organic on neurosis basis. Has frequent claims.

/9 NEVADA Has general safety rule applying to noise exposure.

30 NEW HAMPSHIRE * Anticipate adoption of federal noise esposare standards.

31 NEW JERSEY Large number of cases- 607 in 1969, more in 1970.

32 NEW MEXICO District courts administer the comp eeeee ion act.

33 NEW YORK Formulating safety code after Walsh Healey act. Compensation for tinnitus and non-organic loss if causally related.

34 NORTH CAROLINA Legislation enacted to compensate for noise- induced hearing loss. Effective Oct. I. 1970. Deduction for presbycusis after age 38.

35 NORTH DAKOTA Use Walsh Healey guidelines for noise exposure.

36 OHIO Loss must be total in one or both ears to be compensable.

37 OKLAHOMA Considered as series of traumatic incidents. Permit use of live voice tests at measured distances.

38 OREGON Schedule is based on degrees rather than weeks. Each degree is worth SSS

39 PENNSYLVANIA Most have complete loss of hearing in bolt ears. Receives compensation at rate of 66 per cent of weekly wages for 180 weeks
(180 x $60 = $ 10,800). Legislation suggested but has not been • nacted. Passed specific safety rule for hearing Protection in noise.

40 RHODE ISLAND Legislation effective Sept. 1, 1969. Deduction for presbycusis, /

2
1 dB each year after 40 Benefits were reduced.

41 SOUTH CAROLINA

42 SOUTH DAKOTA Occupational noise compensable as of July I. 1971. Credit for improvement with hearing aid.

43 TENNESSEE

44 TEXAS Texas law effective June 9, 1971 provides comp eeeee ion of noise- induced hearing loss. Compensation for noise•induced hearing
loss is based on both ears only.

45 UTAH Impairment computed on binaural hearing only.

46 VERMONT Department of Health uses Walsh Healey criteria. May pass legislation in 1972.

47 VIRGINIA Hearing table slightly different than AMA. Impairment begins at 17 dB and ends at 80 dB.

48 WASHINGTON Noise exposure criteria like Walsh Healey. Award for tinnitus and non- organic hearing loss on medical evidence.

41 WEST VIRGINIA Presbycusis and tinnitus given consideration based on medical evidence.

50 WISCONSIN Award for tinnitus and non-organic hearing loss if it is disabling and related to employment. Has separate schedule for noise
and traumatic hearing lost cases.

51 WYOMING Rated as percentage loss similar to rating for impairment of vision.

52 PUERTO RICO

53 CANADIAN PROVINCES The statutes and provisions in the Canadian Provinces show slight variations- In I they compensate for noise- induced
hearing loss and have specific guidelines ( Walsh Healey) for noise exposure and for the use of personal hearing protection
devices. Com ion is based upon the percentage of the body as a whole, and varies from three to five per cent for total
loss in one ear to thirty per cent for both ears. They use a modification of the AMA formula as follows: No award for average
loss less than 25 dB, ASA scale or 35 dl, ISO scale. Complete loss if average is 70 dB on ASA scale or 80 dB on ISO scale.
The determination of binaural hearing loss varies from one- to-five to one- to- nine ratio. Most provinces require the applicant to
remove himself from noise exposure to collect an award ( Pension). Some Provinces have • presbycusis deduction of VI dl for
each year over 50: some start at 60, and some provinces have no presbycusis deduction. Ther• is provision for furnishing •
hearing aid where medically indicated, but no credit is allowed for improvement by use of • hearing aid.
Appendix IX

General Radio
Product Directory
The following pages include detailed specifications for General
Radio sound- and vibration-measuring instruments and accessories;
specifications given are subject to change without notice.

Community Noise Measurement

GR Community Noise Measurement Sets 279

Industrial Hearing Conservation

Plant Noise Measurement & Reduction
1565-B and 1563 Sound- Level Meters 280
1556-B Impact-Noise Analyzer 281
Noise Monitors/Hearing Testing
1944 Noise Dosimeter 282
1937 Manual Audiometer 283
1703 Recording Audiometer 283
Calibration
1562-Z Audiometer Calibration Set 284
1562 Sound-Level Meter Calibrator 284
1567 Sound- Level Meter Calibrator 284
Product Noise Reduction and Testing
Measurement
1551-C Sound- Level Meter 285
1933 Precision Sound-Level Meter and Analyzer 286/287
1935 Cassette Data Recorder 288
1940 Power Supply and Charger 288
1553 Vibration Meter 289
1557 Vibration Calibrator 290
Analysis
1523 Recorder/Analyzer 291/293
1568 Wave Analyzer 294/295
1900 Wave Analyzer 296
1564 Sound and Vibration Analyzer 297
1521-B Graphic Level Recorder 298
1910/1911/1913 Recording Analyzer 299/300
Real-Time Analysis
1923 Real-Time Time-Series Analyzers 301/302
1921 Real-Time Analyzer 303/304
1925 Multifilter 305
1926 Multichannel RMS Detector 306
192I-P2 Storage Display Unit 307
1566 Multichannel Amplifier 307
1522 Recorder (for RTA) 308
Other Instruments
Noise Generators
1381/1382 Random-Noise Generators 309
1390-B Random-Noise Generator 310
Miscellaneous System Building Blocks
1569 Automatic Level Regulator 311
1952 Universal Filter 312
Transducers and Accessories
Microphones 313/314
Accessories 315/316
Vibration Pickup 317

278
Community Noise Measurement

GR COMMUNITY NOISE MEASUREMENT SETS

SOUND- LEVEL VEHICLE- NOISE

MEASUREMENT SET MEASUREMENT SET

Whether or not your community has This GR set features simplicity of opera- 1567 Sound-Level Calibrator, included
enacted noise ordinances, GR's 1565- tion and calibration so that reliable meas- in the set. The sound-level calibrator and
9905 Sound-Level Measurement Set will urements can be made by law enforce- sound-level meter are protected during
be avaluable asset. The set consists of a ment officers who have had no prior ex- storage and transportation from site to
1565-C Sound-Level Meter for making perience in acoustic measurements. The site by an impact-resistant, padded car-
noise measurements from 30 to 130 dB, set includes aGR 1565-9015 Sound-Level rying case.
a 1567 Sound- Level Calibrator for rou- Meter that measures in the range of 50 to
tine calibration of the meter, awind- 140 dB. Since most vehicle-noise meas-
Set Specifications
screen, and arugged carrying/storage case. urement regulations specify microphone
location during measurement, with the 1565-9906 Vehicle-Noise Measurement
The 1565-C Sound- Level Meter has been observer out of the sound field, the set Set: Consists of 1565-9015 Sound-Level
specifically designed for community noise includes aremote microphone, extension Meter, 1567 Sound-Level Calibrator, mi-
measurements. It meets the American cable, and tripod. crophone, adaptor for microphone, tripod,
extension cable ( 100 ft), windscreen, carry-
National Standards Institute ( ANSI) Type
2Specifications, which detail the required Calibration of the sound-level meter for ing case, batteries, screwdriver for calibration
instrument performance. The 1565-C is optimum measurement accuracy is easily adjust, and minature phone plug that con-
performed in just afew seconds with the nects to sound-level-meter output.
easy to operate and requires no prior
knowledge of acoustics. Its rugged ceram-
ic microphone, high-impact case, and all
solid-state construction make it extremely
durable, even under the severe operating
conditions encountered during field meas-

f
urements.

Set Specifications
1565-9905 Sound- Level Measurement
Set: Consists of 1565-C Sound- Level
Meter, 1567 Sound- Level Calibrator,
windscreen, carrying case, batteries,
screwdriver for calibration adjust, and
miniature phone plug that connects to
sound-level-meter output.

ser

Sound Level Measurement Set Vehicle-Ncuse Measurement Set

279
Industrial Hearing Conservation

GR 1565-B and GR 1563 SOUND- LEVEL METERS

• 40- to 140-dB range

• meet ANSI and IEC standards

• rugged ceramic microphones

• FET and integrated-circuit design

combine performance with reliability

• convenient pocket proportions —

small and light

1565-8 Sound- Level Meter 1563 Sound- Level Meter

Type 2 Type 3

The best of both worlds The 1565.8 is a full-fledged stand- The 1565-B is the successor to the 1565-A, long popular for
ard sound- level meter - it conforms to both national and inter- rapid measurements of plant, traffic and community noise.
national standards, meets all criteria necessary for the noise pro- The - B version is a total redesign to take advantage of the
visions of the Occupational Safety and Health Act, and includes experience gained with its predecessor and of the latest advances
most of the features usually found in larger, more cumbersome. in components and techniques - it is smaller, 40% lighter, and
and more expensive instruments. Yet the 1565-B fits in the palm easier to use. It offers 50% longer life on batteries that are read-
of your hand and operates in severe environments for up to 50 ily available. In common with the 1565-A, the 1565-6 is ap-
hours on self-contained batteries. There are no line cords to proved by the Bureau of Mines for use in gassy coal mines.
bother with or microphone cords to trip over, and an imaginative Performance and versatility built-in The 156511 uses a rug-
combination of controls permits one- hand operation and rapid ged, yet laboratory-quality, ceramic microphone that can be
interpretation of the result - just aim and read. checked easily, when necessary, by such standard calibration de-
vices as the GR 1562 Sound- Level Calibrator An output jack is
provided for use with headphone or recorders, and alock is pro.
vided so the range control can be fixed in asingle position. The
instrument is housed e a tough plastic case, tapered at the mi-
crophone end to reduce the effects of case diffraction, and rneets
all ANSI requirements for a Type 2 general-purpose sound- level
meter.
Easily- reed meter provides The 1583 is similar to the 156511 but is designed to meet
fast, accurate answers. the less stringent requirements for ANSI Type 3 survey meters.
Other differences include a pressure-calibration restriction to
1000 Hz and the inability to be adapted for use with vibration
transducers, an external microphone, or microphone wind-
screens.

specifications

Sound Level 40 to 140 dB re 20 pN/rn . Supplied Carrying pouch, miniature phone plug to connect to output.
%%166M A, B. and C 1565-B conforms to ANSI S1 4-1971 Type 2and screwdriver for calibration adjust. batteries
IEC 123.1961 1563 conforms to ANSI S1 4-1971 Type 3 Available For 1565-B only, when used with 1560-P96 Adaptor 1560-P52
Met« Rrns response with fast and slow speeds Vibration Pickup. 1560-P73 or P73B Extension Cable for remote micro-
phone connection
Input MICROPHONE Lead-zirconate-titanate ceramic. For 1565-B. a
1560.1,96 Adaptor converts input to 3- pin male A3 connector, for correct Pow« Two 9-V batteries (
Burgess 2U6 or equal) supplied, provide r• 50-h
weighting, source impedance must be 380 pF 5%. INPUT IMPEDANCE operation
rr 13 Mfl//15 pF Metheniml Shielded plastic case DIMENSIONS vvxhxd) 156543.
3.63x 6.5x2.09 in. ( 92 x165x53 mm). 1563. 3.63x5.3802 09 in.
Output: a 1.2 V rrns behind 6200 with meter at full scale, will drive 1556
Impact- Noise Analyzer, 1558 Octave- Band Noise Analyzer. 1521 or 1523 192.137 x53rn). WEIGHT 1lb 10 45 kg/ net. 3 lb 11.4 kg) shipping
recorders, oscilloscopes, or low- impedance headphones HARMONIC DIS- Catalog
TORTION 6 0.5% 10 1% typrcall from 32 Hz to 8 kHz, C-weighted with Number
meter at full scale
158543 Sound- Level Meter 1585-9702
Calibration I with 1562 Sound- Level Calibrator) 1565-8 can be acousti- 1583 Sound- Lev« Meter 1583-9701
cally calibrated at 125. 250. 500. 1000. and 2000 Hz. at 1000 He only for
Windaereens for 1 in microphone on 156511
1563
only, reduce wind noise and protect against
Environmental: TEMPERATURE —10 to 50.0 operating — 40 to 060*C contaminants pack of 4 15869521
storage, with batteries removed, coefficient of sensitivity •• —0.01 dBr C
UMW. ea. for 1565-0 and 1563 lose 2) 18410-3200
at 6 dB below full scale meter reading. HUMIDITY 90% RH. MAGNETIC
FIE LD • IOersted (80 Aim) 50- or 60- Hz field causes fr 45dB C-weighted
indication when meter is oriented to Maximum sensitivity to field

280
Type 1556-B IMPACT- NOISE ANALYZER

• measures electrical and acoustical noise peaks

• stores transient peak and time-average values

• 50-ps rise-time response

The 1556.9 Impact.Nolse Analyzer

attaches to the 1551.0 Sound- Level
Meter as shown here.

This device evaluates the characteristics of impact-type For these applications, the 1556-B operates from the
output of a 1551, 1561, or 1565-A Sound- Level Meter or
sounds and electrical noise impulses, which cannot be
satisfactorily measured with conventional noise-meters. 1558 Octave- Band Noise Analyzer and, when a vibration
Impact Noises include those produced by punch presses, pickup is used in place of the microphone, will measure
forging hammers, fire alarms, pile drivers, office machin- vibration impacts. It will also operate from tape re-
ery, and similar equipment. From the standpoint of hear- corders, and vibration meters.
ing damage, some of these sounds constitute a serious Electrical Noise Peaks in a wire communication circuit
problem for industry. They have hitherto been measurable can be measured with this instrument as one of the tests
only by complicated methods employing oscilloscopes. to determine the adequacy of the circuit for transmitting
The two characteristics of impact sounds that seem data pulses. In such measurements, many peaks may be
most significant are the peak amplitude and the duration, measured in a short time, and, after each peak, the stored
or decay time. This analyzer measures the: signal must be erased before the next pulse occurs. To
facilitate this a RESET pushbutton is provided, which can
• peak value, the maximum level reached by the noise,
also be operated by an ordinary camera cable release.
• "quasi-peak", a continuously indicating measure of the Circuit. A battery-operated, degenerative, transistor am-
high levels reached just before the time of indication, plifier simultaneously drives three ac voltmeter circuits,
and which comprise rectifiers, storage capacitors, and a dc
• time-average, a measure of the average level over a electronic voltmeter. The electrical storage system (a
predetermined period of time, which, when subtracted capacitor charged by a rectifier) makes it possible to mea-
from peak level, is a measure of the duration of the sure three characteristics of an impulse — peak, quasi-
impact. peak, and time-average — with asingle meter.

specifications
Input: Any voltage from 1 to 10 V for normal range. Inputs below Input Terminals: Cord with phone plug at one end.
1 V reduce the range of reading. Accessory Required: A sound- level meter, analyzer, or other cali-
Input Impedance: Between 25,000 and 100,000 0, depending on brated amplifier to supply 1556 input.
the setting of the LEVEL control.
Batteries: One 112 -V sizeD flashlight cell and one 45-V battery
/
Frequency 5 Hz to 20 kHz. are supplied. Typical battery life is 100 hours.
Level Indication: Meter calibrated in dB from — 10 to + 10. Atten-
uator switch increases range by 10 dB. Mounting: Aluminum cabinet; leather Carrying case supplied. Cab-
inet can be fastened directly to one end of a 1551 Sound-Levol
Peak Reading: Rise time is less than 50 os for a value within Meter.
1 dB of peak value ( for rectangular pulses). Storage time at
normal room temperature is greater than 10 s for a 1-de change Di11.11M011$ ( width o height a depth): 71
/
2 x 61
/
2 0 41
/
2 in. ( 190
in value. o 170 x 110 mm).
Quasi- Peak Reading: Rise time of less than /
4
1 ms and decay time Weight: Net, 41
/
2 lb ( 2.1 kg); shipping, 12 lb ( 5.5 kg).
of 600 -T.-120 ms for rectifier circuit.
Time-Average Reading: Charge time of rectifier circuit selected by
'seven-position switch, having times of 0.002, 0.005, 0.01, 0.02,
Catalog
Number I Description
0.05, 0.1, and 0.2 s for the resistance-capacitance time constant.
Storage time at normal room temperature is greater than 1 min 1556-9702 a155$- fl Impact-Noise Analyzer
for a 1-dB change in value. 8410-9590 Set of Replacement Batteries

281
GR 1944 NOISE DOSIMETER

GR 1944 Measuring true accumulated noise exposure The requirements

of the Occupational Safety and Health Act for measuring employee
Noise- Exposure Monitor noise exposure are not always met by single measurements of noise
levels at various work areas. Many workers tend to move about
during the workday and noise levels vary from area to area. General
Radio has a solution for these problems in its 1944 Noise- Exposure
Monitor, a small, lightweight instrument ( 7-1/2 ounces), which is
• small shirt- pocket size worn by the worker and mows with him during the entire workday.
Powered by a single 9-V battery, the 1944 Noise- Exposure
• lightweight, 7-1/2 oz
Monitor detects noise with its built-in microphone, weights it and
• tamper- proof accumulates total noise exposure for the workday, based on OSHA
criteria. Exposure to levels in excess of 115 dBA is also detected and
• conforms to applicable portions
stored by the 1944 for readout at the end of the day.
of ANSI S1.4-1971 Type 2
This instrument performs its measurements with absolutely no
sound- level meters effort on the part of the employee. Of equal importance, its light
weight and small size permit it to be worn with no hindrance to the
• long battery life — 300 h
employee's work. In addition, since its microphone is built in, there
are no exposed wires or microphones to snag, break or possibly foul
in moving equipment. Tamper- proof operation is assured since there
are no displays or controls on the unit. A concealed on-off switch is

MUNI
activated by a pin that is controlled by the supervisor of the noise
measurement program. The pin also serves to lock the battery in
place, and the single 9-V battery provides about 300 hours of
operation.
Displaying accumulated noise-exposure levels Retrieving the in-

/
formation detected and stored in the 1944 Noise- Exposure Monitor
is accomplished with the 1944 Noise- Exposure Indicator. The pro-

4
1 cedure is simple. At the end of the workday, the wearable monitor
is plugged into the indicator. You press a button and the cumulative
noise exposure will be displayed on a light -emitting diode readout.
The number displayed represents the accumulated percentage of
GR 1944 Noise- Exposure Indicator nome exposure a worker has experienced ... 100% being the maxi•
mum permissible in accordance with OSHA. A lamp on the indica-
• built-in sound- level calibrator checks complete system tor lights in the event that 115 dBA has been exceeded at any time
including microphone during the work period.
The 1944 indicator unit also provides pushbutton, acoustic-cali-
• readout available only to authorized persons
bration verification as well as a check of battery condition. Pushing
• all electronic including bright light-emitting- diode dis- two other buttons resets the memory in the monitor to zero, ready
play — no moving parts for the next days use.

SPECIFICATIONS Linearity: t0.4 dB at 1 kHz from 90 to 115 dBA with reference to the 115dBA
level.
1944 NOISE- EXPOSURE MONITOR Sunder*: Satisf ies all applicable sections of ANSI S1.4.1971 for Type 2 Sound.
Noise- Exposure Time: Maximum perrnissible exposure time of 100% in accord- Level Meters. including acoustical, amplifier, weighting and detection specific.-
ance with Occupational Safety and Health Act ( Public Law 91.596) is accura•
lead for the following combinations
Detector:' True rim response with "slow" speed in accedence with ANSI

tre aRL
etra
.v t: ra1tei xpOsure ( bones 1
,4, 4.11/
S1.4.1971. IEC Publication 123. 1961 and the Occupational Safety and Health
Act. Crest- factor capacity at full scale ( 115 dBA) is grater than 15 dB.
90 115 d8A Ecomdad: If 115-dBA sound level is exceaed, this information is
95 4 stored in the monitor unit and read out on the indicator.
100 2
Microphone: Ceramic tea.
105 1
110 0.5 Environmental: TEMPERATURE. - 10 to 0513° C operating ( 14 to 122° F), - 40
115 0.25 to 061f C storage with batteries removed. HUMIDITY Up to 90% RH at 40°C.
Sound level is interpola ad between the above points as shown in the graph. Power: One 9.V battery ( Burgess 206 or equivalent) supplied, provides about
The integrator cuts of1 sharp y below 90 d(34 ( see g aph). 3o0 h operation. MOS counter ar.d latch, which store data. are permanently
connected to battery and can store accumulated percentage and 115-dBA limit
for shelf life of battery.
meienicsh Shielded metal case. DIMENSIONS ( welted) 2.2>4.950.9 in.
(564124x23 nuo). WEIGHT • 7.5 oz ( 0.2 kg) net, 2 lb ( 1kg) shipping.
UI IIIIM111111MIIIIMMIIMI 1944 NOISE- EXPOSURE INDICATOR
III IMBRIUM« Readout: The indicator reads from 1% to 999% on light-omitting-diode display in

II I11.1111111111
1% steps, showing percentage of permissible exposure time with 100% represent-
ing the legal limit. Monitor latches at 999% for accumulations exceeding this
amoum.
115 deft Exceeded: A light shows when 115 dB A.weighted sound level has been
exceeded ( stored in 1944 Monitor).
Mk' Celibration: A sound- level calibrator is an integral part of the indicator. The
calibrator produces a tine sound leal for a period of 45 scorresponding to en
UM . 30.1111111.1111 accumulation of 5%.
Maury Check: A meter à provided for checking batteries in the monitor and
indicator units.
11111M11111
Environmental: TEMPERATURE - 10 to 05(1° C operating ( 14 to 122° F) - 40 to
,IBMIMMIM1
.ecr cstorage with batteries removed. HUMIDITY: Up to 90% RH at 40. C.
Power: Four 1-1/2.V "C" yells supplied. provide 20 h C01111 ,11.101.13 life. In Memel
service, life is typically 6 months or longer.

Horami
s Mecheniml: DIMENSIONS lwahxd) 1147.3x2.1 1279.086>M rnrn). WEIGHT:
39 lb 119 kg/ net, 6 lb 12.6 kg) snipping.

- -
Catalog
Description Number
• -
1944 Noise.Expoture Monitor 1944.9701
Sound-Level Range 90 to 115 dBA It 20 wN/n1 . 1. Batte.y. tor 1944.9701 ( use 1/ 8410.3200
Weighting "A" in accordance with ANSI Standard S1.4.1971 Type 2 and IEC 1944 None.Exposure Indicator 1944-9702
Publication 123.1961. Battery. spare for 1944.9702 use 4) 8410-0100
Amormy: At 115 dB. 1 kHz. 23° C - t 0.5 dB. at 115 dB. 1 kHz over the
temperature range - 10 to + 5( ° C 114 to 122° P1 el dB Pending.

282
GR 1937 MANUAL AUDIOMETER
The GR Model 1937 Audiometer is designed ti,
provide maximum operating ease, accuracy, and
reliability to the audiometric technician who ad-
ministers screening tests. All 1937 controls are
carefully arranged for minimum operator motion,
with the tone presentation bar easily accessible to
the hand manipulating the HTL Control. An inter-
com system is built-in so the operator can rein-
struct the subject by merely pressing abutton and
talking. The level of the intercom is controlled
by the HTL Control,
Attenuator: Low- noise, thick-film, continuously'
The HTL Control employs aunique, continuously- variable, with detents at 5-dB intervals.
variable, thick- film attenuator with detents at 5-dB
intervals — permitting more precise determination Case: High- impact, shatter resistant.
of threshold than is possible with conventional
stepped attenuators. Use of state-of-the-art inte- Size: 14iew x 13'4"d x6'4"h.
grated circuit design assures long-term stability
and reliable operation. The unit is packaged in an Supplied: Headphones, 50 Audiogram forms.
attractive and rugged plastic carrying case with
ample space for earphones and accessories. Available: 3-wire patch cord to connect earphone
output to Audiometric enclosure.
Specifications
Weight: 10 lbs net, 16 lbs shipping.
Power: 115-230 V ( switch- selected). 50-60 Hz, 3W.
Catalog
Test Frequencies: 250, 500, 750, 1000, 1500, Description Number
2000, 3000, 4000, 6000, 8000 Hz 1937.9700
1937 AudiOmeter
3-
Wire Patch Cord ( used to con-
HTL Range: — 10 to +90 dB at all frequencies
nect earphone output to audio-
1937.9600
metric enclosure).
Calibration: ANSI- 1969, ISO 1964

G-S1703 RECORDING AUDIOMETER

The Grason-Stadler Model 1703 Recording Audiometer

is designed especially to implement simple, rapid and
accurate pure- tone hearing tests in situations such as
employment screening and the ongoing monitoring es-
sential to hearing conservation programs. The unit is
manufactured with special attention to reliability, sim-
plicity of operation, and permanence of calibration. We
brought to its design not only 20 years of experience
with psychoacoustic instrumentation, but specific elec-
trical and mechanical assemblies and procedures already
proven in our highly respected diagnostic audiometers.
This, coupled with the expanded capability provided by
our association with General Radio, means better ser-
vice and higher reliability than ever before.
Control: START button initiates test or continues test after stop. STOP button
Signal: Pura Tom, 7frequencies automatically end sequentially presented to the interrupts signals, sets intensity to min NIL, and permits .snual frequency
left then to the right ear. 51)0 Hz, 1, 2, 3. 4,6, and 8kHz,eutornetically retests selection. HOLD button maintains constant frequency and leaves subject in
right ear at 1 kHz at end of sequence. FIfflefer Accuracy: t3%. Timing and control of signal intensity.
Control: After START button is pushed, each frequency is presented for 30 s. Display: 5e8 in. 1127x203 min) audiogram of HTL us frequency. automatically
HOLD button pernXts threshold deternsinatiOn at frequenCy selected and STOP and continuously plotted.
button permits manual Selection of • diff frequency.
Communications: Subject handsvetch.
Intensity: Varies automatically and continuously from — 10 to 090 dB Hearing Supplied: Calibrated subject earphones. SUPjeCt handswitch, 1000 audiearam
Threshold Leoni re 1964 ISO Stnsdard. Accuracy: t3 dB from 500 Hz to 2kHz,
forms, 12 blue fiber-tip pens, power and patch cords.
±4 dB to 4 'alt.
- dB to 8 kHz. Timing and Control: Signal is automatically
pulsed 200 no on and 200 no off or can be presented continuously. Intensity
slutomenullY changes at 10 dB/s until first subject response. 5deis until fourth
response, then at 2.5 dBls thereafter; after fourth response, any failure to re- Number OMM,101 ,011
spond for 5$ returns intensity change rate to initial 5dB/s. At end of test or when
STOP is pushed, intensity n set to min HTL and subject control of intern./ is 1703-9700 11703 thoordine Audiometer
removed.

283
 GR 1562-A and 1562-Z MICROPHONE
and AUDIOMETER CALIBRATORS

• 125 top 2000 Hz

• !A:1.3-dB accuracy
• fits many microphones
• approved by Bureau of Mines

GR 1562-2
Audiometer GR 1562-A
Calibration Set
,i00)Sound-Level Calibrator

specifications

1565-2 AUDIOMETER CALIBRATION SET

1562-A SOUND- LEVEL CALIBRATOR
Supplied: 1565 Sound Level Meter, 1562 Sound- Level Celrbrator, earphone
oupler , spare batteries st orage case Acoustic Output: FREQUENCIES 125, 250.500, 1000, and 2000 Hr. 12%.
Mechanical: DIMENSIONS ( wshxd) 11.25s4.25.10 in ( 286x108x254 SOUND PRESSURE LEVEL, 114 dB re 20 pN/rnT, accuracy at 23 °C and
um,L WEIGHT 51512 3kg/ net. 12 11,16 kg) shipping. 760 elm Hg is, tor WE 640AA or equivalent microphone, 7:0.3 dB at 500 Hz
and 710.5 dB at other frequencies and, for other microphones. /10.5 dB at 500
EARPHONE COUPLERS Hr and ±0.7 dB at other frequencies Temperature coefficient of output
sound level is 0to - 0012 dBP°C, pressure correction chart supplied.
Electrical Output: 1 V ±2% behind 6 kit. flat 2
-2%with < 0.5% distortion.
available at phooe tack.
Environmental: TEMPERATURE 0 to • 40°C operating. - 24 to 055°C
non operating HUMIDITY 95% RH and + 30°C. VIBRATION 0.03 in.
from 10 to 55 Hz. SHOCK 30G, 11 ms. DROP. 30m.
Supplied: Carrying case, adaptors for 5/8. and 1- in. microphones ( lits 1.1/8-
in. microphones:without adaptor), battery .
Available: 1560.9561 Coupler Adapt•
or Set. for coupling 1562 to 1/8. 1/4,
and 1/2- in. microphones.
Power: Battery operated 19 V. Burgess PM6 or equal). 120 huse.
GR 1560-P131 and -P82 GR 1560 P83
Mechanical: DIMENSIONS. 5 in. 1128 mm) long s 2.25 in. ( 57 mm) dia.
WEIGHT 11610.5 kg) net.4 lb 11.9 kg) shipping.
Typa: ANSI Type 1. VOLUME GR 9A I modified version of
6 crn 3 including NBS type 9-A). VOLUME. One r
Ipt lon
equivalent volume of 5.642 cm 3 including volume
15629701 1562-A Sound- Leval Calibrator
microphone AXIAL added by microphone. AXIAL
HOLDING FORCE HOLDING FORCE 450 1560.9561 Coupler Adaptor Set, adapts 1562 A to 1/8.
500 grams, grams nominal 1/4. and 1/2 in microphone
Frequency: 125 Hz to 8 kHz audiometric frequencies, re- 8410 3000 Battery. for 1562-A lutes IT replacement
sponse is equal to NBS9A Coupler within 1dB to 11565-2 AudionteMr Calibration Sat
4 kHz and 1.5 dB to 8 kHz when used with TDH- 1565-9900 with 1560-P82 Earphone Coupler
39 or TDH.49 earphone in MX-41/AR earcushion. 1565-9901 with 1560-P83 Earphone Coupler
Mechanical: 1560-P81 and -P82: DIMENSIONS Coupler. 2.25 in dia is 1.06 1560.9681 1560-P81 Earphone Coupler. ANSI type 1.
M. high ( 57s27 nms), over-all lwxhiecl), 2.2503,13 in. 1571(76)(76 rnm). I1/8 in
WEIGHT. 0.5 lb ( 0.3 kg) net, 2lb ( 1kg) shipping. 16804,83: DIMENSIONS: 1560-9682 1560-P82 Earphone Coupler. ANSI type 1,
Coupler. 2.94 in. dia u 1.25 in. high ( 75)(32 trim), oser - all ( wkhxd).
2.94k3.5x3.5 in. 175090c90 mrm IG iIT 0.5 1010.3 kg) net. 2 ' bit kg) 1560-9683
Shipping. 1560 P83 Earphone Coupler. GR type 9A
84 10-9591 Battery Set ror 1565.Z ( uses 1). replacement

GR 1567 SOUND- LEVEL CALIBRATOR

Economical Calibration The 1567 Sound- Level Calibrate ,

is specially designed for you who wish to check instrument sen
sitivity only ) not frequency response). The 1567 has a single-
level 6utput and is well suited for calibrating the 1563 and
1565-8 Sound- Level Meters at 1 kHz.

SPECIFICATIONS

Acoustic Output: FREQUENCY: 1000 Hz. ± 3%. SOUND-PRES-

SURE LEVEL: 114 dB re 20 pN/rTe; accuracy ( at 23° C and
760 mrn Hg) is 00.5 dB for 1565-B, i1 dB for 1563.

Environment: TEMPERATURE: 0 to 55° C operating. Tempera- Mechanical: Cylindrical housing. DIMENSIONS ( die n h):
ture coefficient of sound-pressure level is zero ± 0.01 dB/° C. 2.38x4.44 in. ( 61x113 mm). WEIGHT: 1 lb (0.5 kg) net, 4 lb
(1.9 kg) shipping.
0 to 23"C; - 0.017 o 0.008 clEIrC, 23 to 50° C. Pressure-correc-
lion chart supplied. HUMIDITY: 0 to 95% RH,
Catalog
Suppiied: Carrying case, adaptor for 1565-B, battery. Number DmmrlptIon
Power: Battery operated, using 9-V Burgess 2U6 or equiva- 1567-9701
lent; 10011 of use. 1567 Sound- Level Calibrator
13410-3200 Battery, spare required)

284
Product Noise Reduction and Testing

Type 1551-C SOUND- LEVEL METER

• 24- to 150-dB measurement range

• meets common standards:
ANSI Standard S1.4-1961
IEC Publication 123, 1961
• 20- Hz to 20- kHz amplifier response
• internal calibration system

specifications

Magnetic Fields: When exposed to a 60- Hz. 1-Oersted ( 80 A)m)

Sound- Level Range: From 24 to 150 dB ( re 20 ,., N mt). field, the sound- level meter will indicate 60 dB ( C weighting)

Frequency Characteristics: Four response characteristics. A. B. when oriented for maximum sensitivity to the magnetic field.

C. or 20 kHz. as selected by panel switch. The A.. B-, and C- Electrostatic Fields: Aluminum case provides sufficient shielding,
weighting positions are in accordance with ANSI Standard 51.4- so that normally encountered electrostatic fields have no effect.
1961 and IEC Publication 123, 1961. Frequency response for the Vibration: Case is fitted with soft rubber feet and amplifier is
20- kHz position is flat from 20 Hz to 20 kHz, so that complete resiliently mounted for vibration isolation. When the instrument
use can be made of very wide-band microphones. is set on its feet on a shake table and vibrated at 10 mils pk-pk
displacement over the frequency range of 10 to 55 Hz, the un-
Microphone: GR Type 1560-P5.
wanted signals generated do not exceed an equivalent C- weighted
Sound- Level Indication: Sound level is indicated by the sum of the
sound-Pressure level of 45 dB when motion is vertical. 60 dB
meter and attenuator readings. The clearly marked, open-scale
when motion is lengthwise, or 40 dB when motion is sidewise.
meter covers a span of 16 dB with calibration from — 6 to + 10 dB.
The attenuator Is calibrated In 10-dB steps from 30 to 140 dB
GENERAL
above 20 » Him ,. Power Supply: Two 144-V size D flashlight cells and one 67W.V
Calibration Accuracy: When amplifier sensitivity has been stand- battery ( Burgess XX45 or equivalent) are supplied. An ac power
ardized, the absolute accuracy of sound- level measurements at supply, the ryPe 1262-B. is available.
500 Hz is within t- 1 dB and at all frequencies is in accordance
Acc sssss ies Supplied: Telephone plug.
with the ANSI Standard.
Accessories Available 1551.P2 Leather Case ( permits operation of
Panel adjustment is provided for standardizing amplifier gain
instrument without removal from case), 1562 Sound- Level Cali-
with internal calibration circuit.
brator. 1560-P95 Adaptor Cable for connecting output to 152111
Absolute acoustic sensitivity is factory calibrated at 500 Hz.
Graphic Level Recorder.
Microphone response and sensitivity are measured in a free field
from 20 Hz to 15 kHz by comparison with a WE 64066 laboratory- Mounting: Aluminum cabinet.
standard microphone with calibration traceable to the National Dimensions (
width • height o depth): 744 944 x 6Ve in. ( 185 •
Bureau of Standards. Complete electrical frequency.resPonse
235 a 160 mm).
measurements are made on each instrument.
Weight. Net. 73
/
4 lb ( 3.6 kg): shipping, 16 ; b ( 7.5 kg), batteries
The 1562.A Sound- Level Calibrator or the 1559-B Microphone
incl. Add 2 lb for leather case.
Reciprocity Calibrator can be used for making periodic over-all

acoustic checks. 1262-8 POWER SUPPLY

Output: 1.4 V behind 7000 U ( panel meter at full scale. The out- Attaches to the 1551-C Sound- Level Meter for ac- line
put can be used to drive analyzers, recorders, oscilloscopes, and
headphones. Harmonic distortion ( Panel meter at full scale) 1%. operation.
Power Required: 105 to 125 or 210 to 250 V, 50 to 400 Hz, 2 W.
Input Impedance: 25 MU in parallel with 50 pF.
Dimensions ( width x height o depth): 5 x 7Va o 31
/
4 in. ( 130
Meter: Rms response, and fast and slow meter speeds in accord-
185 x 80 mm).
ance with ANS S16-1961 and IEC 123. 1961.
Weight: Net, 21
/
2 lb ( 1.2 kg), shipping, 8 lb ( 3.7 kg).

Catalog Description
Environmental Effects Number

Temperature and Humidity: Microphone is not damaged at tem- 1551-9703 1551-C Sound- Level Meter

peratures from — 30 to 1-95 .0 and relative humidities from 0 to 8410-9499 Set of Replacement Batteries

100%. When standardized by its internal calibration system or a 1551-9602 1551-P2 Leather Carrying Case
1562 Sound- Level Calibrator, the instrument will operate within 1262-11 Power Supply
1262-9702 115-V Model
catalog specifications ( for panel- meter indications above 0 dB)
1262-9705 230-V Model
over the temperature range of 0 to 60 .0 and the relative humidity
1560.9521 Windscreen, for 1- in microphone.
range of 0 to 90%. Per pack of 4

285
GR 1933
PRECISION SOUND- LEVEL METER and ANALYZER

The 1933 reflects 50 years of GR expertise and leadership in acoustic measurements ...
... it has no equal.

• three instruments in one

precision sound- level meter

impulse sound- level meter
precision octave- band analyzer

• compact, light- weight, and fully portable

• unique auto- ranging permits one- knob operation

• can be used with low-cost dc recorders

• easy- to- read meter display

• low cost and high reliability due to unusual design

and efficient construction

• microphone readily extends away from operator

A precision sound laboratory The 1933 is a precision meas- an octave- band analyzer. In other analyzers, two are required: an
uring instrument that conforms to all standards for aType 1preci- input range control to set the -all-pass" level and an analyzing
sion sound- level meter, a precision octave-band analyzer, and an range control to provide an on- scale meter indication after the
impulse sound- level meter. It measures true rms values ( there are desired octave filter has been selected. Both are necessary to obtain
no approximations) and warns of invalid readings due to overloads the maximum analyzing range and maximum dynamic range, but
(a panel- lamp is energized from monitors located at two critical in the 1933 a unique automatic attenuator system is used. With
points). this feature, you need only set asingle range control for an on-scale
indication. A second manual control is also provided in cases where
the automatic system rnay not be suitable, as with some measure-
ments of transient signals.
The unusual meter scale also enhances the operating ease. It
spans a full 20- dB range, is graduated linearly over the entire range.
and displays the attenuator setting directly on the meter face. This
reduces the number of range changes necessary and aids in rapid.
error- free interpretations of the readings.

A versatile sound laboratory Set the upper knob to Weighting

and the 1933 becomes a sound- level meter with a pushbutton
choice of A. B. or C weighting or flat response from 5 Hz to 100
kHz. Fast and slow meter speeds are also pushbutton selected. and
another button allows impulse testing according to the proposed
amendment to IEC 179.
Conversion to an octave- band analyzer is equally simple, turn
the knob to the desire(' octave band — there are 10 to choose from.
with center frequencies from 31.5 Hz to 16kHz.
A simple sound laboratory A single control is all that is re-
quired to set the meter range, even when the instrument is used as

A self-contained sound laboratory The 1933 operates for up

to 20 hours on self - contalned batteries. A companion instrument,
the GR 1940 Power Supply and Charger, allows the analyzer to be
operated from the ac line and provides rechargeable batteries and a
charging circuit.
Two microphOnes are available to suit awide range of measure-
ments — a '
h- inch microphone for the smoothest and widest fre-
quency response at moderate and high levels, and a 1- inch micro-
phone for the lowest levels. Both are laboratory-quality, ceramic
microphones designed for flat random- incidence response and
noted for reliable operation even under extreme conditions.

286
 of the recorder. On playback, the level- range setting used for the
sound- level meter is indicated by a digital display on the panel of

the tape recorder. Thus, the tape stores both data and absolute

level information.
A dc output. proportional to the meter deflection in dB, is

provided to drive a low-cost dc recorder for hard-copy records of

the level vs time. This output has a dynamic range of 40 dB plus an

additional 20-dB crest- factor allowance.

An economical sound laboratory The design of the 1933 not
only produced an incredible degree of capability in a small light-
weight instrument. it also resulted in considerable cost savings. A
combination of commercially available integrated circuits and
hybrid circuits ( manufactured by General Radio) accounts for
much of the circuitry. High-cost hand-soldering and conventional'
cabling operations have been virtually eliminated. The result is a
high- caliber instrument for an unusually low price.

The mid, , - 1 atop a telescoping 18- inch extension to

reduce the effects of the instrument and operator on the sound
field — there is rarely a need for extension cables and tripod. Where
these are necessary, however, a 10- foot cable is provided and a

60- foot cable and tripod are available. Measurements are unaffec-
ted by the cables because the preamplifier in the 1933 is detach-
able and connects to the cable at the microphone end to prevent

signal loss.
An expandable sound laboratory Often it is desirable to record
field measurements for further analysis later in the laboratory.
With the 1933 and its companion recorder, the 1935, such record-

ings are exceedingly simple and accurate. In addition to an ac-

signal output to drive one channel of the recorder, the 1933 also
provides a level- range-code signal that is stored on a second channel

tentative specifications

the ': it• Hireau of Standards ON SITE Built in calibrator provides

Specifications meet ANSI SI 4 1971 for Type I ( precision) Sound Level Meters. '• • circuits. GR 1562 Sound Level Calibrator ' saveable for
IEC 179-1965 for Precision Sound Level Meters. 0- C 123-1961 for Sound Level " • all cal brat.° including microphones
Meters, ANSI SI 11-1966 lut Octave. Half Oct ave, and Third Octave Band Tv,
Environmental MPF RA TURI • IC to l'50): C operatimt. 40 to 160"C
0 Class n Filler Set, and IEC 225 1966 tor Octave, Hall Octave, and Thad
SIIII.191' removed HUMIDITY 0 to 9crx. RI-I VIBRATION
Octave Band Films for the Analysis of Sounds and Vibrations
AND MIt • • ICS Conform to apple Mile ANSI aol It C standards
Lwitel , 10 to 150 dB., 20 uNirn' with 1 in microphone 30 to 150 dB with '4 in
microphone Supplied: I.., 19330600 101, microphone e 0,51(111 cable, miniature
Frequency: 5 Fir to 100 kHa essentially fiat response. 10 octave bands with
phone Plog
center: flequencies from 31 5 Hz to 16 kHt. plus A B. and C weighting Power - 1 nu n rechargeable C cells supplied tirovide v 20 h opera
Display: ME TER 20413 scale linearly marked in dB and lower, center, and • ' • ...• )1upply ant Chigger allows line operation of 1933 and includes
upper values automatically indicated on scale RESPONSE Fast. slow. and r. • • • • iim arid charging sower , Flattery i-hick provided on 1933
impulse ( per IEC 179 amendment) pushbutton selected Precise inn detection Mechanical • i rugged hand held case with standard 0 75 20 threaded hole
for signals with u 20 dB crest factor at lull scale ¡Yin, factor rape'', greater lor I DIMF.NSIONS ( w.h.d) 6 25 9 3in 1159 279 76
below lull scale OVI RLOAD Signal peaks monitored at 2 critical Pionll to
mini WE IGH f 5 511,12 5 kg) nett° lbI4 6 kg) Shipping
provide positive panel lamp warning RANGING A utornat lc system ma ir tes
Catalog Nunlber
analyzing range and signal to- noise ratio lor each level range control setting,
manual control provides override 1933 PrIleifiel Sound- Level Meter end Anelyrer
With loin and Iin Y11( rophoriii. 19319700
Filters: WEIGHTING A.B.C. and flat. pushbutton selected OCTAVE BANDS
1933 9701
10 manually selected. with 3 5 e_ 1-dB attenuation al nominal 18 dB With Yi ai onlV
attenuate. al Y. and 7X center hr.:pen,. 70dB ultimate attenuation 84100100
Battery. spare for 1933, uses 4
EXTERNAL FILTERS can be substituted toe internal weighting networks a,1
octave- band filters. connect 1° 2miniature phone lacks 19319601
Microphone extension cable. 60 1t
19319602
Ingot, 1
/r or 1 - in ceramic microphone with flat randorn-incidence response. Miniature phone plug to 1933 microlil'-
mountHi with detachable preamplifier on IEt in extendable mast, or on 10- ft 15609677
Miniature phone plug tr, double banaol
156e9679
extension cable supplied, or on 60 lt cable available input can also be from tape Miniature phone plug to EINC.
1560-9678
recorder INPUT IMP, DANCE IGlill< 3ge Miniature phone plug to standard phone
Output: SIGNAL OUTPUT 0 5 V ins behind 600 I) corresponding to lull scale' yaguiecreene. reduce wind now'. protect agell• '
meter deflection, any load permissible RANGE CODE Contact closure pro contaminants
vides sound level meter range information to 1935 Cassette Data Recorder For 'h in microphone, set of 1 1560-4521
DF IL CTED OUTPUT 45V di - behind 45 kSI corresponding to lull scale For l- in microphone. set of 4 1560-9522
meter deflection output is linear in dB a, 0 IV/dB over 60 dB range 140-dFl 1562-A Sound- Level Calibrator 1562-9701
normal range plus 20,113 crest factor allowance) any load permissible Complete sets with companion instruments and accessories are also offered,
Celibretien , FACTORY Fully tested and calibrated to all sPeciflcations• 00005
see next papa
',cal response and sensitivity are measured in a tree held by comParison With it
WE 6.10AA Laborato, Standard Microphone whose calibration is traceable to

287
GR 1935 CASSETTE DATA RECORDER
This is one of the smallest and most capable data recorders available today.

• excellent companion instrument for the 1933 Precision

Sound- Level Meter and Analyzer

• cassette simplicity

• battery operated and completely portable

• two channels, each with peak monitors for

fool- proof recordings

• memorizes sound- level- meter attenuator settings

Cassette simplicity The 1935 is atwo-channel, two-track mag-

netic tape recorder with aPhilips Cassette format. In use with the
1933 Precision Sound- Level Meter and Analyzer, the 1935 records
the signal on one channel while on the other it records the setting
signal from a second sound- level meter in lieu of the range data.
of the sound- level- meter range control. Thus the signal and range
Each channel includes apanel monitor that displays the absolute
information given on play-back are identical to those available
peak level, regardless of polarity. so that recordings are virtually
from the sound- level meter at the time of the original recording.
fool- proof.
Voice notes can also be recorded by breaking into the range code.
Included with the 1936 are all necessary accessories — tape,
Although intended primarily for use with the 1933, the
microphone for voice notes. tape- head maintenance set, and inter-
recorder can also be used with any instrument that provides an
connecting cables to the 1933.
output of from 0.4 to 2V rms. In this case, the range information is
entered manually. If desired, the second channel can record the

GR 1940 POWER SUPPLY and CHARGER

Valuable companion The 1940 allows either the 1933 Preci•
sion Sound- Level Meter and Analyzer or the 1935 Cassette Data
Recorder to be operated from an ac line, independent of their
internal batteries. It is supplied with five rechargeable cells Ito
replace the ordinary C cells supplied in the analyzer or recorder)
and abattery charger There are no internal connections to make;
the instruments simply plug into the 1940 and are supported at a
convenient angle for bench- top operation,

tentative specifications
Power Sarre.: 5 V for me operation of 1933. 6 5 V for lone operation of 1935.
250 inA max
Charging Source) 200 inA max for charging batteries in 1933 or 1935. Arromar
'call, reduces to w 30-mA trickle charge when batteries are charged Charging
tome n 16h
StfIefeelf 5 rechargeable nickel.cachnium C cells to replace nonxechargoable
batteries in 1933 or 1935
Pam) 100 to 125 or 200to 250V.50to 400 Hr.11 W Catalog
Mechanical) DIMENSIONS lefxhxd) 4 38k4 25.9 44 in Ill 1.106x 240 rnrn)
Number
WFIGHT 3 516 ( 15 kg) net. 5Ib 12 3 kg) shipping 1940 Power Supply end Charger 11M0.9701

GR SOUND- ANALYZER SYSTEMS

Fos the ultimate in convenience. the 1933 Precision Sound- The 1933 Precision Sound- Level Meter and Analyzer is
Level Meter and Analyzer is offered in two systems. Each system is included in both systems, together with the Ya - inch and 1- inch
neatly housed in arugged luggage- type case that provides room not microphones, a 60- foot extension cable and tripod for remote
only for all instruments and their accessories but for all documen- microphone installations, a windscreen for each microphone to
tation as well, with space left over for incidental accessories the reduce wind noise and to protect against contaminants, aheadset.
user may wish to include. and a GR 1562-A Sound- Level Calibrator for on-site calibration
One system includes the 1935 Cassette Data Recorder with all checks.
its accessories; the other does not include the recorder or its acces-
sories.

288
Type 1553 VIBRATION METER

• direct reading in acceleration, velocity, displace-

ment, and jerk
• 2 to 2000 Hz ( 120 to 120,000 rpm) to 20,000 Hz à
with suitable pickup SEL[C,OR

• portable, battery operated, simple to use

specifications

Ranges of Measurement:
Feat to Peat Average
Frequency
Type Re. Quantity Min Max Min Max Units Range ( Hz)

Acceleration 0.3 300.000 0.03 30.000 in./s) 2-2000

Velocity 0.03 30.000 0.003 3,000 in./5 2-2000
1SS3-A Displacement 3 300,000 0.3 30,000 mils 2-2000
Displacement 0.03 30,000 0.003 3.000 mils 20-2000
Jerk 30 300,000 3 30,000 indS ) 2-20

Acceleration 0.01 10,000 0.001 1.000 mar 2-2000

Velocity 0.001 1.000 0.0001 100 m/s 2-2000
1853-AK Displacement 0.1 10,000 0.01 1,000 mm 2-2000
Displacement 0.001 1,000 0.0001 100 mm 20-2000
Jerk 1 10.000 0.1 1.000 m/s , 2-20

Accuracy: -
± 10% of full scale. Mounting: Flip-Tilt Case. Rack-mount versions also available.
Input Impedance: 25 MO. Dimensions (width o height o depth): Portable model. 8 0 944 x
7,6 in. (205 x 235 o 190 mm); rack model, 19 x 10 1
/
2 x 5 in. ( 485
Voltage at Output Jack: 5 V rms. behind 75 KO for full-scale
o 270 x 130 mm).
deflection.
Attenuators: A 10-step attenuator changes the meter-scale range Net Weight: Portable model, 10 1
/
2 lb (4.8 kg): rack model, 14 lb
by a factor of 100,000 to 1. Window readout indicates full-scale (6.5 kg).
values and units ( 10 times full-scale for AVERAGE readings).
Shipping Weight: Portable model, 14 lb ( 6.5 kg); rack model, 31 lb
Calibration: Internal. (14.5 kg).
Allowable Pickup Sensitivity for Direct Reading: 30 to 150 mV/g.
Terminals: A panel jack is provided for plugging in earphones.
1564-A Sound and Vibration Analyzer. 1556-B Impact- Noise Ana-
lyzer, 1538 or 1531 Strobotaco electronic stroboscope, 1568-A
or 1900-A Wave Analyzer, or an oscilloscope.
FCC Description

Power Supply: Portable model, 3 size-D cells and one 67 ,6-V 13534 Vibration Meter, English Units
1553.9701 Portable Model, with batteries
battery ( nor Type )0(45 or equivalent) supplied. Typical Portable Model with 115-V ac supply
1553-9710
battery life, 7 days at 8 h per day. For ac operation, use Type 1553-9711 Portable Model with 230.V ac supply
1262-C Power Supply ( listed below). Rack model, Type 1262-C 1553-9550 Rack Model with 115-V an supply
Power Supply is included. 1553.9551 Rack Model with 230N ac supply
1853-AK Vibraban Meter, Metric Units
Accessory Supplied: 1560-P52 Vibration Pickup. 1553.9819 Portable Model, with batteries
Accessories Available: 1560-P35 Permanent-Magnet Clamp: 15574 1553-9820 Portable Model with 115-V an supply
1553-9821 Portable Model with 230-V ac supply
Vibration Calibrator, high-frequency pickup 1560-P53, and high- 1553-9560 Rack Model with 115-V ac supply
sensitivity pickup 1560-P54. 1553.9561 Rack Model with 230-V ac supply

289
Type 1557-A VIBRATION CALIBRATOR

• calibrates vibration pickups, meters

• generates 1gat 100 Hz
• portable, battery-operated

The calibrator provides a single-frequency ( 100 Hz), brator appears at two pillbox-shaped, 50-gram disks
single- level ( 1 g) check on the GR Vibration Pickups, the mounted on an internal cylinder that projects through the
1553 Vibration Meter, or any pickup whose total mass is sides of the instrument.
300 grams or less. It can provide on-the-spot calibration
of vibration-measuring systems immediately before and
after important measurements and can also be used to
compare transducers or to calibrate working transducers
against astandard transducer.
Operation of the calibrator is simple. A pickup of
known mass is attached to the shaker, either in place of
one of the removable 50-gram disks or to one of the disks
by double-faced, pressure-sensitive tape. The user adjusts
the LEVEL control until the panel meter, calibrated in View of the calibrator with
Type 1560-P52 Vibration Pick.
grams, indicates the mass of the pickup. The pickup will up attached.
then be automatically subjected to an acceleration of 1 g
at 100 Hz,
The 1557-A is a small, battery-operated unit consisting
of a transistorized electromechanical oscillator and a
cylindrical shaker. The acceleration output of the cali-

specifications
OUTPUT Accessory Supplied: Leather carrying case.
Acceleration: 1 g rms = 10%. I g = 386 in./sr (9.81 rn/W). Mounting: Aluminum case.
Velocity: 0.614 in./S ( 15.6 mm/s) rms. Dimensions (width x height o depth): 4 x 8 o 4 in. ( 105 x 205 x
Displacement: 0.000978 in. (0.0248 mm) rmle; 0.00277 in. ( 0.0704 105 mm).
nun) peek. Weight: Net, 3% lb ( 1.5 kg); shipping, 51/
4 lb ( 2.4 kg).

Freq 100 Hz = 1% for 50-gram load; 100 Ho -I-0, —2% for

300-gram load. Catalog
. Description

lotteries: Four RM-4 ( or equivalent) mercury cells. Battery life 1557-9701 185741 Vibration Calibrator
is 100 hours of continuous operation. ( Dry cells optional; pisase 8410-1372 Replacement Mercury Coll, 4 reci'd
specify.) 8410.1050 Replacement Dry Cell, 1 reted

290
 lui 1111181111111
e

1523 Graphic Level Recorder

• 1- Hz to 500- kHz frequency range Precise 1/3-octave- band analysis for
• product- noise reduction
• 100-µV sensitivity • plant- and field-noise studies
• up to 100- dB dynamic range • materials testing
Insert the 1523-P3 Stepped 1/ 3-Octave- Band Analyzer
• up to 0.1- dB linearity plug-in and your recorder becomes a 1/ 3-octave-band
• 4recorders in one with plug-in versatility analyzer with a frequency range of 1Hz to 80 kHz ( ANSI
standard bands 0 through 49.) You can perform analysis
• remotely programmable — asystems natural
on selected portions within the full range, and you can
switch in an all- pass channel to display the over-all level
Automatic measurements— simply and graphically The at the start of a record. You can also select any one oi l1
1523 is not just another recorder; it is a measurement automatic programs to provide recordings with constant
center. It incorporates the latest refinements of the re- averaging times or, as a unique feature, to provide anal-
corder field with those of the sweep-oscillator and sound- yses with statistical confidences of up to 90% to within
analyzer fields and does so in one instrument that elim- ±-0.5 dB. You can record measurements singly on a new
inates the usual bother of trying to keep everything syn- chart, successively on a single chart for comparison, or
chronized. Simply connect your signal or device, set up successively on new charts, simply by the turn of asingle
the desired measurement conditions, and push a button knob.
— the 1523 does the rest, automatically and without
constant attention or control manipulation. Simple response measurements for
• filter and network response testing
Narrow- band wave analysis for • loudspeaker, amplifier, and tape- recorder evaluation
• product-noise reduction programs • performance tests for microphones, hydrophones,
• spectrum-signature work and hearing aids
• vibration studies • general medical and educational applications
• preventive maintenance programs With the 1523-P2 Sweep Oscillator plug-in, which in-
• distortion measurement corporates a sweep oscillator, your recorder produces
• network analysis frequency- response recordings at the push of a button.
The 1523-P4 Wave Analyzer plug-in gives you the capa- You can set the oscillator to sweep the full 1- Hz to 500-
bility to perform high- resolution spectral analysis, swept- kHz range, or various portions of it, at output levels con-
frequency analysis with a tuned detector, and amplitude- tinuously adjustable from 500 N to 5 V behind 600 Q.
vs-time measurements at selected frequencies. A unique and versatile constant-Q mode of operation can
The 1523-P4 provides analysis bandwidths of 10 Hz be selected to speed the recording in many applications
and 100 Hz and an all- pass mode covering the range from by increasing the sweep rate automatically as the fre-
10 Hz to 80 kHz. Most signals can be handled without quency increases. Under many conditions, recordings
preamplification, because the analyzer will accommodate can be made in the constant-Q mode in V2 to V3 the time
inputs over a 140-dB range ( 3»V to 30 V), with capabil- normally required.
ities to detect and display any 80-dB portion. The accuracy and stability of the generator, plus the
The analyzer conveniently provides two frequency dis- resolution of the recorder and the variety of chart speeds
play modes. A logarithmic display with 2.0-, 2.5- and 5.0- and averaging-time programs, permit precise response
inch decades and a linear display with scale factors of 50
measurements of almost any device — performed with
Hz, 500 Hz, and 5kHz per inch. The logarithmic mode is the ease and economy of a single instrument rather than
useful when you use the tracking analyzer feature for net- with the clutter and confusion of two.
work measurements or plotting spectra over a wide fre-
quency range. The linear mode is most useful for obtain- Versatile level recording for
ing a detailed analysis over a narrower frequency range. • reverberation-time measurements
For maximum versatility, this analyzer has controls for • general level- recording applications
setting both the analysis start and stop points anywhere
within the 10 Hz-to-80 kHz range. Select the 1523-P1 Preamplifier plug-in for the best in
Through the use of the plug-in approach, GR offers you general recorder performance. The 1523-P1 gives you
two instruments — a graphic- level recorder and a wave a broad frequency coverage from 1 Hz to 500 kHz, a
analyzer — in a single coordinated package, capable of sensitivity of 100 pV, and 18 chart speeds from as slow
handling an entire measurement requirement from data as 20 hours per inch to as fast as half a second per inch.
acquisition and processing to hard-copy data recording. Continuously adjustable attenuation from 0 to 70 dB
Additionally, the plug-in approach accommodates present provides the utmost in recording resolution, and achoice
needs while allowing for economical future expansion of nine averaging times from 10 ms to 5sallows supreme
with other GR plug- ins. flexibility.

7g1
 Conveniences standard All plug- ins feature remote desired time. For reliability, a stepper motor drives the
programmability, a variety of inputs and outputs to syn- chart ( there are no gears or clutches to wear out, slip, or
chronize recorder operation with that of other instru- jam), and clog- free disposable pens eliminate messy re-
ments, and a choice of several potentiometers with dy- filling and provide clear, easily read, and skip-free traces
namic ranges from 10 dB to 100 dB to tailor the instru- even at the fastest writing speeds. You have a choice of
ment to your specific requirement. colors and a choice of marker types: the fastrak® Marker
for general purposes and the Slow-Speed Marker for par-
For convenience, a chart take-up reel is included but ticularly slow- moving records or those with much re-
the paper also can be fed out directly for immediate in- tracing over a part of the chart. GO/NO-GO limit adjust-
spection and use. For interpretation, an event marker ments are included to provide LO, GO, and Hl electrical
can be recorded by the simple push of a button at the outputs for external alarm or control applications.

SPECIFICATIONS

Recording: CHART SPEED: Automatically set by sweep time

with 1523-P1A (see below) and decade length. Decade length can be set for
Preamplifier Plug-in 2, 2.5, 3, 4, 5, or 10 in./decade.
Sweep Frequency: 1 Hz to 500 kHz; automatically from lower
for level-vs-time recordings to upper frequency. Lower frequency can be set to 1, 2, 5, 10,
20, 50, 100, or 200 Hz, 1, 10, or 100 kHz; upper frequency
•f
requencies to 500 kHz can be set to 10 or 100 Hz, 1, 2, 5, 10, 50, 100, 200, or 500
• up to 100-dB dynamic range kHz. ACCURACY: t--1.% of indicated frequency. STABILITY:
±0.05% over 10 min, ± - 0.25% over 24 h; after 30-min warm-
• 1- MI] input impedance
up. SWEEP TIME: 5 s to 200 ks/ decade in 5, 10, 20 se-
• 100. AV sensitivity quence; or manual sweep. Averaging time decreases with
frequency as follows: 2 s from 1 to 10 Hz, 200 ms from 10
• 18 chart speeds
to 100 Hz, 50 ms from 100 Hz to 100 kHz, and 20 ms from
Input: Chart 0- level can be 0 dB ( 100 µV) to 70 dB; set in 10- 100 to 500 kHz. SWEEP RESOLUTION: 3000 discrete log-
dB steps plus a continuous vernier. See Maximum Input arithmically scaled steps per decade ( 0.08% step). SWEEP
Sensitivity under 1523 Mainframe Specifications. MAXIMUM VOLTAGE: Dc output proportional to log of swept frequency
INPUT: 10 V pk ac to 250 kHz, nt 5 V pk ac to 500 kHz, re available at rear connector.
dc component of ± 350 V max. IMPEDANCE: 1Mn/30 pF at Sweep Amplitude: 500 0V to 5 V rms into open circuit behind
plug-in; 3.35 kit -± 1% direct to potentiometer via internal 600 ff. > 10 mW into 600 n, available at front BNC connector;
switch. CONNECTORS: Front and rear BNC and rear 3.pin A3 set in four decade ranges of 5 mV to 5 Vfull-scale open-circuit
mike connector that also provides power for 1560-P40 or - P42 plus continuous vernier; flat within 00.1 dB to 100 kHz, within
Preamplifier. =1 dB to 500 kHz. DISTORTION: < 0.2% from 1 Hz to 100
Input Frequency: 1 Hz to 500 kHz; flat within - ±0.1 dB to kHz with any linear load. HUM: < 0.03%. SPURIOUS ( dis.
100 kHz, within -
±2 dB to 500 kHz except 0-dB range down crete non-harmonic): - 55 dB. NOISE: > 60 dB below carrier
< 3dB at 100 kHz. Low-frequency and crest-factor cutoffs de- in 100-kHz bandwidth.
pend on averaging times ( see below).
with 1523-P3 Stepped 1/3-
Recording: CHART SPEED: 0.5 s/ in. to 20 h/ in., in 18 ranges
of 0.5, 1, 2, 5, 10, and 20 h, min, or s/ in., plus fast scan of Octave-Band Analyzer Plug-in
2 in. / sand slow scan of 2 in./min.; all synchronized to line
contains 1/3- octave- band filters 4
frequency. AVERAGING TIMES: 10 ms to 5s in 9 ranges, all
for spectrum recording
remotely programmable. Sinusoidal low-frequency cutoff (< 1
dB down) and fundamental cutoff for 20-dB crest factor de- • 1Hz to 80 kHz
pend on averaging times as follows: • Fifty 1/ 3-octave bands
Weighting: A. conforms to ANSI S1.4-1971 Type 1. • 100-,A/ sensitivity
Low•Frequency Cutoff Low- Frequency Cutoff • 1-
Mn input impedance
Av e
Full Crest Avg Full Crest
Time Sinusoidal Factor Time Sinusoidal Factor • averaging times varied auto
10 ms 400 Hz 1kHz 500 ms 2Hz 8 Hz matically for fast constant-
20 Ins 100 Hz 500 Hz Is 1Hz 3.5 Hz conf idence results
50 ms 20 Hz 120 Hz 2s <1 Hz 1.6 Hz
100 ms 10 Hz 35 Hz 5s < 1 HZ 1Hz
200 ins 5 Hz 16 Hz Input: Chart 0- level can be 0 dB ( 100 µV) to 70 dB; set in
10-dB steps plus a continuous vernier. MAXIMUM INPUT:
with 1523-P2 -tr 10 V pk ac to 100 kHz, re dc component of 90 V max.
Sweep Oscillator Plug-in IMPEDANCE: 1Mn/j30 pF at plug-in, 3.35 ka 1% direct to
e potentiometer via internal switch. CONNECTORS: Front and
contains sweep generator for rear BNC and rear 3- pin A3 mike connector that also provides
level-vs- frequency recordings power for 1560-P40 or - P42 Preamplifier.
Input Frequency: 1Hz to 80 kHz center frequencies; fifty 1 / 3-
• freauencies to 500 kHz octave noise-bandwidth ( bands 0 thru 49) 4- pole Butterworth
filters with Class If ( moderate attenuation) Type 0 characteris-
• 100,4V sensitivity
tics that conform to ANSI S1.11-1966 for sound- level record-
• up to 100-dB dynamic range ings, switch-selected flat-response (all pass) channel at start
of chart displays over-all level. Also available, with filter char-
• 1-
Mn input impedance acteristics conforming to IEC 225, as Type 1523-P31. CEN-
TER- FREQUENCY ACCURACY: t .3%.
- LEVEL UNIFORMITY:
, 1dB at center frequencies. PASSBAND RIPPLE: 1dB pk-pk
Input: Chart 0-level can be 0 dB ( 100 AV) to 70 dB; set in 10-
dB steps. See also Maximum Input Sensitivity under 1523 max. NOISE: Equivalent input noise below 0 dB ( re 100 AO
Mainframe Specifications. MAXIMUM INPUT: ± 10 V pk ac to for all pass, and bands 0-39, increasing 1 dB/band to below
500 kHz, re dc component of -± 40 V max. IMPEDANCE: 10 dB in band 49. HARMONIC DISTORTION: < 0.1% for sig-
1M51 , 30 pF at plug-in; 3.35 kif ,t 1% direct to potentiometer nals of 1V rms at band centers. STOP- BAND ATTENUATION:
> 60 d8 for frequencies of > 8 and < 1/8 times center fre-
via internal switch. CONNECTORS: Front and rear BNC and
rear 3- pin A3 mike connector that also provides power for quency. OUTPUT: Filtered input signal available at rear con-
1560-P40 or - P42 Preamplifier. nector. PEAK MONITOR included with panel lamp to indicate
overload.
Input Frequency: 1 Hz to 500 kHz; flat within •±-0.1 dB to 100 Input-Frequency Scan: Adjustable; automatically steps from
kHz, within r-2 dB to 500 kHz, except on 0-d8 range, down lower frequency to ( but not thru) upper frequency. Lower
< 3 dB at 100 kHz. Averaging times programmed automati- center frequency can be set to 1, 2, 10, 20, 100. or 1000 Hz;
cally to avoid low-frequency cutoff; program can be inhibited upper center frequency can be set to 0.1, 1, 2, 10, 20, or
by external input. 100 kHz.

011.1.4.er•
 specifications (cont'en

Analysis Programs: NO. OF PROGRAMS: 11, designated A, B, 0.5. 0.1 Hz/ sand ( b) 5, 10, 50, 100, 500 s/ decade. Chart-
C . . L. VARIETY: Program A provides constant averaging scale factors are 2.0, 2.5, and 5.0 in./decade. Sweep rate,
time per band; others, combinations of the same with aseries automatically controlled. Averaging time and pen lift, also
of steps of constant confidence (averaging time - propor- automatically controlled unless overridden by remote contrm
tional to one over analysis bandwidth). CONFIDENCE: For MANUAL: Frequency may be manually set or swept with START
random noise, for 90% confidence, averaging time is adequate frequency control and range multiplier. LEVEL vs TIME: An-
for ± 0.5-dB accuracy along the "staircase" labeled + 0.5 dB. alyzer may be set manually to any frequency or all pass and
See chart. Similarly, 1.5 dB, for the staircase so labeled. level recorded vs time.
Any part of a program above a staircase has greater confi- Secondary Inputs: Remote control by switch or DTL/TTL
dence. No program operates below the "sine - staircase on the ground closures of numerous factors such as pen lift, chart
left; thus averaging time is always sufficient to keep the de- scan, run, stop, averaging time, event markers, etc.
tector operating above its low- frequency cutoff. AVERAGING Tracking Output: FREQUENCY: Follows analyzer center f re-
TIME: 5 sfor 1-, 1.25-, and 1.6- Hz bands; 5, 2, 1 . . 0.05 s quency. LEVEL: 1 V across 600 n ( or less, by manual con-
for higher-freq bands, depending on selected program. Exam- trol). DISTORTION: < 0.2%.
ples: (See chart). Program B uses 5-s avg'g time below 2 Hz,
Other Outputs: I- F: Filtered signal at 50 kHz, amplitude propor-
2-s for 2 Hz and above. Program H uses 5-s avg'g time up to
tional to analyzer output. RETRANSMITTING: DC voltage pro-
8 Hz, constant confidence (-", 1.5 dB) from 8 through 100 Hz,
portional to pen excursion. SWEEP: DC voltage proportional
and 0.5-s avg'g time from 100 Hz through 100 kHz. Location
to analyzer frequency. DIGITAL: Signals indicating system
of program letter on chart indicates shortest avg'g time for the
status, such as pen lift, chart direction and rate, run, stop,
program. ( Imagine that each arrow, like K, extends to right.)
averaging time, limits. etc.
SEQUENCE: Analyzer steps to a band, dwells for stabilization,
then records the level, at 2 in./ s chart speed. DWELL PE-
RIOD: 6 x "averaging time," except at your selected Start MAINFRAME SPECIFICATIONS
Frequency band of each analysis ( and "all- pass") 18 x avg'g Dyremie Ranee: To 100 dB depending on potentiometer. 5 available, all
time. easily interchanged and all with 5,0 wales except for 60 dB which has
90% CONsiDENCE
12-cm scale 10 dB ( with 013 1.d8 hnearityl. 25 dB 100 15 dB). 50 dB
1.0.25 dB. recommended for normal usel. 60 dB 100 3 dB, for use with

5 AVERAGING - TIME
1523.P3 onlYI. and 100 dB Ito heel Maininurn Input SensitIvIty 100 PV

Iii
2 6 PROGRAMS rms for averaging times 0 1 sec or greeter. I rnV rin, for averaging times
GR 1523- P3 < 0.1 sec. except for 10 dB pot max sensitivIty 1rn1/. minimum e.reging
C time 50 meen DEAD BAND. 0 15% of full scale except t 0.25% on
0.01. 0.02 and 005 sec averaging time. DETECTION True ems, error
D H 01 dB for 15.013 crest factor. < 0 5 dB for full 20-dB erect factor for
frequencies above crest- factor cutoff frequency. NOISE , Equivalent Input
02 E
Noise ieee than 40 pV rms. RETRANSMITTING POTENTIOMETER.
?, 0I F Provide, dc output voltage, PrOpOrtional to ac input, of 0 to 10 4 V dc
12 V/in of pen deflection).
1
' 005
I 2 5 10 20 50 100 200 500 ix 2k 55 nai 20k 50k 1005 Pen Control: Pushbutton switche, or eaternal DTL IT L signals control pen
BAND CENTER TREOuENCy IN nERTZ POution lup, down, or automatocally raned or lowered) Pen status also
inchmted by DTL outputs

au. Control: Pushbutton switches or external DT L - 7TE ground closures

With 1523-P4
start or stop recording. reset paper to start ol sane chart or advance ot to
Wave Analyzer Plug-in start of next chart. and provide fast forward or rekerse Switch settings also
inchmted by DTL output, CHART SPEED (also see Chart Speed und.
for detailed noise Individual ptug-on headongsl Can be externally programmed except on
and vibration analysis 1523.P3. MOTOR StePPer motor moves PeTer 0 0067 'm Increm en t
at rates . ais to 300 pus 12 In ./s) Pulses suPfalted be internal clock o, be en -
lerml DIE • TT L input of A300 pus Pulses also available as an output to
• 80-dB dynamic range
synchtonce other recorders PHOTOCELL DTL ground closure OutPut
• 10- Hz and 100- Hz bandwidth corresponds to black marks printed on paper

• Tracking output Levels and Event Markers: LIMITS 3 DTL outputs provide ht. go. and lo
results of 2 achustable limits vs the recording level for go/no-go using
• 10 Hz-to-80 kHz analysis rang- EVE NT MAREE RS Two, pushbutton watches control one pen to mark
selected events on paper. external DTL.TTL sum' achmt. either or both
• Linear and log frequency
pens.
displays Menem: All plug-in pushbuttoneontrol lunctIons can be remotely Inch.-
led or controlled, other controls cannot be except for Chart Speed and
Averaging Tome controls on PI and - P2 and Sweep Time Per Decade on
Main Input: Chart 0- level can be - 30 dB ( 3.16 erV) to + 50 dB; P2 Levels are standard DIE or TSE ( closure to ground or 00.5 V, • 3.5 to
set in 10-dB steps supplemented by a 12-dB continuous ver- 05 VI and are available at 2 rear double 19- pin etched-board terminals
nier. TOTAL MEASUREMENT RANGE: 140 dB ( 3.16 •V to &Weed: 3 It BNC terminated patch cord. 3 rolls of chart paper, fastrak®
31.6 V). DYNAMIC RANGE: 80 dB, without changing attenu- Marker Set 14 red. 4 green. 4 blue pens), E.nt Marker Set of 4 red and 4
black pens, 3 potentiometer contacts. 2 paper cap assemblies, 50 chart
ators, may be displayed with the 100-dB potentiometer. MAX-
mounting sheets. 2 double 19-tAn etched-board connectors for external
IMUM INPUT: 30 V rms to 80 kHz, re dc component of -± 50 V programming inputs and outputs. Power cord.
max. INPUT OVERLOAD: Detected by peak monitor and indi-
Pewee: 100 to 125 or 200 to 250 V. 50 lo 60 Hz. 90 W typical. 160 W
cated on front panel. IMPEDANCE: IMO/ / 30 pF. CONNEC- Mee.
TORS: Front and rear BNC and rear 3-pin A3 mike connector PA lee Pencil or rack models. DIMENSIONS lwxhxd) Bench
that also provides power for 1560-P40 or -P42 Preamplifier. 19.56X 8.18X 19.56 M. 149602070496 nun). rack, 1907018 44 M.
ACCURACY OF RECORDED LEVEL: 0.5 dB below 10 kHz, 1 1483%178%470 fawn). WEIGHT Bench, includIng plug-in. 63 lb ( 29 kg1
dB above ( with 50-dB pot). net, 98 lb ( 45 kg) shipping. rack, including plug n, 57 lb 126 kg) net, 92 lb
(42 kg) shipping. plug-in when shipped separately. 8 lb 13.7 kg) net. 16 lb
Analyzer Characteristics: FREQUENCY: 10 Hz to 80 kHz. 18 kg, shipping.
SWEEP RANGE: 10 Hz to 10 kHz ( 10- Hz bandwidth); start and Ceram
stop frequencies each set by 4- position switch and 10-turn
dial. BANDWIDTH: 10 Hz, 100 Hz, and "all pass," switch 1523 Graphic Level Recorder, vatt. 50 dB potentiometer but
without plug on
selected. PASS- BAND SHAPES: 3-dB bandwidth 10 -± 1Hz, MAINFRAME. BENCH MODEL 1523-9700
60-dB bandwidth - 60 Hz, 80-dB bandwidth 120 Hz; 3-dB MAINFRAME, RACK MODEL 1523-9701
bandwidth - 100 -, 10 Hz, 60-dB bandwidth 400 Hz, 80-dB REQUIRES ONE OF THE FOLLOWING POTENT1OME
TE RS, potentionetes are easily interchanged so that more
bandwidth " 800 Hz. ACCURACY OF RECORDED FRE-
than One WS. Can be ordered to suit chi ferIng applications
QUENCY: -± 2% or 5 Hz, whichever is greater. STABILITY: 10-dB Potentiometer 1523-9620
(After 30- min warmup) --, 0.05% ( 10 min), 0.25% ( 24 h). 25-dB Potentiometer 15219621
50413 Potentiometer. normally supplied 1523-9622
DISTORTION: Internally generated distortion and hum prod- 60.dI3 Potmhometer. used only voth 1523-P3 1523-9623
ucts 75 dB below full scale. NOISE: < 1 •V in a 10- Hz band- 100.dB Potentoometer 1523-9624
width. IMAGE REJECTION: 80 dB ( 100 to 180 kHz). I-F REQUIRES ONE OF THE FOLLOWING PLUG INS, ohm ics
REJECTION: ?- 80 dB ( 50 kHz). are easily interchanged so that more than one type can be or
dared to suit differing applicatoons
Analysis Modes: AUTOMATIC: Linear sweep between start and Preamplifier Plug-in, 1523-P1A 1523-9601
stop frequency settings with automatic decade ranging. Anal- Sweep Oscillator Plug in. 1523 P2 1523-9602
Stepped 1/3- Octave- Band Analymr Plug-in,
ysis rates are 100, 50, 10, 5, 1, 0.5 and 0.1 Hz/ s. Chart- Conforms to ANSI S1.11-1966. 1523.P3 1623 9603
scale factors of 50, 500 or 5000 Hz! in. Logarithmic sweep Stepped 1/3-Octane- Band Amlyaer Plug In,
over single or multiple decades with automatic decade ranging conforms to IEC 225, 1523-P31 . 1523 9605
provides the following types of analysis: ( a) 100, 50, 10, 5, 1, Wave Analyzer Plug- In. 1523.P4 1523 9604

293
Type 1568-A WAVE ANALYZER

• 20 Hz to 20 kHz
• 1% constant-percentage bandwidth
• portable, battery-operated
• 85-db rejection

The 1568-A is an important new instrument for high. HIGH RESOLUTION

resolution frequency analyses, whether for measuring
Narrow bandwidth permits separation of closely spaced
vibration and noise components or the spectrum of a
frequencies; wide dynamic range, high stop-band attenua-
complex electrical signal. Recent design advances com- tion, and low distortion allow measurement of small com-
bine the excellent filter shape of a wave analyzer with
ponents in the presence of components up to 80 dB
the convenient, simple operation of constant- percentage-
larger. These capabilities are vital to the identification
bandwidth analyzers in a portable, low-cost instrument.
of unwanted vibration and noise components and to the
The voltage sensitivity and input impedance, adequate measuring of discrete frequencies in complex electrical
for most uses, can be improved to 10 microvolts full-scale
waveforms. At low frequencies bandwidth is narrower,
and > 500 megohms, respectively, by the use of a 1560-
stability better, and calibration more accurate than those
P40 Preamplifier. Power for it is supplied at the input
of fixed- bandwidth heterodyne wave analyzers.
connector.
The 1568 will excell in such applications as
• harmonic distortion measurements at low frequencies
• harmonic analysis — 1% bw yields 50 components
• detailed analysis of machinery noise and vibration
• separation of close, discrete, low frequencies

AUTOMATIC ANALYSIS
In combination, the 1568-A and 1521-B Graphic Level
-2.0 - to 0 . 10 - 2.0 Recorder produce spectrum plots with as much as a 70-dB
a f (PER CENT) recording range. Automatic range switching is included
for ease and speed in making spectrum analyses. The
analyzer and recorder are available mounted in a cabinet,
05 10 2.0 3.0 10.0
interconnected, and mechanically coupled as a complete
NORMALIZED FREQUENCY
system, the 1913 Recording Wave Analyzer.
Attenuation cherecteristics of the filter.

294
 specifications
»I. 0 FILTER SECTION
CS V
RS V:142, ri;r1;:relrfC
C'Tirtnel'
;ER CONTR..
IS VOLTS DC

10W 0FILTER SECTION

NTTENIMTO. FILTER ttttt VAT OH

—, « TEN RANGE 001013 rejellEN AMPI. flip PIETER RANGE MSS
10-db STEPS K.L0 ONE N)- dl STEP
TO 30db
e

ceetole

OUTPUT
NYPLIIIIER MAcFIR
Nod AMPLIFIE

Graphic Level Recorder is used. Medium- speed motor is recom-

FREQUENCY
mended. Chart pflPer is Catalog No. 1521-9475. Frequency scale
Range: 20 Hz to 20 kHz In six half-decade bands.
is logarithmic, 10 inches per decade, vertical scale is 4 inches
Dial Calibration: Logarithmic. for 20, 40, or 80 dB, depending on the potentiometer used in the
Accuracy of Frequency Calibration: 1%. recorder.
Filter Cheeneteristice: Bandwidth between 3-dB points on selec- Amplitude Calib IA built-in, feedback- type calibration system
tivity curve is one percent of selected frequency. permits amplitude calibration at any frequency.
Attenuation at 20% above and at 20% below selected fre- Acc eeeee ies Supplied: Power cord; 1568-2090 Detentad Knob and
quency is greater than 50 dB referred to the level at the selected Dial Assembly, used to facilitate measuring the components of
frequency. Attenuation at twice and at one-half the selected fre- an input signal as a percentage or in decibels with an arbitrary
quency is at least 75 dB referred to the level at the selected voltage reference.
frequency. Ultimate attenuation is greater than 85 dB. Acc sssss ies Available: Preamplifier and Adaptor Set 1560-P40.1,
Uniformity of filter peak response with tuning is ,1,1 dB from Link Unit 1521.P15, with Sprocket Kit 1521-P16 for mechanical
20 Hz to 6.3 kHz and 1-.-2 dB from 20 Hz to 20 kHz. coupling to 1521-B Graphic Level Recorder equipped with Drive
Unit 1521-P1013; Chart Paper 1521-9475.
INPUT Power bubbly: 100 to 125 or 200 to 250 V, 50 to 60 Hz. 2 W for
Impedance: 100 kn. normal operation, 3.5 W for batter/ charging. A rechargeable
nickel-cadmium battery is supplied. Battery provides about 20
Voltage Range: 100 µV to 300 V, full scale, in 3-10 series steps.
hours of operation when fully charged and requires 16 hours for
Power is supplied at input socket for the 1560-P40 Preamplifier,
which extends the sensitivity to 10 µV, full scale, and increases charging. Internal charger operates from the power line.
the input impedance to more than 500 MO. Mounting: Flip-Tilt case, rack model available.
Distortion: Input-circuit distortion is lower than — 80 dB relative DiMOTHEi0OS ( width x height x depth): Portable, with case closed,
to input- signal level. 1344 o 13 A 81 /
4 in. ( 340 x 330 x 210 min); reek, 19 x 12V4 x 5
In. ( 485 x 310 x 130 mm).
OUTPUT Weight: Net, 21Vz lb ( 10.0 kg); shipping, 27 lb ( 12.5 kg).
impedance: 6000 0. Any load can be connected.
Catalog
Voltage: At least one volt open circuit when meter reads full scale. liee Description

Crest- Factor Capacity: Greater than 13 dB. 1511114 Ware Analyzer

Output Meter: In addition to normal-speed mode, meter has slow. 1568.9701 Portable Model, 115 V at
speed mode for manual measurements of noise. 1568-9702 Portable Model, 230 V ac
1568.9820 Rack Model, 115 V at
1568.9821 Rack Model, 230 V ac
GENERAL 1580-P401 Preamplifier and Adaptor Set
1560.9510
Analyzing ttttt : 80 dB. Components of an input signal that differ 1521.9615 1521-PIS Link Unit
in amplitude by as much as 80 dB can be measured. 1521-9616 1521-Pie Sprocket Kit
Automatic Repording: Automatic range switching is provided to 1521-9475 Chart eeeee
allow convenient, continuous spectrum plotting when the 1521 8410-0410 Replacement Battery

Frequency spectrum an lysl of a 1.04n pulse at a 7041z repetition rate. The 1% bandwidth
yields high resol lion at low frequencies, shows the invokes It high frequencies

• federal Stock Number 6625.974-3558

295
 Type 1900-A WAVE ANALYZER

• 20 to 54,000 Hz, linear frequency scale

• 3-, 10-, and 50- Hz bandwidths
• 30 IA/ to 300 V, full scale — 30/ with preamp
• 80-dB recording analyzer with 1521 recorder
• outputs: filtered or BFO, 100 kHz and dc recorder
• 1-megohm input impedance on all ranges

eC>.

• a
- ,

is

specifications
FREQUENCY

Range: 20 to 54,000 Hz. The frequency is indicated on a counter Recorder connected, full-scale output is at least 3 V. Dynamic
and a dial with a linear graduation, 10 Hz per division. range from overload point to internal noise is > 80 dB with atten-
uator knob fully clockwise.
Accuracy of calibration: ±(
- Vz% -I- 5 Hz) up to 50 kHz: ±-1%
-
beyond 50 kHz. Recording Analyzer: See the 1910-8 Recording Analyzer and
152143 Graphic Level Recorder.
Incremental- Frequency Dial ( SF): ±- 100 Hz. Accuracy is ±-2 Hz
below 2 kHz, -_
L-5 Hz up to 54 kHz. DC Output: 1 rnA in 1500 0. full scale, one side grounded.
Automatic Frequency Control: At frequencies below 10 kHz, total Filtered Input Component: Output at least 1 V across 600-0 load
for full-scale meter deflection with output control at max.
range of frequency lock is 400 Hz for the 50- Hz band and 150 Hz
for the 10- Hz band, as defined by 3- dB drop in response from Tracking Analyzer ( Indicated Frequency): 20 Hz to 54 kHz, output
full-sca(e deflection. At 50 kHz, the lock ranges decrease to one- is at least 2 V across 600.0 load with output control at man.
half of these values.
GENERAL
SELECTIVITY Three bandwidths ( 3. 10, and 50 Hz). Terminals: Input, binding posts; output, telephone jacks.
Effective bandwidth for noise equal to nominal bandwidth Power Required: 105 to 125 or 210 to 250 V, 50 to 400 Hz, 40 W.
within ±-10% for 10- and 50- Hz bands and ±
--20% for 3- Hz band.
Accessories Supplied: 1560-P95 Adaptor Cable. phone plug, power
3- Hertz Band: At least 30 dB down at -
±-6 Hz from center fre- COrd.
quency, at least 60 dB down at ± 15 Hz, at least 80 dB down Accessories Available: 1900- PI and 1900-P3 Link Units for cou-
at T -25 Hz and beyond.
pling to 1521 Graphic Level Recorder, 1560-P4OH Preamplifier and
10- Hertz Band: At least 30 dB down at ± 20 Hz. at least 60 dB Power Supply Set.
down at ± -45 Hz, at least 80 dB down at ± 80 Hz and beyond. Mounting: Rack- Bench Cabinet.
50- Hertz Band: At least 30 dB down at _-•_: 100 Hz, at least 60 dB
down at ± --250 Hz, at least 80 dB down at - ± 500 Hz and beyond. Dimensions
1544 in. ( 485
( width
x 415 o
x height
390 mm),
x depth):
rack model,
Bench 19
model,
x 1544
19xx1316
/ 4 in.
/
4 1
1 A
INPUT (485 x 400 x 340 mm).

Impedance: I MU shunted by 30 pF on all ranges. Weight: Net, 56 lb ( 26 kg); shipping. 140 lb ( 64 kg).

Voltage Range: 30 AiV to 300 V. full scale, to 3 AV with dreamt,.

in 3, 10 series. A decibel scale is also provided.
Voltage Accuracy: After calibration by internal source, the accu-
Catalog
racy up to 50 kHz is , ( 3% of indicated value + 2% of full scale) Number Description
except for the effects of internal noise when the attenuator knob
is in the maximum.sensitivity position. From 50 to 54 kHz, the 1900-A Ware Analyzer
above 3% error becomes 6%.
1900-9801 Bench Model, 115 V
Residual Modulation Products and Hein: At least 75 dB down. 1900-9811 Rack Model, 115 V
1900.9802 Bench Model. 230 V
OUTPUT 1900-9812 Rack Model, 230 V
100-kHz Output: Amplitude is proportional to amplitude of selected
component in analyzer input signal. With the 1521 Graphic Level
Patent Numbers D 187.740 and 2.581.133.

296
Type 1564-A SOUND AND VIBRATION ANALYZER

• 2.5 Hz to 25 kHz
• 2bandwidths: 1/3- and 1/10-octave
• use direct from microphone or vibration pickup
• ac or portable battery operation
II automatic spectrum plots with 1521 recorder

specifications
Power Required: Operates from 105 to 125 or 210 to 230 V. 50-60
Hz, or from nickel-cadmium battery supplied. Battery provides
FREQUENCY 25 h of operation when fully charged and requires 14 h for
2.5 Hz to 25 kHz in four decade ranges.
charging.
Dial Calibration: Logarithmic.
Ace sssss ies Supplied: Power cord.
Accuracy of calibration: ..
L.2% of frequency-dial setting.
Accessories Available: 1560.P6 Microphone Assembly, 1560-P52.
Filter Characteristics: Noise bandwidth is either Va octave or
.P53, - P54 Vibration Pickups, 1560-P40 and -P42 Preamplifiers
1/10 octave. One.third-octave characteristic has at least 30-dB
attenuation et one-half and twice the selected frequency. One- (power for preamp available at input connector). 1564-P1 Dial Drive
for coupling to 1521 recorder for stepped third.octave analysis.
tenth-octave characteristic has at least 40-dB attenuation at
one-half and twice the selected frequency. Ultimate attenuation Mounting: Flip-Tilt Case. Rack-mount version also available.
is 70 dB or greater for both characteristics. For both band.
widths, peak response is uniform t.-1 dB from 5 Hz to 10 kHz Dimensions (width o height x depth): Portable model, 104. x 81
/
2
o 8 in. ( 260 0 210 o 205 mm): rack model, 19 o 10 1 /
2 0 6 in.
and -±-1.5 dB from 2.5 Hz to 25 kHz. An ALL- PASS, or flat.
characteristic is also included. (485 x 270 x 155 mm).
Net Weight: Portable model, 14:/z lb ( 7 kg); rack model, 151
/
2 lb
INPUT
Impedance: 25 MO in parallel with 80 pF ( independent of ' Dense.
(7 5kll. Weight: Portable model, 17 lb ( 8 kg): rack model, 28 lb
ator setting). Shipping
(13 kg).
Voltage Range: 0.3 mV to 30 V full scale in 10-dB steps.
Microphone: 1560-P6 Microphone Assembly or the 1560.P4OK Pre-
amplifier and Microphone Set is recommended.

OUTPUT
Voltage: At least 1.0 V open circuit, when meter reads full scale.
Catalog Description
Impedance: 6000 0. Any load can be connected.
Number
Meter: Three scales, 0 to 3 V: 0 to 10 V: — 6 to -,- 10 dB.
15044 sssss end Vibration Analyser
Recording Analyzer: Automatic range switching at the end of each
frequency decade allows convenient continuous recording of Portable model. 115 V
1564-9701
spectra with the 1521-B Graphic Level Recorder. Rack Model, 115 V
1564-9820 Portable Model, 230 V
1564-9702
GENERAL Rack Model, 230 V
1564-9821
Amplitude Calibration: Built-in, feedback.type calibration system 8410-0410 Replacement Batter/
permits amplitude calibration at any frequency.
Detector: Rms with three averaging times. Faster two speeds con- Patent Numbers 3,012,197; D 187,740: and 2,966,257.
form with ANS standard for sound- level meters.

297
 Type 1521-B GRAPHIC LEVEL RECORDER

• 7 Hz to 200 kHz
• 1- mV ac sensitivity

• linear dB plot of rms ac-voltage level

• 20-, 40-, or 80-dB range

• convenient, disposable pens

•-=4•,1111

The 1521-B is a completely transistorized, single- The writing speed is sufficiently high for the measurement
channel, servo-type recorder. It produces a permanent, of reverberation time and other transient phenomena.
reproducible strip-chart record of ac-voltage level as a The wide range of paper speed facilitates long- period
function of time or some other quantity. studies of the noise produced by traffic and machinery, as
Most often these are records of the frequency response well as of short- duration transients.
of a device or of the frequency spectrum of noise or of a The frequency response can be extended downward to
complex electrical signal. The Graphic Level Recorder 4.5 Hz at the slower writing speeds. Writing speeds and
can be mechanically or electrically coupled to various GR low-frequency cutoff are selected by a single switch.
analyzers and oscillators to synchronize the frequency Changes of range are easily accomplished by use of a
scale of the chart record with the instrument's calibrated 20- dB or an 80-dB potentiometer in place of the standard
tuning-control dial. Such combinations of instruments 40-dB unit. With the 80-dB unit, the maximum writing
are available factory assembled or as individual compo- speed is 300 dB/second. The slow writing speeds filter
nents to add to existing equipment. out abrupt level variations, yielding a smoothed plot with-
Owing to the high stability of its reference voltage and out loss of accuracy.
amplifier gain, the 1521 can be calibrated and used as a A linear potentiometer is available, which can be used
recorder of absolute level. for dc recording and is easily substituted for the log-
With a sound- level meter, the recorder can plot sound arithmic ac potentiometers.
levels over a wide dynamic range as a function of time.
spec if ications
Recording Range: As supplied, 40 dB full-scale: 20,10 and 80-dB External Cie Reference: An external dc reference voltage of from
ranges are also available. For dc recording, 0.8 to 1 V ( 0.8 to 0.5 to 1.5 V can be applied internally to correct for variations
1.0 mA) full-scale, with zero position adjustable over full scale. of up to 3 to 1 in the signal source of the system under test.
Frequency Response and Writing Speed Detector Response: Rms within 0.25 dB for multiple sine waves.
Level Recording: High-frequency response -.1-2 dB to 200 kHz. square waves, or noise. Detector operating level is 1V.
Low-frequency sine- wave response depends on writing speed, as Chart eeeee 4- inch recording width on 5- inch paper. All rolls are
shown in following table: 100 feet long. See full list of charts below.
Writing Speed (aperos) Low- Frequency Cutoff Hz Accessories Supplied: 40-413 potentiometer. 12 disposable pens
with 0.1- Inch overshoot ( less than 1dB down) with assorted ink colors. 1 roll of 1521-9428 chart paper, power
cord. 1560-P95 Adaptor Cable ( phone to double plug).
20 100 Accessories Available: Potentiometers, chart paper, pens, high-,
10 20 medium-, and low- speed motors, drive and link units.
3 7 ( 3 dB down at 4.5 Hz) Power Required: 105 to 125 or 210 to 250 V. 50 or 60 Hz, 35 W.
1 7 ( 3 dB down at 4.5 Hz)
Mounting: Rack- Bench Cabinet.
Oc Recording: 3 dB down at 8 Hz ( pk-pk amplitude less than 25%
of full scale).
%,"r4r5iIrlee t3'4g Prieincl
ht o dt
erre:ciel, e1
r n9cil
r
' i4fet
'
Potentiometer Linearity 225 o 290 elm)
20-, 40-, 80-013 Potentiometers: ± 1% of full-scale dB value plus a Weight: Net, 50 lb ( 23 kg): shipping, 62 lb ( 29 kg).
frequency error of 0.5 dB at 100 kHz and 1.5 dB at 200 kHz.
Linear Potentiometer: -± 1% of full scale. cato iog
Resolution :t-0.25% of full scale.

Max Input Voltage: 100 V ac. Graphic Level Recorder, 40-dB

potentiometer, high-speed motor
Input Attenuator: 60 dB in 10-dB steps.
1521-9802 1521-13 60- Hz Bench Model
Input Impedance: 10,000 I/ for ac level recording: 1000 LI for dc 1521-9812 1521-B 60- Hz Rack Model
recording. 1521-9506 1521-13O1 50- Hz Bench Model
Men Sensitivity: 1 rnV at 0 dB for level recording: 0.8 or 1 V 1521-9507 1521 - Bal 50- Hz Rack Model
full-scale for dc recording. Graphic Level Recorder, 40-dB
Paper Speeds potentiometer, medium- speed motor
High-speed meter ( normally supplied): Paper speeds of 2.5, 7.5. 1521-9833 1521-B 60- Hz Bench Model
1521-9834 1521-8 60- Hz Rack Model
25, 75 in./ min. Used for high-speed-transient measurements and
with Type 1304 Beat- Frequency Audio Generator. Graphic Level Recorder, 4008
potentiometer, low-speed motor
Medium-speed motor ( supplied on r eeeee t): Paper ' Feeds of 0.5.
1521-9817 1521-B 60- Hz Bench Model
1.5. 5. 15 in./min. Used with analyzers and in level- vs-time plots: 1521-9818 152143 60- Hz Rack Model
must be used with 1564-P1 Dial Drive.
Other potentiometers and combinations for both 60.Hz and
Low-speed motor ( supplied on rectum,* Paper speeds of 2.5, 7.5,
50- Hz models available: Inquiries Incited.
25, 75 in./h. Used for level- vs-time measurements from 1 to 24 h.
Patent Number 2,581.133

298
 FREQUENCY- RESPONSE 81 SPECTRUM RECORDER ASSEMBLIES

Several GR instruments can be used with the 1521-8 be ordered under a single catalog number and include all
Graphic Level Recorder for automatic plotting of the fre- accessories normally needed. Or the component items
quency response of devices or the frequency spectrum of can be ordered individually to convert existing equipment
acoustic or electrical noise or of a complex electrical into fully automatic recording assemblies.
waveform. Automatic plotting with these instruments re- Custom assemblies of GR analysis equipment and sound
places tedious point-by-point manual methods and pro- and vibration instruments can be built to order to meet a
vides much more information in the form of finer-resolution variety of special requirements.
curves. Listed below are several such assemblies that can

RECORDING ANALYZERS
Catalog

1910-A RECORDING WAVE ANALYZER

The 1910-A is particularly useful in analyzing and re-

cording the frequency components present in mechanical
vibrations, acoustic signals, and in complex electrical sig-
nals including random noise. Its linear frequency scale,
20- Hz to 54-kHz range, three bandwidths (3, 10, and 50
Hz), and 80-d13 dynamic range permit higher-order, closely
spaced and weak components to be found with ease.
The complete assembly Includes the following:
1901 Wave Analyzer, including 15110- PIS Adaptor cable and
ot er accessories
1521 Be(opre-teisantliefmoret5er H
12
151100210r
a
p p
p dhi
cLevelaced
medium-speord er
matar
w i th
1521-P3 80-dB Potentiometer ( 1521-9603)
1521-P1013 Drive Unit ( 1521.9467) ( installed)
1900-P1 Link Unit ( 1900-9601) ( installed)
1900-P3 Link Unit ( 1900-96031
Chart ,,,,,, 10 rolls ( 1521-9464), scale 0-10 kHz
Chart Paper, 10 rolls ( 1521-9465), scale 0-50 kHz
Accessories Available: 1560-P40 and 1560-P42 Preamplifiers; choice
of vibration pickups or microphones.
Mounting: Rack- Bench Cabinets; Includes end frames for bench
use and supports for rack mounting. 1910-A Recording Ware Analyzer,
Dimensions (width x height 4 depth): Bench. 19 x 25 19 4 15V4 in. 60- Hz 115-V Model 1910-9701
(485 x 645 4 390 mm); rack, 19 x 24:9 x 1319 in. ( 485 4 625 x 60- Hz 230-V Model 1910-9711
340 mm). 50- Hz 115-V Model 1910-9493
Weight: Net, 116 lb ( 53 kg); shipping. 227 lb ( 104 kg). 50- Hz 230-V Model 1910.9494

1911-A
RECORDING SOUND AND VIBRATION ANALYZER
This assembly will generate continuous frequency plots
of the 1/3- or 1/10-octave spectrum of sound and vibra-
tion signals over the range of 4.5 Hz to 25 kHz. Thus
1/3-octave measurements can be made in accordance with
several common military and industrial noise-control speci-
fications. While the third-octave bandwidth is convenient
for testing compliance to a specification for maximum al-
lowable noise or vibration level, the 1/10-octave bandwidth
permits identification of individual frequency components,
leading to their reduction or elimination. The analyzer
will accept signals from a sound- level meter, vibration
meter, or bther stable amplifier, or directly from a micro-
phone or vibration pickup. It includes a storage drawer
and oower system control.
The 11111.9 consists of the fellewIng:
1584-A Sound and Vibration Analyzer, rack model
152141 ( or - B(21 for 50-114 supply) Graphic Level Recorder with
40-dB Potentiometer ( 1521-9602) and medium- speed motor
1521-P109 Drina Unit ( 1521-9467)
1521-P15 Link Unit ( 1521-9615), with 16- tooth sprocket in-
stalled ( standard 24-tooth sprocket al. included)
Chart Paper, 10 rolls ( 1521-9469). calibrated 2.5-25 normalized.
logarithmic
Adaptor Cable, double plug to offset phone plug.
Accessories Available: 1560.P40 and 1560-P42 Preamplifiers; 80- da
potentiometer; choice of vibration pickups.
1911-A Recording Sound and Vibration Analyzer,
Mounting: Completely assembled with end frames for bench use. 1911-9701
Hardware for rack mounting is supplied. 60- Hz 115-V Model
60- Hz 230-V Model 1911-9711
Dimensions ( width o height o depth): 1944 n 3144 0 1544 in. ( 500 50- Hz 115-V Model 1911-9493
o 8004 400 mm). 50-Hz 230-V Model 1911-9494
Weight: Net, 101 lb ( 47 kg); shipping, 158 lb ( 72 kg).

299
RECORDING ANALYZERS
(‘Ont'd)

Catalog
Number

1913
RECORDING WAVE ANALYZER — 1% Bandwidth
This constant- percentage- bandwidth recording analyzer
will make high-resolution spectrum plots from 20 Hz to
20 kHz. It is easy to use, having automatic range switch-
ing and few controls. Wide dynamic range and the 80-d8
potentiometer reduce the need to change sensitivity man-
ually to accommodate widely varying amplitudes. Narrow
bandwidth permits separation of closely spaced low fre-
quencies without forfeiting high-frequency resolution;
typically, the fiftieth harmonic can be identified. See
description of 1568-A Wave Analyzer for more details.
The 1913 is supplied assembled and includes a storage
drawer and system power control, which switches the
analyzer battery supply as well as the ac line.
The 1813 includes the following:
1588-A Wave Analyzer, rack model, and accessories
1521-B ( or - 801 for 50- Hz supply) Graphic Level Recorder with
40-dB Potentiometer ( 1521-9602) and medium-speed motor
1521-P3 80-dB Potentiometer ( 1521-96031
1521-PlOB Drive Unit ( 1521-9467)
1521-P15 Link Unit ( 1521-96151. with 16-tooth sprocket in-
stalled ( standard 24-tooth sprocket also included)
Chart Paner. 10 rolls ( 1521-94751, scale 2-20 log, normalized
Accessories Available: 1560-P40 and 1560-P42 Preamplifiers.
Mounting: Assembled in cabinet.
Dimensions ( width x height x depth): 192/
4 x 31 1
/
4 o 15-3
/
4 in. ( 500 1913 Recording Wave Analyzer
x 800 x 400 mm). 60- Hz 115-V Model 1913.9700
60- Hz 230-V Model 1913-9702
Weight: Net, 110 lb ( 50 kg), shipping. 165 lb ( 75 kg).
50- Hz 115-V Model 1913-9703
SO- Hz 230-V Model 1913-9701

300
T/D 1923 REAL-TIME TIME- SERIES ANALYZERS

• 0- to- 50 kHz frequency range

• dynamic range > 70 dB
• frequency resolution to 0.025%

• automatic analysis
• fully calibrated displays

tion — there is no need to know or to learn computer programming

Fast, accurate, and versatile — these analyzers represent the and no need for constant cable patching or control manipulation.
most capable and sophisticated signal- processing equipment The system design allows you to construct any desired compound
processing and input/output operations for automatic or repetitive
available anywhere — yet they can be operated by anovice. data- reduction routines. Parallel processing in both the processor and
controller permits wide-band performance. Complex, repetitive se-
quences can be initiated automatically or at the push of a button,
which eliminates the need for atrained operator to set up and super-
Tailored analysis The Time/Data 1923 analyzers provide a broad vise each measurement.
range ot time series-analysis techniques including such operations as
spectral analysis, correlation, filtering, and probability-density analy-
sis. The analyzers are offered in two basic models to suit widely
divergent needs, each model can be tailored to aspecific application
by a variety of options and additional equipment — all fully inte-
grated into aunified system.
One model combines the speed of a micro- programmed fast-
Fourier- transform processor with the flexibility of adigital control-
ler while the other model uses software for the operations. Both
tql C17111111 Iii
models include a full software package and, in addition, custom
programming is available for specialized needs.
Real-time analysis These analyzers permit real-time, continuous
processing — you process all the data all the time for maximum statis-
tical accuracy. Widest useful dynamic range is preserved by means of
16- bit words, double- precision calculations, and a unique dynamic
scaling operation that automatically adjusts the scale factor during
processing.
Easy analysis A task-oriented control console, meaningful displays,
and acomplete software package provide true " one- button" opera-

301
 specifications
T/D 1923A TM 1923C
Time-Series Analyzer FFT Analyze,
GENERAL
Data Word Size: 18 bits 18 bits
Date Frame Size Single channel 16 to 4096 words (
optional to 8192) 18 to 2048 word. (
optional to 8192)
Dual channel 16 to 2048 words ( optional to 4096) 18 to 1024 Words (
optional to 40961
Memory Size: Fourier Processor 4096 words ( 8192 max) contains no processor
Digital Controller 4096 words 128672 max) 8192 words 128672 max)
Calculation %swan: 1111.biL tixed.point arithmetic for Fourier Transform, 32 bits for Auto ( Power) Spectrum:
double-precision selectable for some operations
Dynamic Scaling: Block scaling is selectable during the Fourier Transform calculation. An option allows the
Digital Controller to adjust the scale factor to preserve the greatest useful dynamic range.

Continuous Proomeime Input data can be loaded continuously d all processing operations on adata frame are com-
pleted within the time required to load one frame, allowing gap-free, continuous processing.
PROCESSING OPERATIONS
Spectral Andyels Processing speed
and real-time bandwidth.
Fourier Transform 12 me for 1024 reel words, direct or inverse <1.0 .for 1024 real words, direct or inverse
Auto ( Power) Spectrum 12 m. for 1024 real words 1efor 1024 real words
38 kHz max for continuous processing 500 Hz max for continuous Processing
Cross Spectrum 17 nu for 512 real words 1.1 sfor 512 real words
14 kHz max for continuous processing 200 Hz max for continuous processing
Two-Channel Spectra Simultaneous Auto and Cross Spectra for Transfer- Function measurements.
Frequency Range 0.1 Fla to 50 kHz selected on 6decade ranges plus multipliers of 1, 2, 2.5.4. and 5. Iwo simultaneous channels.
Dynamic Range >70 dB. voltage or power >70 dB. voltage or power
Frequency Resolution, also see table below. 8 to 2048 lines ( optional to 4096) i 8 to 1024 lines ( optional to 4096)
Input Data Uniform or Henning weighting of input data
Averaging Mode Uniform or exponential weighting, normalized, selectable from 1to 2" averages or continuous averaging.

Resolution per Line in Hz Supplied: Operator training, installation, and acceptance testing at customer's
Input Frame Frequency Analysis Frequency Range in Hz • facility. STANDARD SYSTEM includes a digital controller IDEC PDP.11)
laze in words) Lines 10000 20,000 26,000 40.000 so,000 with core memory and hardware multiply/divide, dual analog- to-digital con-
yeller, control console, keyboard, digitaldo•onalog converter, oscilloscope
8192 . 4096 . 2.5 . 5. 6.25 . to. 12.5. display unit, and high-speed tape reader 11923A only). STANDARD SOFT-
4096 . 2048 . 5 • 10 . 12.5 • 20. 25 • WARE includes programs for all standard analysis operations, computer diag-
2048 1024 10 20 25 40 50 nostics and maintenance programs, and operating, programming, and main.
1024 512 20 40 50 80 100 tenance manuals.
512 256 40 80 100 160 200
Amildde Input scanners, aliasing filters. dc recorders. data printers. X-
Y
256 128 80 160 200 320 400 Plotters, teletypewriters, digital tape and custom software.
128 64 160 320 400 640 800
64 32 320 640 800 1280 1600 Payer: 110 to 125 or 220 to 250 V.50 or 60 Hz.
32 16 640 1280 1600 2560 3200 64.ehenial: WEIGHT 1923A, 1200 lb ( 544 kg) net, 1923C. 675 lb 1306 kg) net.
16 8 1280 2560 3200 5120 6403 Cataba
'May be option I. see *bow. • ' For lower analysis frequencies. divid Resole. Number Dewription
tien oar Lino by appropriate power of 10.
1923.9710 TbD 1923A TimeSaries Analyzer
ADDITIONAL OPERATIONS 1923.9712 1/0 1923C FFT Analyzer
Simultaneous AUTO and CROSS SPECTRA. TRANSFER FUNCTION. and Recommended Options - 07. - 19. -20
COHERENCE FUNCTION measurements. AUTO/CROSS CORRELATION. Standard Options Available
continuous CONVOLUTION. CEPSTRUM, and RECURSIVE FILTERING -01.8 Els Processor Memory 11923A only)
(optional). .02S 8k Controller Memory ( 1923A only)
Waveform Mending: 4 to 8192 points on waveform are averaged with -02-12 12k Controller Memory
uniform or exponential weighting, normalized, selectable from 1 to 2" .02-16 16k Controller Memory
averages or continuous averaging. AMPLITUDE HISTOGRAM. Selectable .02-20 20k Controller Memory
resolution to 1024 cells. SPECTRAL SMOOTHING, Block convolution of -02.24 24k Controller Memory
spectra with choice of smoothing function. OCTAVE ANALYSIS ( optional) -02.28 28k Controller Memory
Choice of 1/3 and 1/15 octave. ARITHMETIC OPERATIONS: Addition. -03 1566 Multichannel Preamplifier, with cable
subtraction, real and complex multiplication, complex conjugate multiplica- .04 1522 DC Recorder, with cable
MOO 2016 Printer
tion. binary scaling, automatic scaling, division, decimation, integration. CO-
ORDINATE CONVERSION Polar/rectilinear. log/exponential. -06 ASR.35 Teleprinter and Control Unit
.07 ASR.33 Teleprinter and Control Unit
OTHER CHARACTERISTICS .08 Anti-Aliasing Filter, per channel
.09 ROM Bootstrap Loads.,. for POP- 11
Analog Input: CHANNELS. 2. with simultaneous wrnpling. SENSITIVITY. -10 TU - 10 Digital Tape Unit with Controller. 45 ips.
2100 mV, ± 200 mV. ± 400 m1.1. ± 1.0 V. ± 2.0 V. calibrated. fullocale input 800 bpi.9.track IBM compatible
with 6.c1B vernier range. IMPEDANCE too-an input. COUPLING: an 15 .11 Digital Input, binary or BCD. per channel
Hz) or dc. DIGITAL RESOLUTION' 10 bits including sign. OVERLOAD -12 Analog Output Interface, per channel
indicated by warning lamp and alphanumeric record. TRIGGERING Internal .13 7034 X- Y Recorder, with dc preamps, null
or external with variable level, polarity, coupling, and delay. SAMPLING detector, and point plotter
RATE: Selectable from 0.1 Hz to 100 kHz or external. ALIASING FILTERS -14 Extra Cabinet Rack
(Option 08): One filter set per channel from 10 Hz to 20 kHz in 1, 2. 5 -15 RS- 11 Disk Memory, 256k words with RF•11
programmable steps. Controller, fised.head. 16 as/vvord transfer.
Digital WPM: From PAPER TAPE at 300 characters/s or from decimal 17 ms average access time
KEYBOARD. Additional RS- 11 Disks. 7 maximum
-16 RS.64 Disk Memory. 64 words with RC- 11
Didaley: 8X10 cm storage display oscilloscope with log or linear coordinates
on either axis and automatic calibration. Alphanumeric display of test param- Controller. fixed.head, 16 sis/word transfer.
eters. polar/rectilinear displays. and phase-plane displays. 16.1 ers average access time
Additional RS-64 Disks. 3 maximum
Control: CONSOLE Simplifies machine operation and allows awide range of
•17 TU.56 Single DEC tape Transport, with controller
operations to be performed by a single pushbutton. KEYBOARD. System -18 TU -56 Dual DEC tape Transport, with controller
flexibility is maximized through the use of a keyboard for entering, storing, .19 1762-0460 High-Speed Paper- Tape Reader
and linking software subroutines developed by T/D or the user for custom -20 63-M. Cabinet Rack with Shelf
application. Complex. repetitive sequences can be performed automatically.
PROVISION is also made to allow parallel processing in the Fourier processor
and digital controller. In addition to standard analysis software, custom
software is available to tailor the T/D 1923 to your exact needs.

302
Type 1921 REAL- TIME ANALYZER

• 3.15 Hz to 80 kHz
• 30 to 45 1/10-, 1/3-,
or 1/1-octave bands
• 70-dB dynamic range,
60 dB displayed
• 100-mV sensitivity
• calibrated attenuators
• digital detection provides
true rms answers
• 9known integration periods
• corrected spectrum displayed
directly

HmIIIIII

1111111111111111111111119/1/1e

Long aneed, now areality Since the inception of the first control and interpretation of the measurements with a
sound and vibration instruments, the need for rapid, on- minimum of operator attention. Even less work and faster
line measurements was apparent. But it took some thirty results are possible when auxiliary instrumentation is
years to fulfill the need, adequately and economically, and wedded to the 1921, instrumentation such as computer
the fulfillment came with the introduction of real-time control, scanner input, dc recorders, or data printers — all
analysis embodied in the 1921. of which interface with ease.
The 1921 is at least 30 times faster than its contemporary A unique solution The input signal is applied to a set of
serial analyzers, which means it can analyze more data analog filters (from 30 to 45 depending on your require-
faster and can measure non-stationary signals without tape ments) that cover a frequency range from 3.15 Hz to 80
loops. Its 70-dB dynamic range readily accepts the ran- kHz. These filters include, as an option, individual attenu-
dom or totally unspecified signals routinely encountered ators to permit pre-whitening or other signal conditioning;
in acoustic work. The answers are not only fast, they are they are housed in aunit that can be purchased separately
repeatable and statistically reliable because of the ana- if desired.
lyzer's -±0.5-dB accuracy and known integration times. The outputs of the filters are processed in another unit,
Optional attenuators provide ±- 0.25-d8 accuracy for dy- the rms detector. This detector is unique in that it proc-
namic-range extension, transducer flattening, subjective esses the signals from the filters digitally; i.e., the outputs
correlation, etc., and the filters incorporate A-, B-, or are sampled, the sampled data are converted to digital
C-weighting networks. The entire capability has been spe- binary form, and the binary numbers are fed to a digital
cifically designed to permit simple, rapid, and precise processor that computes root-mean-square levels.

303
The averaging method is true (linear) integration with a Complete systems, tailored to your needs from transducer
choice of nine accurate integration times from 14 second to final data storage, can also be supplied. Such asystem
to 32 seconds. This scheme not only ,produces answers could include acomputer, display scope, dc recorder, and
faster than the running-average circuits found in analog magnetic tape recorder to provide on-line calculations and
devices (which "waste" time and aren't very useful for comparisons such as:
transient signals) but also make it possible to determine
exactly what events in time have affected the answer. •Spectrum comparisons from 3.15 Hz to 80 kHz
The computed band levels are stored in digital memory to •Stevens loudness calculations per ANSI S3.4-1968
be retrieved at a rate limited only by the output recording •ARF loudness calculations
or storage device. The analyzer simultaneously provides •Perceived-noise- level ( PNL) computations recommended
both digital and analog outputs. by the FAA
A versatile solution The 1921 is available in bench or •Speech-interference- level (SIL) computations
rack models, in four standard frequency ranges, and with •Noise criterion levels per L. L. Beranek
or without attenuators. Custom versions with up to 45
octave bands, either 1/10, 1/3, 1/1 or a mixture of the •ARI, AMCA, ASHRAE, and STC ratings
three, are available on special order as are models with
special bandwidths. -See GR Experimenter for May-June 1969 and reprint E122.

specifications

Frequency: 3.15 Hz to 80 kHz. able: thin Mylar• sheets, of same size and scale factor as attenua-
tor display, attach to window with self-contained adhesive and can
Bandwidth: 1/3-octave standard; 1/10 and 1/1 octano available. be used to record position of dots in window with grease pencil
or other marker.
Amplitude: 70-dB dynamic range, 60 dB displayed; 100- mV rms
nominal full-scale sensitivity. Response: 30 6- pole Butterworth filters with I / 3-octave effective
(noise) bandwidth that meet ANSI $ 1.11-1966 Class Ill ( high
Linearity: -±- 0.5-dB deviation from best straight-line fit over top attenuation) and IEC 225-1966 standards. Accuracy of center
50 dB of display range, - ± 1 dB over entire 60-dB range. frequency. ± 2%. Level uniformity, within ±- 0.25 dB at 25 .C, -
±0.5
dB from 0 to 513°C. at center frequency with attenuator at + 25 dB.
Digital resentation: Band information is displayed on high-inten- Passband ripple, 0.5 dB max Bk- ph, Noise, < 15 uV equivalent
sity ne n-readout tubes. Band number per ANSI S1.6 and 51.111 input noise. Harmonic distortion at band centers. < 0.25% at 1-V
is displayed on 2 tubes and available as 2 BCD digits on rear output for bands centered below 25 Hz, < 0.1% at 1-V output at
panel. Band level from 0 to 159 dB in 0.25-dB steps is displayed 25 Hz and above. Weighting, A, B, and C conforming to ANSI S1.4,
on 5 tubes and available as 5 BCD digits on rear panel with over- IEC R123, and IEC R179,
load indicated as 8 or 9 in left digit. Rear-panel data are 1-2-4-8
weighted at standard 5.V nL levels (...5 + 0.5 and + 3.5 V) and Detection: RMS with true ( linear) integration. Integration times,
available from all bands sequentially at a 50-Pin type 57 connector. • V:, 1, 2, 4, 8, 16, or 32 S; pushbutton controlled. Sampling.
sampling rate is changed during integration time to eliminate
Analog Presentation: Band number is available as 0 to + I V linear coherence effects; 1024 samples taken during integration times
ramp at rear BNC connector. Band level from 0 to 60 dB is avail- from 1 to 32 s; below 1 s number of samples reduced in propor-
able as 0 to + 1 V ± 10% signal at rear BNC connector with over- tion to integration time to a minimum of 128. Repeatability, bet-
load indicated by superimposed fitter. Peak monitor, a peak ter than 1 dB (1 14 limit) for tone burst with duty factor of 1/100
detector senses lends at two circuit points and drives a panel (equivalent to crest factor of 23 dB) when rms levels are < 13 dB
meter calibrated in dB to overload lev. A signal pro- below full scale.
portional to meter indication is available at a rear connector to
drive a dc recorder; 1mA for full-scale reading. Programmability: All panel controls, except output display rate,
attenuators, and gain control, are programmable by closures to
Input: Connects to rear BNC or mike connector. Sensitivity, 100 ground applied to a rear 50- Pin type 57 connector.
mV rms nominal for full scale. Can be increased to 5 mV with
1560-P40 or - P42 Preamplifier. Dynamic range, 60 dB displayed Supplied: Power cord; 24-pin, 36- pin. and 50- pin type 57 plugs to
plus 10-dB crest- factor margin at full scale. Maximum input, 35 mate with rear connectors; 5 Transmission Record Sheets .
Vdc, 17 V peak ac. Impedance, 100 kn. Calibration, full-scale and
zero- level self calibration provided in two auxiliary channels; Available: 1560-P40 and - P42 Preamplifiers, 1566 Multichannel
panel control allows a calibration factor to be added to digital Amplifier ( input scanner), 1522 Dc Recorder, 1921-P1 Storage
output; full-scale indication is adjustable from 6 to 159 dB in Display Unit, Houston Instruments 6400.024 series plotters,
1-dB steps. Mohawk Data Sciences model 800 High- Speed Printer.

Attenuation: 18-dB continuous gain adjustment common to all Power: 100 to 125 and 200 to 250 V, 50.60 Hz, 135 W.
channels plus, optionally, 0 to 50-013 attenuation in 1-dB steps
±-0.25.dB accuracy ( re + 25- dB setting) by means of a panel
with - Mechanical: Bench or rack models. Consist of 1925 Multifilter and
thumbwheel switch for each band. Attenuation of each band is 1926 Multichannel RMS Detector. Dimensions Iwo ho dl: Bench.
indicated on panel display end represents the transmission be- 19.5 x 19020 in. ( 495 x483 x 508 mm); rack, 190 17.5 016 in.
tween input and summed output of multifilter. Display has stand- (483 x445 x406 mm). Weight, Bench, 95 lb ( 44 kg) net, 190 lb
ard 50-dB- per-decade scale factor, 10 dB per in. vertical, 5 in. per (87 kg) shipping; rack, 80 lb ( 37 kg) net. 120 lb ( 55 kg) shipping.
decade horizontal. 1925.9870 Transmission Record Sheets avail- 'Registered trademark of E. I. du Pont de Nemours and Co. Inc.

Catalog Number
YAM Attenuator Without Attenuator
Bench Rack Bench Rack Description
1821 Real-Time Analyzer
One-Third-Octave Bands
1921-9700 1921-9701 1921-9708 1921-9709 25 Hz to 20 kHz
1921-9702 1921.9703 1921-9710 1921-9711 12.5 Hz to 10 kHz
1921-9704 1921-9705 1921-9712 1921-9713 3.15 Hz to 2.5 kHz
1921-9706 1921-9707 1921-9714 1921.1915 100 Hz to 80 kHz

Description

111254470 Transmission Record Sheets, pack of ten

304
Type 1925 mULTIFILTER

Spectrum shaper or analyzer building block The 1925 peak and indicates in decibels referred to the overload
Multifilter contains up to 30 parallel octave- band or one- level.
third-octave-band filters from 3.15 Hz to 80 kHz and can Attenuators for each band Optional attenuators for each
be supplied with or without attenuators that permit in- filter broaden the Multifilter's usefulness. Each attenua-
dependent control of the gain in each band. With the tor provides 50 dB of gain control in 1-dB steps accurate
attenuators, you can use it as an equalizer or spectrum to 0.25 dB. Thumbwheel switches control the attenuation
shaper to simulate or to compensate for irregularities in and a panel display indicates the "transmission" of the
the frequency response of electrical or acoustical trans- instrument. This display has the same scale as the 1521-
mission systems or transducers. With or without attenua- 9463 chart paper used with the 1564-A Sound and Vibra-
tors, you can use it as the basis for a serial or parallel tion Analyzer ( 5 in./decade horizontal, 10 dB/in. vertical).
frequency analysis system. A key-operated lock guards against unintended changes in
the attenuator control settings.
A variety of outputs The outputs from the individual
filters are presented simultaneously in parallel, summed Filters meet American and international standards The
in a single output, and selected individually by manual filters, built on plug-in etched boards (three per board)
switching, by external switch closure, or by a remote for easy interchange, are available with either octave or
scanner control unit. Additional outputs provide the un- one-third-octave bandwidths that conform to both Ameri-
filtered input signal and the signal with A, B, or C weight- can and international standards. The A-, B-, and C-
ing imposed. Peak detectors located before and after the weighting characteristics also conform to the require-
filters drive a metering circuit that selects the highest ments of the various standards for sound- level meters.
-See GR Experimenter for May-June 1969 and reprint E122.

specifications
Frequency: 3.15 Hz to fil) kHz. band is indicated on panel display and represents the transmis-
sion between input and summed output. Display has standard
Bandwidth: 1/3 or 1/1 octave standard; 1/10 octave available.
50-dB per decade scale factor; 1040 per in. vertical, 5 in. per
Peak Monitor: A peak detector senses levels at two circuit points decade horizontal. Lock on panel prevents accidental changes in
and drives a panel meter calibrated in dB referred to overload attenuator settings. 1925-9670 Transmission Record Sheets avail-
level. A signal proportional to meter indication is available at • able: thin Mylar* sheets, of same size and scale factor as atten-
rear connector to drive a dc recorder; 1 mA for full-scale reading. uator display, attach to window with self-contained adhesive and
Input: Connects to rear BNC or mike connector. Gain, 0 dB can be used to record position of dots in window with grease
nominal. Maximum input, 35 V de, 17 V ac. Impedance, 100 ka. pencil or other marker.
Attenuation + 6 to — 2-dB continuous gain ad ustment common Response: 30 6-pole Butterworth filters with 1/3-octave effective
to all channels plus, optionally, +25 to — 2549 attenuation in (noise) bandwidths that conform to ANSI S1.11-1966 Class Ill
1-dB steps with -±0.25413 accuracy ( re +25-d0 etting) by means (high attenuation) and IEC 225-1966 standards or with 1/1-octave
of a panel thumbwheel switch for each band. Attenuation of each bandwidths that conform to ANSI S1,11-1966 doss II ( moderate
rate but highest for octave-band filters) and IEC 225-1966 stand-
ards. Accuracy of center frequency, ±- 2%. Leve/ uniformity,
within -1.-0.25 dB at 25•C, dB from 0 to 50•C, at center fre-
M111111111111111111111!filiMMZUM1111111111 quency with attenuator at + 25 dB. Passband ripple, 0.5 dB max
pk.pk. Noise, < 15 0)/ equivalent input noise. HallTIOrliC distortion,
IIIIIMIUM111111d1111111111111M1111111111 < 0.25% at 1-V output for bands centered below 25 Hz, < 0.1% at
MaMBIllIt1111111111BLIMIIIII111111111 1-V output for 25 Hz and above. Weighting, A, B, C, conforming to
ANSI 51.4. IEC R123, and IEC R179.
MMEM11111;11111,111111\1111111111111111111 Outputs: Parallel band outputs, ±-4.2 V max ( 3 V rms) behind 20 a
"MIEMIIIIM111111111111111111MMI111111111 nominal; 3 ka min load for max output voltage. Scanned band
output, ±4.2 V max ( 3 V rms) behind 20 a; 3-ki) min load for max
output voltage. Two chassis can be wired in parallel for up to 60
scanned outputs. Summed output .( for equalizing and shaping
IIBMIIM11111111111111111•30111111111111111 applications). J-4.2 V max open circuit behind 600 a; impedance
of load does not affect output linearity. Weighted and unfiltered
outputs, 0-013 nominal gain at 1 kHz, behind 20 12 nominal; 30k0
1. -mmommunfoiniumumaN1111111111 min load for max output voltage.

MIIVAIIIIIE111111111MIIMMUIC111111111 Supplied: Power cord, two 36- pin type 57 plugs to mate with mar
connectors.
'
SBMIIMIIIM1111111111111111•011111111101111111 Agymapillaifiheler: (in15
p6u0-
tscP4aannnner).
Oaenr0). -P42 Preamplifiers, 1566 Multichannel

MIA111111111511111111111MM11111111011111 Power: 100 to 125 and 200 to 250 V, 50-60 Hz, 17 W.

IITAIMP1111111111111111MMEIMI111111911 Mechanical: Bench or rack models. Dimensions ( w x h xd): Bench,

IA111111111Efill1111111111•111111111111‘111111111É 19.75 9.13 x 12.25 in. ( 502 02320311 mm); rack, 198.75
12.25 in. ( 483 x222 311 mm). Weight: Bench, 49 lb (43 kg) net, 58
lb ( 27 kg) shipping; rack. 39 lb ( 18 kg) net, 47 lb ( 22 kg) shipping.
t
f

305
 specifications iciiiire)
Catalog Number
With Attenuator Without Attenuator
Bench Rack Beach Rack Description
1925 Multi6Iter
One-Third- Octave 9
1925.9700 1925-9701 1925.9712 1925-9713 25 Hz to 20 kHz
1925-9702 1925-9703 1925-9714 1925-9715 12.5 Hz to 10 kHz
1925-9704 1925-9705 1925-9716 1925-9717 3.15 Hz to 2.5 kHz
1925-9706 1925-9707 1925.9718 1925-9719 100 Hz to 80 kHz

Octave
1925-9708 1925-9709 1925-9720 1925-9721 31.5 Hz to 16 kHz
1925-9710 1925-9711 1925-9722 1925-9723 4 Hz to 2 kHz

bond models,

Type 1926 MULTICHANNEL RMS DETECTOR

• 1 Hz to 100 kHz

-.311- 11111111111111•1111111MIIIIIIke. • 70-dB dynamic range,

60 dB displayed
• digital detection provides
25 53 5 o true rms answers
• fully programmable

Digital detection The 1926 is the digital detector section ful as a multichannel true-rms detector in other applica-
of the 1921 Real-Time Analyzer and is available separately tions such as monitoring the sound pressure at a number
for use with the 1925 Multifilter or other multichannel of points or measuring the electrical noise of multichan-
filters where real-time analysis is required. It is also use- nel systems.

specifications
Frequency: 1 Hz to 100 kHz (3 dB down M < I Hz for bands 1 to Detection: RMS with true ( linear) integration. Integration times,
10, at < 6 Hz for all other bands; < 1-dB frequency-response error / ,1
2
1 / ,1
4 / ,1, 2, 4, 8, 16, or 32
2 pushbutton controlled. Sampling,
at 100 kHz). sampling rate is changed during integration time to minimize
coherence effects; 1024 samples taken during integration times
Amplitude: 70-d13 dynamic range, 60 dB displayed, 1 V rms :C10% from 1 to 32 s; below 1 s number of samples reduced in propor-
funscale sensitivity. tion to integration time to e minimum of 128. Repeatability,
better than 1 dB ( 1 8 limit) for tone burst with duty factor of
Linearity: ±-0.5-0B deviation from best straight-line fit over top 1/100 (equivalent to crest factor of 23 dB) when rms levels are
50 dB of display range, -2.-1 dB over entire 60-dB range. < 13 dB below full scale.

Digital Presentation: Band information is displayed on high- inten- Programmability: All panel control functions, except output dis-
sity neon- readout tubes. Band number per ANSI S1.6 and 51.111 play rate, are programmable by closures to ground applied to o
is displayed on 2 tubes and available as 2 BCD digits on rear rear 50- pin type 57 connector.
panel. Band level from 0 to 159 dB in 0.25-dB steps is displayed
on 5 tubes and available as 5 BCD digits on rear panel with over- Supplied: Power cord, 24, 36, and 50-pin type 57 plugs to mate
load indicated as 8 or 9 in left digit. Rear-panel data are 1.2-4-8 with rear connectors.
weighted at standard 5-V TTL levels (± + 0.5 and 3.5 V; 15-V
DTL available on special request) and available from all bands Available: 1522 DC Recorder, 1921-P1 Storage Display Unit, Hous-
sequentially and automatically in 720 vs to 45 s for 45 bands or ton Instruments 6400-024 series plotters, Mohawk Data Sciences
manually by pushbutton or external command, at a rear 50- pin model 800 High- Speed Printer.
type 57 connector.
Power: 100 to 125 and 200 to 250 V, 50-60 Hz, 135 W.
Analog Presentation: Band number is available es 0 to -I-1-V linear
ramp at rear BNC connector. Band level from 0 to 60 dB is avail- Mechanical: Rack model only. Dimensions ( w x h xd): 19 8.75 e
able as 0 to + 1 V ± 10% signal at rear BNC connector with over- 17.44 in. ( 483 x222 x443 mm). Weight: 47 lb (21 kg) net, 55 lb
load indicated by superimposed jitter. (25 kg) shipping.

Input: 30 or 45 channels. Sensitivity, 1 V nns -± 10% full-scale.

Dynamic range, 60 dB displayed plus 10-dB crest.factor margin at
full scale. Maximum input, 100 mVdc, 3 V peak ac for linear opera-
tion. Impedance, w 5 Ittl; 0 to 30 CI allowable source impedance.
Calibration, full-scale and zero- level self-calibration provided in 1926 Multichannel RMS Detector, rack model
two auxiliary channels; panel control allows a calibration factor to 1926-9701 30 Channels
be added to digital output; full-scale indication is adjustable from
1926 ,9703 45 Channels
0 to 159 dB in 1-d111 steps.

306
Type 1921-P2 STORAGE DISPLAY UNIT

lo especially useful with real-time analyzers

• clear, bright spectrum display
• large 8x10cm viewing area
• bench or rack models

Spectrum at aglance As an accessory for the 1921 Real- inch bistab e storage CRT, built-in vertical and horizontal
Time Analyzer, the 1921-P2 displays the results of the deflection amplifiers, and Z-axis modulation capability. It
spectrum analysis— amplitude vs frequency — in asingle, is supplied with inter-connecting cable for the 1921 ana-
easily interpreted format. The display functions of the lyzer and graticules marked with 1 2 -
/ octave standard band
unit are programmable and necessary control and deflec- numbers, center frequencies, and a decibel scale.
tion voltages are provided by the analyzer.
The 1921-P2 is a slightly modified Tektronix* Type 603 •Registered trademark of Tektronix Inc.
Storage Display Unit. It has all solid-state circuits, a 5-

specifications
Frequency: Useful to 100 kHz for waveform display of Y or T plot Power: 90 to 136 end 180 to 272 V, 48-440 Hz, 57 W.
with 1. phase shift between X and Y. Mechanical: Bench or rack models. Dimensions ( w x h x0): Bench,
Display: 8 cm vertical X 10 cm horizontal display on 5- in, flat- 8.5 X 6 X 17.5 in. (216 x 152 x 445 mm); rack, 19 x 5.25 x 17.5
faced bistable storage tube phosphor similar to Pl. Linearity, in. (483 X 133 x 445 mm). Weight: 18 lb ( 8 kg) net, 25 lb ( 12 kg)
2% vertical and 5% horizontal difference between any 2 cm. shipping.
Speed, > 5 cm/ ms. stored line-writing speed. Storage. Z-axis on-
time should be be 9 es to ensure good storage. Viewing time up to Catalog
15 m recommended; erasure becomes more difficult if information Number I Description
is stored longer. Erasure, 200 ms erase time.
1921-P2 Storage Display Unit
Supplied: Five graticule grids graduated vertically every 5 dB from
Bench Model
0 to 60 dB and horizontally in 30 bands, 1 each for bands 5 to 34,
Rack Model
11 to 40, 14 to 43, and 20 to 49, and one with bands unmarked.

Type 1566 MULTICHANNEL AMPLIFIER

• 16 channels
• manual or remote channel selection
• 2- Hz to 100-kHz response
• 55-dB gain, manually or remotely adjusted
• calibration noise source built in

Many inputs — one output Many sound and vibration vides a built-in pink-noise calibration source that speeds
measurements can be simplified by use of a scanner that not only the check out of the scanner but also that of any
connects, in sequence or in any arbitrary order, the outputs analyzer connected to it. The 1566 is particularly useful
from a number of transducers to a single analyzer. A scan- with the 1921 Real-Time Analyzer. This combination can
ner system can be set up to measure signals individually automatically analyze the spectrum from each transducer
scanned or it can measure the space-averaged spectrum
or to average all signals.
using 2, 4, 6, 8, 10, 12, or 16 microphones. This feature
The 1566 scans up to 16 channels ( up to 99 with aspecial makes possible automatic real-time sound-power meas-
additional unit), amplifies each by up to + 55 dB, and pro- urements.

specifications
Channels: 16 plus 1 for calibration, expandable to 99 ( additional is — 3 dB per octave. Noise signal applied to internal calibration
channels housed in a special unit). Control: Active channel is se- channel is adjustable from 30 to 100 mV rms. Rear- panel noise
lected manually or by external 1-2-4-8 BCD signal, or automati- output is fixed at 100 mV rms and can be loaded by 0.05 µF
cally scanned in sequence with range of channels to be scanned without affecting spectrum up to 100 kHz.
selected by thumbwheel switches; dwell time adjustable from Supplied: Power cord, two 24-pin data plugs.
100 ms to 10 s or infinity (channel advance initiated by external
signals); scan set to occur once or repetitively and started, stopped Available: 1560-P40 and - P42 Preamplifiers ( 1566 provides power
on active channel. or reset to lowest channel by pushbuttons or for up to 99 of either). 1566-9500 Cable Set for connection to 1921
external closures to ground. Display: Two, high- intensity neon Real-Time Analyzer, microphones, vibration pickups.
readout tubes display active channel. Power: 100 to 125 or 200 to 250 V, 50 to 60 Hz, 30 W.
Frequency: 2 Hz to 100 kHz, flat within i-0.5 dB.
Mechanical: Bench or rack mount. Dimensions ( w x h x d): Bench.
Sensitivity: 1.8 mV to 1.6 V for 1-V output; gain set in 1-dB in- 19 1
/ x 5 x 20 in. ( 495 x 130 x 510 mm); rack, 19 x 31
2 /
2 x 18 1
/
2 in.
crements by panel control or 1.2-4-8 BCD signal at standard OIL (485 x 89 x 470 mm). Weight: Bench. 32 lb ( 15 kg) net, 47 lb
levels ( logic 0 c ground, logic 1 • 3 V). Rear-panel adjust-
(213 kg/ shipping; rack 26 lb ( 12 kg) net, 41 lb ( 19 kg) shipping.
ment provides 10-de continuous control of gain for all channels
for calibration. Each channel includes a 6.dB gain adjustment Ca talog
for transducer sensitivity equalization. Description
Impedance: Input, 100 ka. Output, 600 0.
Noise: < 10 µV equivalent input noise in each channel when gain 1588 Multi:funnel Amplifier
is maximum and source impedance is ,`" 100 D. 1566-9700 Bench Model
Calibration: Built-in pink- noise (± 0.5 dB) source with symmetrical 15669701 Rack Model
Gaussian distribution from 2 Hz to 100 kHz. Spectrum- level slope 1566-9500 Coble Set

307
Type 1522 DC RECORDER

• 2-mV/inch and 0.2- eA/ inch sensitivity

• 50-inch/second writing speed
• 0.25% linearity, 0.5% accuracy
• programmable writing functions
• plug-in versatility: grounded or
differential input

Bench model shown with 1522-P1 Preamplifier.

specifications

INPUT WITH 1522-P1 PREAMPLIFIER Linearity: -±- 0.25% of full scale.

Controlled by range switches, polarity switch, and con-
tinuous control with calibrated position that operates on all ranges. Deedband: -± 0.15% of full scale.
DC Voltage, 2 mV/ in. to 100 Win.; 15 ranges, 1-25 sequence. DC
Current. 0.2 . A/ in. to 100 mA/ in.; 18 ranges. 1-2-5 sequence. Zero Adjustment: 10-turn pot, can be set over full range.
Accuracy: -. 0.5% of full scale.
Chart Speed.: 0.5, 1, 2, 5, 10. 20 seconds, minutes, hours per
Linearity: , 0.25% off full scale, including recorder linearity. inch; 18 speeds. Chart moves in 0.0067- in. increments and can be
Stability: < 0.01%/day drift typical in 02 V/ in. range after warmu0. started or stopped in one step.
Input Isolation: > 1000 MD de front LOW to GROUND terminal
at 200 V max dc; 0.22 µF ac. Voltage, 200 V meat dc or peak ac. Chart-Speed Synchronization: Sync outputs permit other 1522 re-
corders to run at identical speed or at other standard speeds in
Common- Mode Rejection: 70 dB dc with 1411 source impedance:
synchronization with master recorder.
40 dB ac at 60 Hz.
Input Resistance: Voltage, 1 MD, Current, 1 I) to 10 k0 depending Programmability: All chart control functions fully programmable
on scale as follows: and outputs provided for full system integration.

Scale Resistance Remote-Control Functions: Require switch or solid.state closure to

ground. Controls: pen lift and pen down; two event markers, all
0.2µA/ in. to 2,01/ in. 10 ka chart speeds; chart start, stop, forward, reverse; fast scan (2 in./s)
20 x lEiruV/ in. with pen- lift: slow scan ( 2 in./min) with pen- lift; record command
5i:A/in. to 100.A/in. (drops pen, starts chart at selected speed); pen motion stopped in
Scale in 01/in.
position by servo blanking.
20mV/ in.
0.2MA/in. to 10mA/in. Remote-Control Outputs: Start, stop, forward, reverse, servo-posi-
scale in m11/ in. tion error voltage 10 state for position error of <0.5%), retrans-
20mA/in. to 100mA/ in. 10 mitting potentiometer, three independent solid-state closures cor.
responding to lines printed on paper: pen down; 300 pps sync and
motor- speed sync selected by chart- speed controls.
Offset and Drift: Voltage. Adjustable o zero. Drift, ±-25 .1.r/ •C from
0 to 50°C after warmup; warmup drift <0.5 mV. Current (bias). Other Outputs: Power for two additional stepper motors, power for
0.1 nA at 25*C; doubles each rise of 11 .C. externally controlled dc reference voltage.
INPUT WITH 1522-P2 DIFFERENTIAL PREAMPLIFIER Accessories Supplied: 274- NQ 3-ft double.plug patch cord, fastreke
Frequency: - 3 dB at 50 Hz ( 3rd order Butterworth response). Marker Set of 12 assorted-color pens. Event- Marker Set of 4 red,
4 black pens, 2 chart- paper rolls type 1522-9640, 2 potentiometer
Ranges: Controlled by range switches, polarity switch, and con. contacts. 2 paper cap assemblies, power cord. 50 chart mounting
tinuous control with calibrated position that operates on all ranges. sheets.
voltage, 2 mV/in. to 100 Win.; 15 ranges, 1-2-5 sequence. Current,
0.2 iiA/ in. to 100 MA/in.: 113 ranges, 1.2-5 sequence. Accessories Available: 1522-P11 Limit- Switch Set provides two ad-
justable limit stops; pen at limit closes reed- relay contacts with
Accuracy: ±..-03.5% of full scale when in calibrated position.
50-V. 500-mA dc rating, 150.V breakdown rating.
Linearity: ±
--0.25% off full scale, including recorder linearity.

Power: 100 to 125 or 200 10 250 V, 50-60 Hz, 90 W.

Input Isolation: frorn GUARD terminal to ground, in parallel
with < 500 pF. Voltage, 500 V do or peak ac.
Mechanical: Bench or nick models. DIMENSIONS Iwishxd): Bends,
Common- Mode Rejection: 160 dB dc. 80 dB 60 Hz, undriven guard, 19.7505.97015.88 in. 150201520429 rem), rack, 1905.22018.78 in.
typical, 180 dB ac up to 20 kFlz, driven guard, typical. 148301330427 mm). WEIGHT: Bench, 43 lb 120 kg) net. 58 10 ( 27 leg)
shipping, reek 39 lb lie kg) net. 54 lb 125 kg) Mifflin. -Pl. 15 lb 10. 7 lag)
Offset and Drift: Voltage, Adjustable to zero. Drift, ± (
25 iiVi*C net, Bib C.7 kg) shipping; -P2, 3.5 lb 11.8 kral net, 1010 (4.6 kg) shipping.
0.005% of full scald ' C) from 0 to 55°C. Current ( bias). 0.1
no at 25°C; doubles each rise of 11 .C.
Input Resistance between HIGH and LOW terminals: Voltage. 1 MO. Description
Current, 0.11 to 40.0 11 depending on scale as follows:
1522 DC Recorder, less preamplifier
Bench Model
Scale Resistance Rack Model
10 ii1// in. 0.06
1522-P1 Preamplifier
02 µA/in. tO 0.2 mA/ in. 1522-P2 Differential Preamplifier
scale in isA/ in.
1522-P11 Limit-Switch Set, for two adjustable limit
20 rnVi in. stops
0.5 mA/in. to 10 mA/ in. 1522-8813 Extender Board Kit, for servicing ease
scale in mA/in.
1522-9870 Cable Set, for connection to 1921
1522.8612 Event-Marker Set. 4 black. 4 red Dens
20 rnA/ in. to 100 mA/ in. 111
1522-9638 Mounting Sheets, 8:,,z x 11 in. sheets
with adhesive strips to mount charts for filing
RECORDER RESPONSE ( with 2- in. eatursi.) in 3- ring notebooks, 50 charts Per Pace
Fast Writing Speed: 50 in./a with < 3% overshoot. fastrake Marker Sets

Slow Writing Speed.: " I 301"I 10 I5 I 2 1 I0.5 in./ s 1522-9614 Red. 4 pens
Servo Bandwidth ( 3 dB) 30 20 15 7.5 4 1.5 0.75 0.4 Hz 1522.9615 Green. 4 pens
(for Va -
in. excursions/ 1522.9616 Blue. 4 pens

308
Other Instruments

RANDOM- NOISE GENERATORS — Type 1381, Type 1382

bleflPiet)
it#

a 2 Hz to 2, 5, or 50 kHz, Gaussian distribution • 20 Hz to 50 kHz, Gaussian distribution

Type • adjustable clipping Type • white, pink, or ANSI spectra
1381 1382 • 3-V rms output, balanced, unbalanced,
• 3-V rms output
or floating

Electrical noise is, by definition, any unwanted dis- normally occur in experimental measurements — the nor-
turbance and its reduction in communications circuits is mal or Gaussian distribution. In general-purpose noise
a constant aim of the engineer. Noise from a controlled generators the design objective is random noise that is
source, however, is useful in studying the effectiveness of Gaussian and has a uniform spectrum level over the speci-
systems for detecting and recovering signals in noise. fied frequency range.
Well defined random noise is, moreover, a remarkably The General Radio random-noise generators produce
useful test signal that has, for many measurements, prop- electrical noise at high output levels, each model having
erties that are more useful than those of asingle-frequency been designed for specific uses. The 1381 is useful for
signal. Its wide frequency content sometimes permits one many audio-frequency applications, and also in vibration
test with random noise to replace a series of single- testing as its spectrum extends well into the subaudio
frequency tests. Noise is also useful in simulating speech, range. The 1382 is intended for audio-frequency electri-
music, or communications circuit traffic.
cal, acoustical, and psychoacoustical applications. The
Noise is called random if its instantaneous amplitude
1390-B is useful at higher frequencies because its spec-
at any future instant is unpredictable. Random noise is
specified by its amplitude distribution and by its spec- trum extends to 5 MHz. The 1383 generates wide- band
trum. Many types of naturally occurring electrical noise noise of uniform spectrum level and is particularly useful
have the same distribution of amplitudes as do errors that for tests in video- and radio-frequency spectrums.

specifications
Output Voltage: > 3 V ohs max, open-circuit, for any bandwidth.
SPECTRUM
1381: Flat ( constant energy per hertz of bandwidth) cl dB from
Output Impedance: 600 O. Can be shorted without causing distor-
2 Hz to 1, 2.5. or 25 kHz; upper- cutoff frequency ( 3-dB point) tion. 1381 output is unbalanced; 1382 output is floating, can be
can be switched to 2, 5, or 50 kHz. Spectral density at 3.V output
connected balanced or unbalanced.
and for 1- Hz bandwidth is approx 64, 40, and 13 'TN, respectively,
for 2-, 5., and 50.kHz upper cutoff. Upper cutoff Slope is Amplitude Control: Continuous adjustment from full output to
12 dB/octave. See curve. approx 60 dB below that level.

1382: Either ( a) white noise ( constant energy per hertz bandwidth) Terminals: 1381 output at front-panel binding posts and rear- panel
rt 1 dB, 20 Hz to 25 kHz, with 3-0E1 points at anomie 10 Hz and BNC connector; 1382 output at front-panel binding posts and rear-
50 kHz; ( b) pink noise ( constant energy per octave bandwidth)
panel jacks for double plugs.
C.1 dB, 20 Hz to 20 kHz; or ( c) ANSI noise, as specified in ANSI
Standard 51.4-1961. See curve. Accessories Supplied: Power cord, rack- mounting hardware with
rack models.

W OOOOOO ins
Amplitude- Density Power Required: 100 to 125 or 200 to 250 V, 50 to 400 Hz, 6 W.
Gaussian Probability.
Voltage Density Function Distribution of 1381 1382
Mounting: Convertible- Bench Cabinet.
o 0.0796 0.0796 It- 0.005
ntn 0.0484 0.0484 ± 0.005 Dimensions ( width x height x depth): Bench, 8Va x 31/
2 o 91
/
2 in.
±-20 0.0108 0.0108 0.003 (220 x 99 0250 rnol): rack. 19 3'c o 9 in. ( 485 89 x 230 mm).
±-30 0.000898 0.000898 00 0002
±40 0.0000274 0.0000274 ± 0.00002 Weighs Net. 7 lb ( 3.2 kg), shipping, 10 lb ( 4.6 kg).

Catalog
These data measured in a - window" of 0.2n. centered on the in- Description
Number
dicated values; n is the standard deviation or mu value of the
noise voltage. Random-Noise Gen rrrrrr
1381-9703 2 Ho to 50 kHz, Bench Model
Clipping: The output of the 1381 can be clipped internally to re-
1381-9701 2 Hz to 50 kHz. Rack Model
move the occasional wide extremes of amplitude. Clipping. if
1382-9700 20 Hz to 50 kHz. Bench Model
desired. is adjustable to approx 2. 3. 4. Or 5n. Such clipping has
negligible effect on the spectrum or the rms amplitude. 1382-9701 20 Wz to 50 kHz, Rack Model

309
 Type 1390-B RANDOM- NOISE GENERATOR

• 5Hz to 5MHz
• 3O-»Vto 3-V output
• ± 1-dB audio-spectrum- level uniformity

*41iMm• • e

specifications

typical Spectrum
Frequency Range: 5 Hz to 5 MHz.
Range
Output Voltage: Max open-circuit output is at least 3 V for 20-kHz Spectrum Level Uniformity •
range. 2 V for 500-kHz range, and 1V for 5- MHz range.
20 kHz 5 mV for 1- Hz band within del dB, 20 Hz to 20 kHz
Output Impedance: Source impedance for max output is appro. 500 kHz 1.2 mV for 1- HZ band within 2
-e3 dB, 20 Hz to 500 kHz

900 G. Output is taken from a 250047 potentiometer. Source im- 5 MHz 0.6 mV for 1- Hz band within ±-3 dB, 20 Hz to 500 kHz;
pedance for attenuated output is 200 G. One output terminal within etill dB, 500 kHz to 5 MHz
is grounded. •Noise energy also present beyond these limits. Level is down 3 dB
at 5 Hz. Se plot.
Waveform, Noise source has good normal, or Gaussian, distribu-
tion of amplitudes for ranges of the frequency spectrum that are
narrow compared with the band selected. Over wide ranges the
distribution le less symmetrical because of dissym,netry intro-
duced by the gas tube. Some clipping occurs on the 500-kHz
and 5- MHz ranges.
°
Voltmeter: Rectifier-type averaging meter measures output. It is .00
calibrated to read rms value of noise.
IM(01.10.
Attenuate::: Multiplying factors of 1.0, 0.1, 0.01, 0.001, and 0.0001.
Accurate to eL-3% to 100 kHz, within - ± 10% to 5 MHz. TYP ea spectrum- level characteristics.

Power Required: 105 to 125 or 210 to 250 V, 50 to 400 Hz, 50 W.

Amessedm SMPlied: CAP-22 Power Cord. Weight: Net, 12 lb ( 5.5 kg); shipping, 16 lb ( 7.5 kg).

Accessories Available: Rack-adaptor set ( 19 o 7 in.); 1390-P2.

Mounting: Convertible- Bench Cabinet.
gt.tee.1
Dimensions
190 x 250 min).*
( width x height o depth): 124Is x 71
/
2 x 93
/
4 in . ( 325 1360-13 Randoen-Nohe °mentor
1390-9702 115.v Model
1390-9703 230-V Model

"1-1— — 1
• I

,
0 15 250, 100
10 SS

(Left curre) White noise output of the 13908 Random.Nolse Generator as measured by a one-
third
octave.bandwidth filter and ( right curve) pink-noise output.

310
 Type 1569 AUTOMATIC LEVEL REGULATOR

• 2 Hz to 100 kHz
• 50-dB control range
• acoustic-system accessory

VOLTAGE
The 1569 Automatic Level Regulator is intended as an '"V°L. CONTROLLED •IIPLIPIER
INPUT ATTENUATOR
accessory for an oscillator or for a source of narrow-
band noise. Its primary functirn is to control the signal
level in swept-frequency sound and vibration tests.
The regulator senses a control voltage from a micro- LP PA-TER WIT., CONTROL
AVERAGE
phone, accelerometer, or other transducer monitoring the ADJUSTAGLE
CUT- 0, DETECTOR
SIGNAL
•MPLIPIE
5
,41
On
sound or vibration to be controlled and adjusts its output ARTOUENCT

to maintain constant level (see diagram). Output level is 018V3111 of 1569 Automatic Level Regulator
indicated by a panel meter with a linear-dB scale, show-
operating frequency and magnitude-phase conditions in
ing the operator where in its 50-dB control range the
regulator is operating. Regulation is such that a level the control loop.
The 1569 can also be used to regulate voltage from
variation (without the regulator) of 25 dB, for instance, an oscillator or other signal source. In this mode, the
is compressed to a variation of 1 dB. The control rate
is adjustable by means of a panel control to suit the control range is limited to 15 dB.

— See GR Experimenter for April 1968.

specifications
Noise: Typically better than 65 dB below 3 V in 100-kflz band.
Frequency Range: 2 Hz to 100 kHz. Harmonic Distortion: < 1% total for < 1-V output level.
Control Range: 50 dB. Automatic " Shot- Damn": A loss of drive ( input) voltage from sig-
Compression Ratio: 25 ( 0.04 dB per dB). nal source causes the output voltage to drop to zero to protect
DRIVE ( INPUT) equipment connected to output.
Voltage Required ( for normal operation): 1 V. CONTROL-SIGNAL INPUT
Impedance: 100 Id/. Voltage: 5 rnV to 4 V required.
OUTPUT Impedance: 25 MU.
Voltage: 3 V max to 10 rnV min. Control Rates and Corresponding Min Operating Frequencies:
Impedance: 600 0 Any load impedance can be connected with- 1000 dB/s 1200 dB/s 1100 dB/5 130 dEl/s 1 10 dB/s 1 3 dB/s
out affecting linear operation of output circuit.
600 Hz F20060 1 60 Hz 1 20 Hz 1 6 Hz 1 2 Hz

Power Required: 100 to 125 or 200 to 250 V ( switch selected),

50 to 60 Hz, 4 W.
Accessor ies Supplied: Power cord, mounting hardware with rack
or bench models.
Accessories Available: GR 1560-P40 Preamplifier ( power for pre-
amplifier available at rear- panel input connector): 1304 Beat-
Frequency Audio Generator, 1521 Graphic Level Recorder; micro-
phones and vibration pickups.
Mounting: Rack- Bench Cabinet.
Dimensions ( width e height x depth): Bench model, 19 x 3,/a x
13 in. ( 485 x 99 x 330 mm); rack model, 19 x 31 / x 10 1
2 / in. ( 485
2
x89 x275 mm).
Weight: Net, 13 lb ( 6 kg): shipping, 30 lb ( 14 kg).

Catalog 1
Number I Description

l 1569 Automatic Lem Regulator

1569-9700 Bench Model
1569-9701 Rack Model
Typical Measurement System Using 1569

311
Type 1952 UNIVERSAL FILTER

• 4- Hz to 60- kHz tuning

• low-pass or high-pass,
band-pass or band- reject, ganged for easy tuning
• high attenuation rate— 30 db/octave
• line or battery operation

The 1952 Universal Filter will perform as a low-pass, ate the effect of limited bandwidth upon signal intelligi-
high-pass, band-pass, or band-reject filter at the turn of a bility and data-transmission accuracy. As a high-pass filter
panel switch. It consists of a low-pass and a high-pass it can reduce power- line- related components, as a low-pass
filter that can be employed singly, in cascade, or in par- filter control high-frequency noise, or as a notch filter
allel, to provide the assortment of over-all characteristics. eliminate single-frequency components. The 1952 can
The cut-off frequencies of the two filters can be controlled also act as part of a spectrum analyzer or distortion
independently or ganged together to provide constant- meter and, with a random- noise generator, produce con-
percentagie bandwidth for band-pass or band- reject tuning. trolled bands of noise as test signals. It is recommended
This filter is of value in many signal-conditioning ap- as an accessory for the GR 1142 Frequency Meter and
plications. For example, it can be used to control system Discriminator and the 1561-R Precision Sound- Level
bandwidth for reduction of extraneous signals or to evalu- Meter.
— See GR Experimenter for April 1968.

specifications
FREQUENCY RANGE
Cut-off Frequencies: Adjustable 4 Hz to 60 kHz in four ranges.
Pass- Band Limits: Low-frequency response to dc lapprox 0.7 Hz
With ac input coupling) in LOW PASS and BAND REJECT modes.
High-frequency response uniform - L-0.2 dB to 300 kHz in HIGH
PASS and BAND REJECT modes.
Controls: Log frequency-dial calibration; accuracy -
±2% of cut-off
frequency ( at 3-dB points).
FILTERS
Filter Characteristics: Filters are fourth-order ( four-pole) Chebyshev
approximations to ideal magnitude response. The nominal pass-
band ripple is .± 0.1 dB ( - ±-0.2 dB max); nominal attenuation at
the calibrated cut-off frequency is 3 dB; initial attenuation rete
is 30 dB per octave. Attenuation et twice or at one-half the se-
lected frequency, as applicable, is at least 30 dB.
Tuning Modes: Switch selected, LOW PASS, HIGH PASS, BAND
PASS, and BAND REJECT.
Ganged Tuning: The two frequency controls can be ganged in Low-pass and
BAND PASS and BAND REJECT modes so the ratio of upper to high-pass filter
lower cut-off frequencies remains constant as controls are ad- characteristics.
justed. Range overlap is sufficient to permit tuning through
successive ranges without the need to reset frequency controls
if ratio of upper to lower cut-off frequencies is 1.5 or less. Noise: < 100 •V in an effective bandwidth o 50 kHz.
Minimum Bandwidth: 26% ( approx ‘,5 octave) in BAND PASS Distortion: Max harmonic distortion, with all components in the
mode. pass band, for a linear load, is less than 0.25% for open-circuit
voltages up to 3 V and frequencies up to 50 kHz.
Null Tuning: In BAND REJECT mode, setting the frequency con-
trols for a critical ratio of upper to lower cut-off frequency ( in- Power Required: 100 to 125 or 200 to 250 V ( switch selected).
dicated on dials) gives a null characteristic ( point of infinite at- 50 to 60 Hz, 2.5 W. Or 19.2 V, approx 20 rnA from rechargeable
tenuation) that can be tuned from 5 Hz to 50 kHz. nickel-cadmium batteries ( not supplied), about 10-h operation.
Connections for ezternal battery.
INPUT Accessories Supplied: POwer cord, bench- or rack- mount hardware.
Gain: 0 or — 20 dB, switch selected. Accuracy of gain is - ±-1 dB,
Accessories Available: Rechargeable batteries ( two required) and
of 20-cIB attenuator is dB.
1560-P60 Battery Charger.
Impedance: 100 kO.
Dimensions ( width x height o depth): Bench, 19 a Pre x 15 in.
Coupling: Ac or dc, switch selected. Lower cut-off frequency (485 o 99 o 385 min): rack, 19 a 31 / x 1144 in. ( 485 x 89 x
2
(3 dB clown) for ac coupling is about 0.7 Hz. 300 mm), charger, 43/ a 33
4 / coin, (
4 110 0 960205 mm).
Max Voltage: Max sine-wave input is 3 V rocs ( 8.4 V pk-pk) or 30 Weight: Net, 20 1
/ lb ( 9.5 kg), shipping. 25 lb ( 11.5 kg).
2
V rms with input attenuator at 20 dB. Max peak input voltage
for dc coupling is V. For ac coupling max peak level of ac
component must not exceed ±- 4.2 V and dc component must not
exceed 100 V. Input can tolerate peak voltages of -±- 100 V with- Wee Description
out damage. An LC filter at input limits bandwidth to 300 kHz, 1952 Universal Filter
thus reducing danger of overloading active circuits at frequencies 1952-9801 Bench Model
above normal operating range. 1952-9811 Rack Model
GENERAL 8410-1040 Rechargeable Battery ( 2 req'd)
Output: 600-0 impedance. Any load can be connected without 1580-P60 Battery Charger
affecting linear operation of output circuit. Temperature coeffi- 1560-9660 115 volts
cient of output offset voltage is between 0 and +4 mV) * C. 1560-9661 230 volts

312
 MICROPHONES

MICROPHONES

15601.5 and -P6 1-INCH CERAMIC MICROPHONES

-See OR Experimenter for May-June 1967

rr.,
:rreer.2 2 kHz
MITIResseewommeM
10
seekH z, flat -±
-
r?-dB
. me creseref
cpidr=cc rerezo.nni41led
.E
it
,

is typically 0.08 s.
Senertieifle -62 dB min, -60 dB typical, re 1 Wader. TEMPERATURE CO-
EFFICIENT, w -0.01 dElrC. MAXIMUM SOUND-PRESSURE LEVEL, 150 dB
With < 1% distortion; at - 184 and + 174 dB peak, microphone fails.
at
2
1 regvel;Ei
eriee CO5E 'fICIE4rol at
log*Î.'
21p
5M 6 irCl
i0
rn5l3Pt
o el2 61e. %
F

.y100%52
%sti may coc ar . with no damage, at 95 •C
0.5-dB permanent
- 55

I•10... •
Mechanicel: TERMINALS, 3-pin mike connector; both terminals can be
floated with respect to ground for hum reduction. DIMENSIONS: cartridge
only, 1.13 in. ( 29 nun) long, 0.936 Tt .002 in. ( 23.7 mm ± 50 p.m) die: 1560-
P5 assembly. 2.31 in. ( 59 mm) long, 0.94 in. ( 24 mm) die; 1560-P6 as- Random.incidence response for 1560.P5 and • P6 microphones
sembly, 11.75 in. (298 mm) long, 0.94 in. ( 24 mm) dia. WEIGHT: 1565-P5,
0.2 lb (0.1 kg) net, 1 lb (0.5 kg) shipping; 1560-P6, 0.7 lb (0.3 Ice) net, 2 lb 1560.P5 Microphone 1560-9605
10.9 kg) shipping. 1560-P6 Microphone Assembly 1560-9606

1560-9570 1-INCH CERAMIC MICROPHONE CARTRIDGE Same cartridge

as 1560.1.5 and -P6 above, with adaptor to fit 1560-P42 Preamplifier.
%/MOW 1-in. Commis Microphone Cartridge 1560-9570

1972 % ARCH CERAMIC MICROPHONE

Tentative Specificatioes lide..0.4.

Frequency: 5 Hz to 20 kHz, flat T-

.2 dB random.inCidence response, ti dB
from 5 Hz to 12.5 kHz ( see curve). Time constant of pressure-equalizIng
!» ;›
leak is 0.08 stypical.
Sensitivity, -80 dB nominal, -85 dB min, re 1 Weber. TEMPERATURE
COEFFICIENT: v.0.01 clIfIrC. MAXIMUM SOUND-PRESSURE LEVEL:165dB
With < 1% distortion; at + 184 and - 190 dB peak, microphone may fall.
Impedance: 395 pF - ± 15% at 23 .C. TEMPERATURE COEFFICIENT: 2.2
pFi fC from 0 to 50 .C.
Environmentelf TEMPERATURE: -40 to +60 .0 with no damage: at
+95 .0 a 0.5-dB permanent sensitivity loss may occur. HUMIDITY: 0 to
100% RH.
Mechanical: TERMINALS: coaxial with 0.46-60 thread for preamplifier
mounting. DIMENSIONS: 0.5 in. dia x 0.78 in. long ( 13020 mm).
WEIGHT: 0.5 co ( 14 g) net, 0.5 lb ( 0-3 kg) shipping. IA-in. Ceramic Microphone 1972-9601

The following microphone sets include a microphone cartridge, a 1560-P42

Preamplifier, all adaptors necessary to mate the cartridge to the preampli-
fier and to a 1562 Sound.Level Calibrator, and a carrying case for all corn.
ponents including the preamplifier.
DIMENSIONS ( w x h x dl: 10 x 2 x 7.25 in., 254 51 x 184 mrn.
WEIGHT: 2.5 lb ( 1.1 kg) 961.61k ( 2.7 kg).
Nominal Nominal typIcal
Frequency Range Sensitivity' atoado R
re 1 V/ obar re 20 WV/ mf

15604531 1-INCH CERAMIC MICROPHONE SET Uses same cartridge as

1560-P5 and -P6 Microphones above.

5 Hz to 12.5 kHz -60 dB 22 to 145 dB 1,0. Ceramic Microphone Set 1560-9531

15604530 WINCH CERAMIC MICROPHONE Set Includes 1972 micro-

phone above.
/
2
1 -
in. Ceramic Microphone Set 1560-9530
5 Hz to 20 kHz -80 dB 42 to 165 dB

15609532 % ARCM CONDENSER MICROPHONE SET, flat perpendicular

response.
20 Hz to 40 kHz -60 dB 31 to 145 dB Vo -
In. Condenser Microphone Set 1560-9532
•
1580-9533 IA- INCH CONDENSER MICROPHONE SET, flat randorn.incidence
response.
20 Hz to 20 kHz - 60 dB 31 to 145 dB /
2
1 -
in. Condenser M,crophone Set 1560.9533

1560.9534 IA- INCH CONDENSER MICROPHONE SET, flat perpendicular Ve.m. Condenser Microphone Set
response.
1560-9534
50 1-1. to 100 kHz - 74 dB 49 to 158 dB

1560-9535 IA-INCH CONDENSER MICROPHONE SET, flat random- incidence te.in. Condenser Microphone Set
response.
50 Hz to 70 kHz - 80 dB 55 to 165 dB 1560-9535

15609536 St-INCH CONDENSER MICROPHONE SET, flat random- incidence . in. Cond EEEEE Microphone Set
response.
70 Hz to 140 kHz -93 dB 69 to 178 dB 1560.9536

•SenSitivity in the microphone- preamplifier combination is given for XI ( unity. gain) setting of the 1560.P42 Preamplifier, sensitivity à increased by 20 dB by use
of the ItIO setting of the 1560-P42.
"A-weighted noise level to maximum rms sinewave signal without clipping.

313
MICROPHONES
Catalog
Number

ELECTRET CONDENSER MICROPHONES (

tentative specifications)

1961 1- INCH ELECTRETCONDENSER MICROPHONES

Frequency: Curve shows typical response and guaranteed limits: individual response
curve supplied with each microphone. Microphone is essentially omnidirectional
SeeNitivity: - 56 dB nominal, re I Webar. TEMPERATURE COEFFICIENT .
no 03 clBrC born 0 to 055° C at IkHz UPPER SOUND-PRESSURE LEVEL 160
dB absolute max
Inspedence: 100: 10 pF at 26'C and 1kHz. < :02 pFeC temperaturecoeff relent Oa Wee al[gr.
hla 1151C150.015.
at 1kHz
A .00 203 5 en 90 se
Environmental: - 25 to •55° C and 90% RH operating. 1- year exposure in an envi-
ronment of 055° C and 90% RH causes negligible sensi twity change Arm.
Mechenieral: TERMINALS Coaxial with 0907-60 thread DIMENSIONS 0 936 e 1961 Eleccret Condenser Microphones
0 001 in dia ix 0.855 0.001 in long 123 77 e 0 025 21 72 0025 non/ Flat random incidence response. 1- inch . . 1911-9601
WEIGHT 1oz 1289) net. 1lb 10 5kg/ shipping Flat perpendicular- incidence response. 1- inch . 1961-9602

161211-INCH ELECTRETCONDENSER MICROPHONES

Frequency: Curve shows typical response and guaranteed limit. individual response
corvesupplied wit heach microphone Microphone is essentially omnidirectional
Sensitiony: - 62 dB norninal, re I Weber TEMPERATURE COEFFICIENT xx II ... ----------- .----...--- e
..
,
. t?...r:Si-
s —
.0 03 dBrC Iron, 0 to 055° C at 1 kHz UPPER SOUND PRESSURE LEVEL . -- ,••••• --------------- t —,--
170 dB absolute max f

Impedence: 35 e5pF at 25° C and 1kHz. < 00 1pErC temperaturecoellicient al 1

kHz
1s —
p., mom tLEZTIKT 951.0.001r1 7 - I

ro so ice 205 900 It 5

CralifffligléAtd: -25 to . 55° C and 90% RH operating. 1•year exposure in an envi- reseutve•.
ronment of + 55* C and90% RH causes negligible sensitivity change rom.•••••••......... ne.....m............ MI.
PAeclunical: TERMINALS Coaxial with 04660 thread DIMENSIONS 0 500 e 1962 Electron Condeneer Microphones
00005 in dia a 0682: 0001 in. long 112 70 : 00127 a 17 32 e 0025 rnm) Flat random incidence response. I/r inch 1962.9601
WEIGHT 05 oz 114 gl net. I lb 10 5 kg) shipping Flat perpendicular incidence response. Ye inch 1962-9602

1963 X.INCH ELECTRET CONDENSE R MICROPHONE

Frequency: Curve shows typical 'usu.'s, aria a...ran/ix/a limits. individual response
curve supplied with each microphone Response is essentially omnidirectional
Sensitivity: - 76 dB nominal, re 1 Wear TEMPERATURE COEFFICIENT .
50 03 clBr C from 0 to : 55"C at 1kHz UPPER SOUND- PRESSURE LE VEL 170
dB absolute max.
Irispedence: 12e IpF at 25° C and 1kHz. < : 005 ter C minperaturecoefficient at 0..993 CAC/WV 10.500..
IEH? «MAN
09 i05 .0 SOO
Environmental: - 25 to 05ec and 90% RH operating. 1- year exposure in an env, .9150...••A
roomer, of .- 55° C and 90% RH causes negligible sensmvity change
Mocheniod: TERMINALS Coamal with 022460 thread DIMENSIONS 0.250
00005 in dia o 0 355 e 0001 in. long 635e 00127 x 9.02 10025 mm) 1963 Electnet Condoned Microphone
WEIGHT 0.25 oz17 g) net. 1Ib10 5 kg/ shipping Fiat perpendicular Incidence response 4 inch 11163-11602

ELECTRET CONDENSER -MICROPHONE SETS ( tentative specifications)

The following electret Condenser microphone sets include, microphone cartridge. a

1560-P42 Preamplifier. all adaptors necessary to mate the car tridge to the prearnoh•
fief and to a 1562 Sound- Level Calibrator . and acarrying case for all components

Fiessuency Sendthrity• Dynentie Rome° .

nominai nominal, re IV/pbar typical. re 20 uN/m )

11NCH ELECTRET CONDENSER- MICROPHONE SETS

Electret Condenear.Mkrophone Sets
5Hz to 15 kHz - 56 dB 18 to 140 dB Flat perpendicular incidence response, I inch
5Hz to 12 kHz - 56 dB 18 to 140 dB Flat random- incidence response. 1 inch

RANCH ELECTRET CONDENSER MICROPHONE SETS

Electret Condoner-Microphone Sets
5 Hz to 24 kHz 63 dB 29 to 150dB Flat perpendicular- incidence response, Winch
5 Hz to 16 kHz -- 63 dB 29Io 150dB Flat random- incidence response. Tr inOh

RANCH ELECTRET CONDENSER MICROPHONE SET

Elect.« Conelemer.MicroMme Set
5 Hi to 35 kHz -78 dB 49 to 160 08 Flat pexpendiculat - incidence response, % inch

•Sensitivity is for Ow microphone preamplifier corm:Men:in end ri given for th• X Iluni tv - gain/ Salting of the 1560- ha? Pream011fler.sen•ifixitY if
chimed by 20dB ltiv use of fluf % 10 outing of Ow 1560.P42.
• • AIx/eight. noise level to maximum rms sinewava signal without clipping.

Prices on topmast. met f08 Concord, KIA. USA, subiect to quantity chscOunt.

314
 PREAMPLIFIERS
Catalog
Number

,REAMPLIFIERS

Impedance: INPUT, 6 pF; xx500 MO at low audio frequencies, driven

1560.1742 PREAMPLIFIER shield reduces input-capacitance loading for condenser microphones.
OUTPUT. xx15 Si in series with 3.3 µF.
For condenser and ceramic microphones and vibration pickups Output SIGNAL, up to 12 V pk.pk to 10 kHz into open circuit with 15-V
supply, decreasing to 2 V 9k-9k for 1:1 gain and 1 V p6- ph for 10 ,1 gain
at 100 kHz. Up to 10-mA rms output with 15604'62 Power Supply.
The 1560-P42 Preamplifier is a high- input impedance, lowtoiSe Preampli - POLARIZING VOLTAGE + 200 V t 5% behind 300-M0 dc source resist.
fier. It is particularly well suited for amplification of the output of capaci- ance, on-off slide-switch Controlled; temperature coefficient 0.1%/"C:
tive sources, such as condenser and ceramic microphones and piezo- frequency >50 kHz.
electric vibration pickups, and for use with GR sound- level meters and Noise: < 3.5-µV equivalent input with 390.pF source capacitance, C.
analyzers when a long cable must be used between the microphone and the weighted, 10-kHz effective bandwidth.
instrument. It is also a useful probe amplifier for other electrical signals Distortion: <0.25% harmonic distortion at 1 kHz with 1-V rms open-
where high input impedance and low noise are necessary. For esmPle, circuit output; < 1% at 10 kHz with 1-V rms output into 0.1 tiF ( equivalent
It can increase the sensitivity and input impedance of analyzers, recorders. to 2000 ft of cable).
amplifiers, null detectors, counters, frequency meters, voltmeters, and low. Terminals: INPUT. 0.460 x 60 thread for direct connection to 1 /-
2 in, micro-
frequency oscilloscopes. phones and adaptors. ACCEPTS INSERT CALIBRATION SIGNAL. 10 0 ±
20% insert resistance, <0.54B nominal loss between connector and
It is a three-stage amplifier with a low-noise FET input stage, a class AB
output stage, and full dc feedback for stability. Switched ac feedback microphone terminals. 1-V rms max insert voltage. OUTPUT, 10-ft
selects X1 or )( 10 gain and a self-contained oscillator supplies the polariz- cable with 3- pin 63 mike connector, separate ground and shield reduce
ing voltage for condenser microphones. This oscillator operates at a su- sensitivity to interference.
personic frequency to reduce interference and can be switched oft when Available: Microphone sets that include 1560-P42, vibration pickups, tri-
the preamplifier is used with ceramic microphones. Output from the pre- pod, cables and adaptors.
amplifier is through an attached 3-wire shielded cable and the required dc Power: + 15 to + 25 Vdc, 1 to 2 mA idling ( 200 V off) or 3 to 5 mA idling
supply voltage is applied froto one of the wires to ground. (200 V on). Available directly from 1558, 1568. 1564, 1909. 1911, 1912,
1913, 1921, 1922, or 1923 Analyzers, 1525 Recorder, 1561 Sound-Level
Meter, 1934 Walsh-Healey Noise Exposure Meter, 1566 Multichannel
Gain: 1:1 or 10:1 ( 20 dB 1 - ±-0.3 dB at 25C, slide-switch controlled; Amplifier, on from 1560-P62 power supply when preamplifier is to be
<±-0.3-dB gain change, from that at 25 .C, from - 30 to +65'C. used with 1565 or 1551 Sound-Level Meter, 1553 Vibration Meter, and
Frequency ( at 1-V rms open-circuit output behind 600 0, - 30 to + 55•C): 1900 or 1910 Analyzer.
3Hz 5Hz 20 Hz 100 kHz 300 kHz 500 kHz sgeetianiaat DIMENSIONS ( less cable): 6.5 in. ( 165 mm) long X 0.5 in.

8 -
I-
(13 min) dia. WEIGHT (with cable): 1 lb 10.5 kg) net, 3 lb ( 1.4 kg)
f„ I 13 1.548%-1- -15M I ±2 d shipping.

ISI100606
MRS MON ISSN RSIR ••• NI•Nio•

Iha, Cgs,
IORITEISSN RUB

I/2. seR0
11141(111331or 413E1

-1 1
-

ION 1660 IS? RS PTE.

1560.P47 Preamplifier 15 ,119642

-
1560-P40 PREAMPLIFIER
For ceramic microphones and vibration pickups Gain: Same as 1560-P42 except temperature range is - 50 to + 50'C.
Impedance: INPUT, 6 pF, >500 MO at low audio frequencies. OUTPUT,
The 1560-P40 Preamplifier is a high input- impedance, low noise pream- O in series with 3.3 riF at 1:1 gain, 100 U in series with 3.3 µF at
plifier similar to the 1560-P42 Preamplifier above except it produces no
polarizing voltage and therefore cannot be used with condenser micro- 10:1 PAK
Noise: < 2.5.µV equivalent input with 400-pF source capacitance. C.
phones. weighted, 10.kHz effective bandwidth.
It is a three.stage negative- feedback amplifier that makes full use of the Distortion: <0.25% harmonic distortion at audio frequencies with 1 V
tow- noise and high-input.impedance characteristics of an FET. The feed- pk-pk open-circuit output; 1% at 1 kHz with 5 V ph- 9k into 0.01 aF
back can be switched to obtain 1,1 or 10,1 voltage gain and a GR 1- inch (equivalent to 200 ft of cable), 1% at 10 kHz with 2 V Pk-isk into 0.01 eiF•
ceramic microphone cartridge plugs directly onto the input end of the Aveilable: Ceramic microphones, vibration pickups, tripod, cables, and
case. Output from the preamplifier is through a 3-terminal shielded adaptors. 1560.1.96 adaptor converts input to accept 3-pin mike con-
connector, and the required dc supply voltage is applied from one of nectors such as those used on 1560.P95 or 1560-9665 through -9667
these terminals to ground. cables.
Power: + 15 to + 25 V dc, 1 to 2 mA. Available from same sources as
Frequency: 3 to 5 Hz 20 Hz 250,kHz 500 kHz 1560-P42.
Mechanical: DIMENSIONS: 6.88 in. ( 175 mm) long x 1.56 in. ( 30 rb01)
Wizr,'.",„ .
t3 dB } ?. 4'ÓI3 1±0.25 d
±e .
s50 1
_,.5 dB
, dia. WEIGHT: 0.6 lb (0.3 kg) net, 3 lb ( 1.4 kg) shipping.

IIN c01....0 NIKES

ice 1560 P6. f6. 95701

IRO
101666110191.1050.5
ice 151/6452. TU. 956
1510•96[]-
60691.0“ 1560-P40 Preamplifier 1560-9640
<•BLE5.7.131.19,
56C900.1.01.4. CONNECTORS
1560-P96 Adaptor, to mike connector 1560-9696
IGR . 5[0 9665.6 7 SI

315
 TRANSDUCER ACCESSORIES

Cata log
Number

POWER SUPPLY

1560-P62 POWER SUPPLY Required with 1560.P40 or - P42 Preamplifiers charges. RIPPLE. < 5 mV rrns in CHARGE-OPERATE mode. CHARGE TIME.
when they are used with instruments that do not include a source of power 14 to 16 h for completely discharged battery, constant 22—mA battery.
such as the 1551 and 1565 Sound-Level Meters or 1900 and 1910 Analyzers. charging current. Rear-panel slide switch selects internal or external
Also useful when long cables are to be driven at high levels and as a battery.
charger for rechargeable batteries in the 1561 Sound-Level Meter or 1952 Interlace: INPUT, power to, and signal from, preamplifier made by Switch.
Universal Filter. craft type 83 3-pin microphone connector. OUTPUT, signal from pream-
A single front-panel control selects operating mode, OFF, CHARGE ONLY, plifier and remote power control made by SwItchcraft type Al 3- pin micro.
CHARGE ANO OPERATE. OPERATE ONLY, REMOTE ( off or operate-only mode phone connector.
selected remotely by instrument such as 1561 or 1564 analyzer', and BAT- Supplied) 1560-9665 4-ft cable to connect to 1551, 1561, 1564, etc; 1560--
9668 4-It adaptor cable to connect to 1900, 1910, etc, and cable to con-
TERY CHECK. The batteries are easily removed by a slide-out clip and fit nect to 1561 charging terminals.
into the same type of holder used in the 1952 Universal Filter.
Remote Operation: With line voltage not connected, preamplfler can be set
to OPERATE-ONLY mode by signal of + 15 to 25 V at 300 KA.
Input 100 to 125 or 200 to 250 V, 50 to 60 Hz. Environmental: TEMPERATURE, — 15 to + 50 .0 operating.
Mechanical: Convertible-Bench cabinet. DIMENSIONS ( w x h X di: Beret),
Output 18 to 21 V dc, 15 mA max; automatic limiting protects supply and 8.5 x 3.84 x 5.5 in. ( 216 x 98 x 140 mm); rack, 19 x 304 X 6.02 in.
prevents deep battery discharge. BATTERIES, two rechargeable Ni-Cd bat- (483 x 98 x153 mm). WEIGHT: Bench, 3 lb ( 1.4 kg) net, 5 lb ( 2.3 kg)
teries provide up to 225 met.hours operation at room temperature between shipping; rack, 5.5 lb ( 2.5 kg) net. 8 lb ( 3.7 kg) shipping.

40.0101%0 '6600668
'04 MX 9 N e/F05
IME061,110E C•XtE

F
1M60 9665 6
040 St 565 SO NO E el
',MEE
ME110.11566 OuIRES 1560
Inemerele
40 ce •90 61000001

1.0%66 6 E
.
to 40 106a Ibel 1. 1922 1923 ANAE Yee HS.
See KO Ot E E
461ILE
19» NOISE EePOSEERE Me Ten

1560-1.62 Power Supply, Bench Model 1560-9575

480 Rack.Adepter Set 0480-9742

MICROPHONE WINDSCREENS
These microphone windscreens reduce the effects of ambient wind
noise and protect the microphone diaphragm in oily, misty, or dusty
:e41014,00, emaweE
environments. They attach easily to any 1-Inch ceramic or condenser
microphone and do not appreciably alter the sensitivity or frequency
response of the microphone. The windscreens are made of reticulated — 1

polyurethane foam and can be conveniently removed, washed, and re-

placed if they become soiled.
Wind-Noise Reduction: 20 dB In winds ollO mph.
Microphone Sensitivity Loos( 0 dB to 3 kHz, (40.5 dB to 5 kHz, w.2 dB to
12 kHz, see curve. Microphone Windscreens, 4 each per pack
Windscreens are also available for 1 2 -
/ inch microphones. Their specifica- For 1- in microphones 1560-9521
tions are similar to those for 1- inch microphones. For Va- in microphones 1560-9522

TRIPOD
15604580 TRIPOD This Is a versatile tripod designed to accept a variety
of equipment. A 1 4 -20 threaded stud fits all GR sound- level meters and
/
electronic stroboscopes, a 1- in. sleeve accepts the 1560-P40 Preamplifier,
and a 45- in, sleeve accepts the 1560-P42 Preamplifier. Also included are a
locking nut, wrench, and spare setscrews.

a
Tripod 1560.9580

CABLES AND ADAPTORS

CABLE Shielded 3-wire cable terminated in Switchcraft type A3 3- pin
microphone connectors, for use between preamplifier output and analy-
zer. Mates directly to input and output connectors of 1560-P62 Power
Supply, 1558, 1564. or 1568 analyzers, 1525 recorder, or 1561 Sound-
Level Meter.

Cable, a It 1560-9665

Cable, 15 ft 1560-9666

Cable, 100 It 1560-9667

1560.9561 COUPLER ADAPTOR SET

to 1562 Sound-Level Calibrator.
Adapts 45, WI, and Vrin, microphones
II Coupler Adaptor Set 1560-9561

1560-9581 ADAPTOR Adapts GR 1-in, ceramic microphone cartridge to

the 1560-P42 Preamplifier.

1560.9649 ADAPTOR Adapts 1560-P42 Preamplifier input to Switchcraft

or Adaptor 1560-9581

type A3 3-pin microphone connector.

Adaptor 1560-9669

1560-9542 ADAPTOR Provides proper mechanical and acoustical cou-

pling between a standard 1- inch condenser microphone, such as the
Western Electric 640AA, and the 1560-P42 Preamplifier.
Adaptor 1560-9542

316
 VIBRATION PICKUPS AND SYSTEMS

• accessories for sound- level meters

• select for:
high-frequency performance The 1560-Pli8 vibration Pickup System
with the 1551.0 Sound- Level Meter.
high sensitivity
general application, economy

For the measurement of solid- borne vibrations with the and probe tips are provided. A permanent-magnet mount
sound- level meter a vibration pickup is used in place of is also available.
the microphone.
Each of these Vibration Pickup Systems consists of a TYPE 1560-P13
vibration pickup, a control box, and a connection cable. For measurements at higher frequencies than the -PIIB
The vibration pickup is an inertia-operated, ceramic device, system affords, the -P13 combination is recommended,
which generates a voltage proportional to the acceleration consisting of the 1560-P53 Vibration Pickup and the 1560-
of the vibrating body. By means of integrating networks P23 Control Box. A small holding magnet is included.
in the control box, voltages proportional to velocity and This system with the Type 1551-C or - 13 Sound- Level
displacement can also be delivered to the sound- level Meter provides the flat frequency response and low-noise
meter. The desired response is selected by means of a operation required by MIL-STD- 740 ( SHIPS) for vibration
three-position switch on the control box. Conversion data measurement. (The holding magnet is not used for meas-
are supplied for translating the decibel indications of urements according to that standard.)
the sound- level meter into the vibration parameters of
displacement, velocity, and acceleration.
TYPE 1560 -P14
TYPE 1560 -P11B The vibration pickup used in this system has approxi-
This system uses a lead-zirconate.titanate pickup, iden- mately 10 times the sensitivity and 10 times the imped-
tical with that used on the 1553-A Vibration Meter. Probe ance of the 1560-P52.
General Purpose High Frequency High Sensitivity
Pickup Systems 1560-P118 1580-P13 1560-P14
Vibration Pickup System Vibration Pickup System Vibration Pickup System

Ranges at 0.3 tO 390,000 ( 1000 g)t 0.01 to 3900 ( 10 01

0.1 to 39,000 ( 100 de

Rms velocity (in./s) 0.001 to • 0.001 to 1000 0.0001 to •

Rats DisplaceMent (
in.) 0.00003 to • 0.00003 to 30 0.000003 to •

Frequency Range

Response characteristics for

constant applied (1) accel-
eration, (
2) velocity, and (
3)
displacement.

Net Weight of System ( lb) 4 (

13
/ 00 XII) ita (0111 kg) 2(1 kg)
Shipping Weight ( lb) (2.3 kin 5(
2.3 IMO (03 ITS)
Catalog Number 1560-9922 1560-9613 1560-9614

Pickup Characteristics

Pickup Type Number 1560P52 1560.1•53 1560- PM

Sensitivity ( mV/g), nominal 70 70 700

Temp Ceelf of Sens ( d11/n) <-0.01 <0.02 0.01

Resonant Frequency ( HS) 3200 27,000 5000

Capacita nce ( PM 10,000 350 700

Temperature Range ( n) — 18 to 100 —54 to 177 — 18 to 120

Relative Humidity Range (%) 0to 100 0to 100 0to 100

Cable Length (ft) 5(1.55 m) 8 (2.5 m) 8(2.5m)

Dimensions (in.) 13
/
4 X 1N. X X. W ( hex) X 0.7 19). (dia) X W.
(mm) 42 X 37 X 15 15.5X 18 31 x 27
Net Weight ( eg) 1.6 (45 grams) 1.1 (31 grams) 3.1 (
40 WORD)
Catalog Number 15609652 1560-9653 1560-9654

•Upper limit Of displacement and velocity measurements depends upon frequency and is determined by the maximum acceleration possible before nonlinearity occurs (
100 g
for l560-Plig, 10 gfor 1560-P14).
re= acceleration of grad?),

317
 Index

A, standard tone
3 Anechoic
"A" weighted ( sound levels)
conversion to from octave bands 77 defined
256
defined Rooms ( Chambers)
8 155,191
limitations Annoyance (see also loudness)
44-45 considerations of
to rate noise
42-45 ANSI Standards 20
typical values
4 270
Absorption ( Sound) Anti-aliasing Filters
Acceleration 162-163 80-81
Anti- Friction Bearing Manufacturers Association . 273
defined Applications-Chapters 5, 7, and 8
level 12,13,256
Architectural Acoustics
12 13 ARI 162-167
measurement ( See also Vibration, measurement
273
techniques) 118-120 Articulation Index
37
Accelerometer ( See also Vibration measurement ASA Standards ( see ANSI Standards)
techniques) 112-114,119-120 ASHRAE
ASTM 153,272
Acoustical Materials 162-164,218-222 162 164 271-272
Acoustic Power ( see Power Level) Association of Home Appliance Manufacturers 273
Acoustical Calibration ( see Calibration( Atmospheric Pressure
3
Acoustical Properties of Rooms ( see also Atmospheric Pressure, effects of 187-188,189a02
"Anechoic" and " reverberant") Attenuating Structures 163-1 64,166-167,218-222
Audiogram ( and Audiometer)
effects of 177-1 78,192-196,201-203
how to measure defined
162-166 64,257
ADC uses
273 64 68-69
AFBMA Audiometer and Audiometry 63-69,257
273 calibration
Age- Hearing variations with 18-19
66,136-137
AGMA background noise, effects of 66-67
273
Air Conditioning and Refrigeration Institute 273 Audiometric Examination Room 67-68
Air Diffusion Council Autocorrelation
273 81-83,91,257
AMCA Autospectrum
273 Average 91 ,97,99-100,105,257
Aircraft Noise
154 161-162 •84-85,96-101
Alias Averaging, of meter readings 101
256
Aliasing
79-80,104 "B" weighted ( see " A" weighted)
Aliasmg Filters
80-81 Background Noise
Ambient Noise ( see Background Noise) 179-180
Baffle
Air Moving and Conditioning Association 273
Defined
American Academy of Opthalmology and 257
in use
Otolaryngology ( AA00) 60 219-222
Balancing of Rotating Machinery 160,226
American Gear Manufacturers Association 273
American National Standards Institute 270 Band Levels ( see Octave, 1/3 Octave, etc.)
Bandwidth
American Society for Testing and Materials 162-164,271-272 73-76,95-97
American Society of Heating, Refrigerating, Barometric Pressure ( see Atmospheric Pressure)
Beckesy Technique
and Air-Conditioning Engineers 153,272 66
Bell-shaped Curve
Amplitude Distribution 85-89,256 86-88
Bending Vibration
Amplitude Domain ( see Analysis)
Bits 10
Analog- to- Digital Converter 78-79,91
Block 78
Analog vs Digital
78 105-106 Blurring Effect 81
Analysis ( see also Octave, One-Third
Buffered Mode 95-96
Octave, Narrow, Noise, EFT)
103
amplitude domain 73,74,85-89
in bands of frequency 73-77,89-96 "C" weighted ( Sound Level) (see " A" weighted)
frequency domain 73-77,89-106 effect of, on Octave Band Readings 77,185
Cables
time domain 114
73-74,81-85 effects of long
Analyzers ( See also Analysis) 173
CAGI
1/3 -Octave- Band 153,273
73,75 ,
76,90,96,122-125, Calibration
defined 153 ,
15 6,158,160,162-165
183-185
Acoustical Chapter 7 1 36-137,183-184
FFT56 2 At high freque nc i es

126-127 184
Impact-noise Electrical ( circuit) 183
Narrow Band 131,159 Instruments for
125-127 Audiometer
Octave- Band 74-75,101 ,
12 1,152-155,158-159 136-137
Microphone
One-percent bandwidth 76,125 136
real-time Sound- Level
serial 103 123 126-127 Vibration 136
90,94,103 Vibration 137
Vibration
137 213-214
Wave 122-123 Catalog Section
125-126 278

318
 Excess Level 96
Cathode- Ray Oscilloscope
Exposure, Noise ( See Noise Exposure)
described 134-135
used 166 Far Field ( see Sound Fields)
Fast Fourier Transform ( see FFT)
Ceramic Microphones ( see Microphones)
FFT 91-93,95,97,104,1 26427 ,
159 .
259
Chart Recorder
Fast Meter Characteristics 181-182
described 128-129
Federal Coal Mine Health and Safety Act 160
in use, Chapter 7
Fields, Sound (see Sound Fields)
Circuit Noise, effects of 180
CNEL (Community Noise Equivalent Level) 162 Filter
Continuously adjustable 132-133
Coherence 105,257
defined 259
Comparison Method of measuring power 203-204
Components 77 81,92 Filtering, direct 89-90
Flash-Delay Unit 138
Compressed Air and Gas Institute 153,273
Flat response 7,8
Computers ( see also Systems) 140-148
Flexural Vibration 10
Condenser Microphones ( see Microphones)
Fluctuating Sounds 102-103
Confidence Limits 97-98,257
for Octave bands 101 Flutter
measurement of 236
Constant Bandwidth (see Narrow-band) Folding 80-81,259
Contours, Sound Pressure ( see Sound Pressure Level)
Fourier Transform (see FFT)
Control of Noise & Vibration, Chap 9 217ff Frame 81 91-92,259
Converter, Analog-to- Digital 78-79,91
period 91
Convolution 83-84 size 81.259
Correction for Frequency Response 183,184
Free Field ( see Sound Fields)
Correlation 81-83,104
Frequency
Coupler (see Earphones)
Analysis in bands of (see Analysis)
Criteria (see individual listings: Noise, etc.)
defined 3,259
Critical Bandwidth 4041 range of hearing 17-19,21-22
Critical Ratio ao response testing 128-129
Critical Speed 226 shifting 89 94-95
defined 257 translation 94-95
Cross Correlation 81-83,104,257 Frequency Distribution ( see Amplitude Distribution)
Cross Spectrum 104-105,257
Frequency Response
Cycles per Second ( now " Hertz") 1/3- and 1/10-Octave 121
defined 3 Microphones 110-112,173 -176
"D" weighted levels 4243 Narrow-Band Analyzers 93,95.125
Data Window 257 Octave 121
Dead room 257 Sound- Level Meters 8
Decay Rate Vibration Meters 120
defined 202,264 Vibration Pickups 120 207-212
in reverberant room 202-204
measurements 164-165 defined 113-114,259
Gaussian 86-88,259
Decibel ( dB)
conversion tables 244-251 Graphic Level Recorder
described 128-129
dB ( A) see " A" weighted
defined 4-5 257-258 in use, Chapter 7
how to combine 9.252 Hanning 93,98,260
Definitions 256-268 Harmonic Analysis ( see Analysis)
Degrees of Freedom 97-98,258 Harmonic Index 8-9
Detection Theory 16 Harmonic Motion 9-11
Differential Sensitivity 41 Hearing
Diffuse Field ( see also Reverb. Room) Damage Risk Levels 59-62
defined 192 frequency range of 17-19,21-22
Digital Filtering 90 loss, defined 260
Direct Filtering 89-90 from noise exposure. Chapt 4, App VIII 160-161,275
Directional effects microphone 173-176 measurements to prevent, Chapter 4 . . . . 160-161
Directional Gain Presbycusis 18-19
defined 201,258 Threshold shifts 58-60
Directional Source in free field 189-190 Masking 36-40
Directivity ( of Sound) other characteristics 4042
Factor 191 201 258 Pitch and Mels 4142
Discomfort Threshold 18 Threshold of 15 17-20,260
Disk Storage 148 Hearing Conservation Chapter 4,60-66,68-69
Displacement Hearing Monitoring (see Audiometer)
defined 11,258 Hearing Threshold Level ( HTL) 64-65,260
measurement ( See also Vibration measurement Helmet 62-63
techniques 119 Hemispherical Measurements. 198-200
Distribution ( See Amplitude Distribution) Hertz ( formerly cycles per second)
Double Amplitude 11 defined 3
Duration, Effects on Annoyance 33-34,46 Heterodyning 89 94-95
Dynamic Range 79 103-104 High Frequency Noise 171
E Weighted Levels 42-44 High Sound Level 171,181
Ear 57 Histogram 85-86,89,260
Ear Muffs 62-63 HTL ( see Hearing Threshold Level)
Earphones 258 Hum Pickup
calibration of 66,136437 Instruments 180-181
use of 186 microphones 173
Ear Plugs 62-63 Humidity, effects of on
Ear Protection 62-63 microphones 171-172
Echoes, Testing for 166 Hybrid System 106
Effective Bandwidth 95 Hydrophone 112
Effective Perceived Noise Level ( EPNL) . 34,124-125 IEC 271
Effective Sound Pressure 259 IEEE 272
Electret Microphones Ill Imaginary Part 91
Electrical circuit calibration 183 Impact
Environmental Noise, Section 7.3 160-162 defined 260
Equal Loudness Contours 8,21-23

319
Impact Noise Analyzer 13i Motor Vehicle Noise 161
Impulse Noise 131,159 Mounting
Incidence device under test 186-187,213
grazing 111 of vibration pickups 207-212
perpendicular 111 Multichannel Amplifier 115-116
random 111 "N" Weighted Levels 42-43
Indoor Noise Levels 38 - 39 Narrow- Band Analysis
Institute of Electrical and Electronic Constant Bandwidth 76 91-95
Engineers 272 constant percentage bandwidth 74-77,90
Instrumentation examples of 225,233-234,236,238,240
Specific Measurements ( Chap. 7 and 8) NC Curves 34-35
Listings and Descriptions ( see Table of Contents.) National Electrical Manufacturers
Instrument Society of America 272 Association 273
Interfacing 141,142 National Fluid Power Association 274
Interference 80-81 National Machine Tool Builders Association 274
Internal Noise 103 Near Field ( see Sound Fields)
International Electrotechnical Commission 271 NEMA 273
International Organization for Standardization 271 Newtons per square meter 3
ISO 271 NNI 35,262
Isolation 217-223,227-229,240 Noise ( see also Noise Level)
defined 260,267 Analysis ( see Analysis)
Jerk uses for 73ff
defined 12,260 entire Chapter 5and 7
Journals, list of 274 and Number Index ( NNI) 35,262
Kilohertz 3 background 179-180
Leakage 92-94 control, Chapter 9 , Chapter 10
Legislation ( See also OSHA, definition 262
Walsh-Healey Act) 275-277 effects on hearing, Chapter 4
Level ( see individual listings: Loudness, effects on work output 4647
Noise, Power, Reference, etc.) non-auditory effects 47
defined 260 "Pink" 95 132 263
Level Distribution 89,260 Random 22 , 86-88,95-103,131-132
Level Regulator 134 rating 42-45
Line Component 77,260 (see also Loudness, Perceived Level,
Live Room ( see Reverberant Room) 261 Perceived Noise Level, Speech Inter-
defined 261 ference Level)
Loudness and Loudness Level 20-27 reduction 156-157
Calculations 24-33 and entire Chapter 9 and 10
by Stevens Method 24-27,28-33 regulations 60-62,69-70
by Zwicker Method 27 41 Single Number Ratings for 4245
contours of equal 8,21-23 stationary 87
defined 261 white 95 132 268
loudness index 24-27
Noise Exposure Meter 117-118,160-161
measuring of 156
Noise Exposure Permissible 60-62
sones 23 24-27
Noise Legislation ( See also OSHA
Tables 25-26,253
Walsh- Healey Act) 275-277
Loudspeaker Noise Levels
defined 261 combining ( chan for) 252
Low Frequency Noise 171 Criteria for
Low Sound Levels 171,180 damage risk, Chapter 4
Machinery Maintenance ( see also Preventive hearing loss from Noise, Chapter 4
Maintenance) 48,49-51,229-231 indoor 38-39
Criteria 50-51 office 38-39
measurements for 49-51 residential 4547,161-162
Magnetic Digital Storage 148 speech interference 36-40
Magnetic Tape Recorders ( see Tape Recorder) telephone use 38-39
Maintenance ( see Machinery Maintenance) definition 262
Mark VI 24-27 "Perceived" 27-33
Mark VII 28-33 Prediction of 2(. .M-205
Masking 36-41 Noise and Number Index ( NNI) 35
defined 261 Noise Pollution Level 35
Materials, acoustical 162-164,218-222 Normal Distribution 86-88,259
Mean-Square 82 87 NOYS ( see also Perceived Noise Level) 27-28
Measurement Points for Acoustic Power 198-200 definition 262
Mechanical Impedance Nyquist Frequency 78
defined 261 Nyquist Interval 78,87,262
Mechanical Shock Observer, Effects of , on Measurements 177-178
defined 261 Occupational Noise Exposure 60-62
Mels 41-42,261 Occupational Safety and Health Act ( see OSHA)
Meter Pointer Behavior 100-103,181-182 (see also all of Chapter 4)
Microbar (see also Newton/in' ) Octave ( Band)
defined 3,261 Analysis 74-75,101,152-156,158-159
Micronewton/meter' to control noise 219-222,235-241
defined 3 to rate product noise 152-154
Microphone analyzers 74-75,101,121,152-156,158-159
ceramic 110-111,171-176 confidence limits of 101
choice and use of 171-176 conversion to " A" weighted levels 77
condenser Ill 171-176 conversion to spectrum level- 76
defined 261 defined 74,262
electret 111 preferred series of 74
precautions 171-176 Office Noise-rating chart for 39
sensitivity 113 Omnidirectional 111,173
wind screen 115,173-174 One-Third Octave ( see Third-Octave)
Microphonics, effects of 181 Oscillation
Minimum Audible Field ( MAF) 17-19 defined 262

320
 î
Recorder ( see Tape and Graphic Level)
Oscilloscope Reference Levels
described 134-135 Sound 4,5,7,13
uses for 166 vibration 12f
OSHA ( Occupational Safety and Health Act) . 60-62,69-70,160
Reference Signal (see Calibration)
"Overall" ( see Sound Pressure Level) Residential Noise Levels 45-46,161
Pain ( see Threshold of Pain) Resolution 91
Paper Tape 147-148 Resonance ( defined) 264
Parallel Filtering 89-90 effects of 178 212-213,241
Peak to Peak Value 213,226 Resonance Frequency 264
defined 262
Response ( see also Frequency)
described 11 20"kHz" 7
Perceived Level 28-33 defined 264
Perceived Magnitude 28-33
flat 7
Perceived Noise Level ( PNL) testing 165-166
calculations 27-28
Reverberant
defined 27-28,262
Field 194
measurement of 154
Periodic Quantity 263 Room
measurements in 175-176,201-204
Peripheral Equipment 147-148
Reverberation Time ( see also Decay Rate) 204
Permanent Threshold Shift ( PTS) 59-60,69
defined 264
Phon ( see also Loudness Level) 22-23
measurements of 164-165
defined 263 RMS ( ems) 82,87,213,264
Photoelectric Pickoff 138
ROC ( Receiver-Operating Characteristic) 16
Pickups ( see also Vibration measurement
Room Constant 193-195
techniques) 112-114,119-12 0 Room, effects on measurement . . . . 178,192-195,201 -204
mounting 207-212
Root- Mean-Square ( see RMS) 272-273
"Pink" Noise 95 132 263
SAE 272-273
Pitch 41-42,263
Sample (see also Sampling) 78,81,102
"PNdB" ( see Perceived Noise Level) Sampling 78-81,102,264
Point Source ( see Simple Sound Source) Scanner 116
Position of Microphone (see also Measurement
Sensitivity, Transducer 113-114
Points) 170,176 Serial Filtering 89-90,94
Power Level ( acoustic) Shakers, vibration 135
Calculation of 201-204
Shear Vibration 10
corrections for Temp and Pressure 189,202
SI weighted level 42 44
defined 5,263
Sigma (a) ( see Standard Deviation) 87,266
in free field 189-191
SIL ( see Speech-Interference Level)
Measurement of 153 195-204
Simple Sound Source 188,265
Comparison Method 203-204
Simple Tone 265
Standard Source for 203-204
Simulated Free Field 155,191
reference 4,5,7,13
Slow Meter Characteristics 182
typical 6
Society of Automotive Engineers 272-273
Power Spectrum 91 Sociocusis 18,265
Preamplifier Chapter 7,114-116
Software 125 142-147
Precision Sound- Level Meter ( see Sound Level Meter)
Sone 23-24,265
Predicting Noise Levels 204-205
Pre-employment Screening 68 Sonics
defined 265
Preferred Frequencies 74-75
Presbycusis (see also Hearing) 18-19 Sound
absorption 162-163
defined 263
attenuation in air 193
Pressure
defined 3,265
corrections for Power Level and
reflected 192-196
Sound-Pressure Level 189,202
effects of Atmospheric 187-188,189,202 speed of 3 13
summary of terms 13
level, sound (see Sound- Pressure Level)
transmission loss 163-164
reference ( also µbar and µWm' ) 4,7,13
Sound Fields 188-205
standard atmospheric 3
diffuse ( see Reverberant)
Pressure Spectrum Level directivity in 191 195 201
defined 263
free field 188-191,259
Preventive Maintenance 11 49-51,158-160,229 -2 31
directional source in 189-190
(see also Machinery Maintenance)
power level in 188
Primitive Period simulated 155,191
defined 263
near field and far field 190
Privacy 40
point source in 188-189
Probability Density Distribution 87,25E
reverberant ( see Reverberant)
Product Noise, Chapter 9 152-159
in rooms 192-196
Production-Line Testing 155-156
PSIL 37-40,263 Sound Intensity
defined 265
Psychoacoust ics 15
Psychophysical Methods 15-17 Sound Level
"A" weighted 8 42-45,152
PTS ( see Permanent Threshold Shift)
typical 4
Punched Cards 148
calibrator 136-137,183 -184
Pure Tone defined 8,265
defined 265 how to measure, Chapter 8
Q meter 7,8
directivity factor 191,201 weighting networks for 7f 42-43
Quantization 78-79,104,263 Sound-Power Level ( see Power Level)
Radiated Power (see Power Level) Sound- Pressure Level
Random Incidence 111 contours 189-190
Random Motions 10 corrections for temp and pressure . 187-188,189,2 02
Random Noise (see also Noise) 22 182-183,264 defined 7,265
generator of 131-132 for directional source 191
Rate of Decay (see Decay Rate) how to measure, Chapter 8,7
Real Part 91 overall, " 20 kc," " flat" 7
Real Time 103 123 126-127 reference ( levels) 7 13
Record of Measurements 205-206

321
sound Source in Reverberant Room . . 178,192-195,201-205 Transducer ( see also microphones and vibration pickups)
Spectrum defined 267
defined 266 Transducer Sensitivity 113-114
Spectrum Analysis ( see Analysis) Transfer Function 105,267
Spectrum Averaging 96-97 Transient Vibration 267
Spectrum Level 76,266 Translational Vibration 10
corrections for 76 Tripod 115
Speech Interference Level ( SIL) ( PSIL) Truncation 92-93
defined 36-37,263 ITS (See Temporary Threshold Shift)
table 38-39 20-kHz Response ( See Sound Pressure Level) 7
telephone usability 38 Ultrasonics
Speed defined 267
of processing 103 Uniformly distributed points ( measuring
of sound 3 13 radiated power) 198-200
SPL ( see Sound Pressure Level) USA Standards ( see ANSI Standards)
Standard Atmospheric Pressure (see Pressure) Variance 87,267
Standard Deviation 87,266 Vector 91
Standard Source (see Power Level) Velocity
Standards, Sound and Vibration 270-274 defined 11-12,267
Standards, Sound-level Meter 7-8,185 level 12-13
Standing Waves ( see Sound, Reflected) measurement ( See also Vibration, measurement
defined 266 techn iques) 119
Stationary 87,266 Vibration ( see also: accekration,displacement, jerk, and velocity)
Statistical Stability 91,96-102
Analyzers for, Chapter 5and 7
Stethoscope 186
background 208,213
Stroboscopes 137-140,160
Subjective comparisons of Noise ( see calibration and calibrator 137 213-214
Psycoacoustics) Classes of 10
Summation Analysis 84-85,98-99 conversion charts for 254-255
Superposition 83-84 control 222-231,236,240-241
Systems 140-148,156 defined 9,267
effect of pickup 212-213
Tape Loops 94
effects on man 47-49
Tape Recorder
intermittent 10
described 129-131
isolation 217-223,227-229,260,267
in use, Chapter 7
measurement techniques 206-215
Telephone Influence Factor ( TIF)
meter 118-119,268
defined 267
nature of motion 10
Telephone Usability 38-39
nonsinusoidal 12f
Telety pewriter 147
pickups 112-114,119-120
Temperature effects on microphones 172-173
reduction, Chapter 9 157
Temporary Threshold Shift (TTS) 59-60,69
shakers 135
Test Codes, noise
sinusoidal 11-12
listing 272-274
specifications (criteria) 48-51
measurements for 152-157
what it does 47-51
Third-Octave ( see also Analyzers)
to Tan 4749
Analysis . . . 73,75,76,90,96,153,156,158,160,162-165
to machine 49-51
corrections for spectrum level 76
other effects 51
preferred series of bands 75
Walsh- Healy Act
Threshold
regulations under 60 69-70,160
of hearing 17-18,266
Wave Analysis ( see Analysis, frequency domain)
of pain 18
Wave Analyzers 125-126
shifts of ( see Hearing)
Wave form 73 79-83,268
Tickle
averaging 84-85,98,268
defined 267
Wavelengths ( see also speed and frequency)
TIF ( see Telephone Influence Factor)
of sound in air 190
Time-Compression 94
Weighted Sound Levels ( see Sound Levels)
Time Domain ( see Analysis)
Weighting Networks
Time-Series 78 90-91,267
in sound-level meters 78 42 43,268
Time-Series Analysis ( sec Analysis)
White Noise 95 132 268
Tolerance, of Noise ( see "Threshold")
Williams-Steiger ( see OSHA)
Tone
Wind Effects 173-174
defined 267
Window 91-93,257
Tone Burst
banning 93,98,260
generator 133-134
raised cosine (see banning window)
for echo testing 166
Wind Screen 115 173-174
Tone Corrections 33
Words used to describe Sounds 269
Torsional Vibration 10
Work Output, effects of noise on 4647

322