Intensity Measurements

in Building Acoustics

' * ? * * * * 4 * * * t t t t * t t I * I 1

. I t

Briiel & Kjaer

BO 0147-11

Intensity Measurements in Building Acoustics

by Torben G. Nielsen, Briiel & Kjaer

Introduction
The procedure outlined in the ISO
140 standards for measurement of
sound insulation rests on relationships
between incident and transmitted
sound power (Ref.[l]). The sound pow
ers are estimated from measurements
of spatially averaged sound pressures,
and it is assumed that the sound fields
are either diffuse or free. Sound inten
sity is a measure of sound power per
2
m , so with an Intensity Analyzer it is
possible to measure the sound power
directly. The intensity method has
some inherent advantages compared
to traditional methods:

Contributions from various flank

ing paths can be quantified

Individual
contributions
from
parts of composite elements may
be determined

Sound leaks can be traced

The method uses a non contacting

transducer
Fig. L Sound Intensity Analyzer Type 4433 being used to measure the sound power emitted

In this note, the classical method of

measuring apparent sound reduction
index (transmission loss) is first re
viewed briefly. The intensity method
is then outlined, and in the following
section the battery operated intensity
analyzer and its probe are described.
After an outline of the general meaurement procedure the last two sections give detads of m-situ measurements m buildings carried out with
tlZl% \ TTT
^ T
?
erences [2, 3, 4, 5, 6] more information
on Sound Intensity measurements in
building acoustics are found.
2

from

wail

Apparent

Sound

Reduction

The Classical Approach

Apparent sound reduction index
apparent transmission loss) is defined in terms of the difference beW 6 e n th 6 P

Wer inddent

n t h e Parti

!
,

"
tion m the transmitting room and the
total power transmitted into the receiving room (Fig. 2). If it is assumed

Index

that the sound fields in the source and

receiver room are diffuse and that the
power entering the receiving room is
absorbed by the absorption area A in
t h e reCeivin

room

then

'
^ index can
be expressed in terms of the difference
between the averaged sound pressure
levels in the two rooms. A correction is

made for the absorption area A in the

receiving room. The procedure accord
ing to ISO 140 is to measure the sound
pressure levels in both rooms, using a
rotating microphone boom for exam
ple to provide the spatial averaging.
The absorption area of the receiving
room is determined by measurement
of the reverberation time T. The ap
parent sound reduction index can be
measured in the field to check on insu
lation specifications and work prac
tices. In the event of the insulation
specifications not being met, it is use
ful to identify the faulty building com
ponents; this is however not an easy
procedure using the traditional meth
ods.
The I n t e n s i t y A p p r o a c h
In the intensity approach (Fig. 3)
the sound power incident on the parti
tion on the transmitting side is mea. . . . . J o v o / l t i T 7 j_u covv,^
,
o^
*
sureu
in
exactly
m
e
s
a
m
e
way
as
m
,,,>,
i i i

i
the classical method, by measuring the
average sound pressure in the trans
mitting room. The power transmitted
into the receiving room is however
measured directly using a sound in
tensity analyzing system. Measure
ment of reverberation time T is not
necessary, and one does not have to
rely on a diffuse field assumption in
the receiving room. The intensity ana
lyzer measures the net
sound
2
power/m . The sound power emitted
from a given surface is therefore the
average sound intensity measured
over the surface, multiplied by the
surface area. In this way the partial
contributions of power injected into
the room from the different bound
aries (walls, floor, ceiling) may be de
termined. It is also possible to mea
sure contributions from windows,
doors, etc. Sound leaks reveal them
selves as spots with high levels of in
tensity. All contributions may be add
ed up to give an apparent sound re
duction index that can be compared
with the result of a classical measuremen
t-

Tig. 2. The classical method

of
measuring
sound
insulation
is
based
on
pressure
level

,,
,
.
,,
,,
,
rn,
r. , ,
measurements in the transmitting room and the receiving room. 1 he sound fields in
h()th

woms

should

be

diffuse

Fig. 3. Sound insulation measured using the intensity technique. The sound field in the
transmitting room should be diffuse, but this is not necessary in the receiving room,
nor is it desirable

Instrumentation
The Sound Intensity Analyzer Type
4433 is ideal for use in on-site building
acoustics investigations. The 4433
weighs less than 6 kg and runs for
more than 7 hours continuously on its
internal batteries. Its small size
(138 mm x 251 mm x 300 mm) allows
it to be brought right to the measurement site even when space is restricted.

The analyzer allows measurement

of pressure, particle velocity and intensity to be done in octaves from
63 Hz to 8 kHz as well as broadband
measurements (linear and A-weighted). It is also possible to A-weight the
octave measurements directly. Automatic scanning of the filters and setting of the input and output amplifiers makes the instrument easy to use.

Stored spectra may be transferred to

external equipment via the built-in serial and IEEE interfaces.
The analyzer is designed to be used
with a probe consisting of two phase
matched microphones. For measurements at low and medium frequencies
half-inch matched microphone pair
Type 4177 or Type 4183 can be used.
3

The distance between the two micro

phones in a pair may be changed to
accommodate different parts of the
frequency range. Details on this are
found in the data sheet for the 4433
analyzer and the probe 3520 (Refs. [7,
8], and in Appendix A. The Portable
Sound Intensity Analyzer Type 4433
is shown along with the Intensity
Probe Type 3520 in Fig. 4.

General Measurement
Procedure
Whereas the classical measurement
of Sound Reduction Index or Trans
mission loss only allows one spectrum
representing all the different trans
mission paths to be determined, the
intensity method makes possible a
quantification of the individual trans
mission paths that contribute to the
sound field in the receiving room. The
transmission through party walls and
flanking walls are measured separate-

Fi

S- 4. Sound Intensity

Analyzer

Type 4433 and Probe Type 3-520

ly thus allowing an evaluation of the

relative importance of the transmis
sion paths. Also, in facade insulation
measurements, doors, windows, win
dow frames, ventilating units etc. can
be measured separately.
Measurement Surface
Common to all these measurements
is the determination of the sound
power radiated from a surface. The
sound power is found by measuring
the average intensity normal to a mea
surement surface enclosing the radiat
ing surface and then multiplying this
average intensity Iau by the area of the
measurement surface.
In Fig. 5 the procedure is illustrated
with the measurement of the sound
power radiated from a window in a
measurement of facade insulation,
where two possible measurement surfaces (Si and S2) are shown. The
choice of surface is determined from
practical considerations. Si is obviously the simplest surface to measure
on since it consists of only one plane,
whereas S 2 consist of 5 planes. On the
other hand, since it is rather close to
the window the sound field may vary
considerably with position making the
determination of the average intensity
normal to the surface difficult. In
practice, the measurement distance is
determined in a preliminary investigation where the probe is swept along
the window at different distances with
4

Fig_ 5

Possibie

measurement surfaces for determining sound power transmitted through a

window

its axis normal to the window. Small

changes in intensity levels indicate a
suitable distance. Typically a distance
of 10 - 20 cm will be adequate. In
some cases, where the measurement
surface parallel to the window is much
larger than the other four surfaces and
the sound energy is believed to propagate mainly perpendicular to the wall,
these four areas may be omitted.
The average intensity is determined
by first subdividing the measurement
surface into a number of areas and
then measuring the normal intensity
level within each area by holding the
probe in the middle of the area or by

slowly sweeping the probe as if painting the area. Choice of area size and
probe technique will depend on how
much the sound field varies with position along the wall and how detailed
information is required. The fixed
point technique has high repeatability
whereas the sweeping approach is
faster, and inaccuracies due to nonsteady probe motion can be minimized
by selection of a manageable area size.
As shown later, an area of approxi2
mately 1 m gives almost identical resuits with point and sweep measurements of the sound power radiated
from a concrete party wall.

C h e c k on A c c u r a c y
The repeatability may be checked
by comparison of a number of "identical" measurements at one point or
over one area. Just as in measurements of sound pressure an increase in
the averaging time will improve repeatability. A good averaging time to
start with is 8 sec.

tion with area S the equations reduce

to:

A possible bias error may be

checked by comparing measurements
where the probe has been turned 180.
The results should be the same but
with opposite sign (opposite direction). If the results show a difference
of e.g. 2 dB the measurement has a
bias error of 1 dB (Ref. [9]). Other bias
errors in the intensity estimates are
caused by two factors: The absorption
of the radiating surface and the reverberant field in the receiving room. The
absorption coefficient a of the radiating surface should be low and the reverberation time T should be kept
small ( T < 0,5 s) in order to facilitate
the measurements. If T is too high
initially, it may be reduced by placement of absorbent material in the
middle of the receiving room. Details
on these precautions are found in Appendix B.

Measurements of sound insulation in

two different houses will be described.
The first set of measurements was
done in a two storey building belonging to the Building Research Establishment in Watford, England. Measurements of sound insulation were
made using both the classical method
and the intensity method so that a
comparison could be made. The other
measurements were carried out on a
party wall and an adjoining flanking
wall in a newly built two storey apartment house in Denmark.

C o m p u t a t i o n of A p p a r e n t S o u n d
Reduction Index
In the ISO standard ISO 140, part
IV, an apparent Sound Reduction Index R' is defined. It is called "apparent" because the equation for i?', as
shown in Fig. 2, defines the Sound Re
duction Index as if the whole trans
mission takes place through the party
wall.

R' = R'n = LPs - 6 dB - LIn

(3)

U8.SG o l U Q l G S

C a s e I: M e a s u r e m e n t of a p p a r ent Sound Reduction Index

A ground plan drawing of the building belonging to BRE, Watford, UK is
shown in Fig. 6. The party wall, consisting of 225 mm bricks with plaster
on both sides, extends up to the roof,
so no significant transmission was estimated to take place via the ceiling.
Neither the concrete floor nor the
backwall were likely to contribute very
much either so it was decided to measure only the party wall and the two
flanking walls.

Measurement Procedure
The significance of wall absorption
and Reactivity Index were first investigated. The absorption coefficient of
the walls was estimated to be around
0,01. The average reverberation time
in the receiving room with 3 persons
present was 1,4 sec. Expecting that
one quarter of the total power is emit
ted from each of the flanking walls, e n
(error due to absorption) and LK (reactivity Index) for these walls were
found to be - 0 , 5 dB and - 1 9 dB (using equations B l and B2 in Appendix
B). Wall absorption could then be neglected but it was necessary to introduce additional absorption in the
room to decrease the magnitude of LK.
From Fig. 10 it is seen that the
4433/Probe combination allows measurements with less than 1 dB error to
be done with L x > - 1 4 d B at 2 kHz.
Foam blocks were now placed in the
room and the average reverberation
time decreased to 0,5 sec and LK s
- 1 5 dB was found to be close enough
for a start. During the measurement
the foam was placed along the wall
behind the operator to efficiently provide more absorption. The Reactivity
Index LK was noted while measurements were being made and was found
to be - 8 to - 1 0 dB for the party wall,
and - 1 0 to - 1 3 dB for the flanking
walls.
The sound power passing through
the party wall was first determined.
The wall was divided up in 30 areas,

A similar equation may be set up for

the apparent sound reduction index
based on intensity measurements.
When all contributions are added the
result should be similar to the classical
result.
The partial apparent sound reduction
index R'n for wall n with area Sn is
given by the equation shown in Fig. 3:
ffn = L p f i - 6 d B - L / n + 101og(S/S n )

(1)

R' is now found by proper addition of

all the R'n:
N

12'= - 1 0 log E 10
n =\

10

(2)

In the laboratory, where there is

only transmission through the parti-

plg. 6. Ground plan drawing of buildi ng where measurements were made

5

0,3 m each, and the normal intensity

was measured at 30 points about
20 cm from the wall. The distance was
not critical and it turned out that
there was very little variation of the
intensity level along the surface, so
much less than 30 points could have
been used. The flanking walls were
then divided in 10 and 11 segments
respectively, and the segments were
laid out to follow the door and the
window. With segments of approxi
l
mately 1 m in size it was decided to
move the probe in a circle instead of
doing a point measurement. The level
in the receiving room was very low and
a true sweep measurement tended to
create too much background noise
from the operator. The frequency
range from 125 Hz to 250 Hz was mea
sured with a microphone spacing
Ar = 50 mm whereas Ar = 12 mm was
used for the rest of the frequency
range (Refs. [8, 9]).
D i s c u s s i o n of M e a s u r e m e n t
Results
T h e m e a s u r e m e n t results are s h o w n
, . ,_ T .
,
,
.
in r i g . /. it is seen, t h a t t h e r e is very

Fig. 7. Measurements of apparent sound reduction index R'.

"
!
Classical measurement. *-* Intensity measurement (party wall + 2 flanking walls) u
u intensity measurement, party wall. A
_1 Intensity measuremerit, flanking wall with window. -D Intensity measurement, flanking wall with

good agreement between the classical

measurement and the sum of the con
tributions from the party wall and the
two flanking walls from 250 Hz up to
4 kHz, as determined using the inten
sity method. In the bands around
250 Hz and 500 Hz the major contri
bution comes from the party wall
whereas the flanking walls are just as
important at higher frequencies. The
discrepancy between the two sets of
measurements in the 125 Hz octave
band is probably due to measurement
inaccuracy of the classical method.
The uncertainty is known to be about
2 dB at 125 Hz in this building.
Conclusion
The portable 4433 Sound Intensity
Analyzer has been used to measure
sound insulation between two rooms
in a house. The information obtained
about the relative importance of
flanking transmission, and the overall
apparent sound reduction index shows
very good agreement with results ob
tained by the classical method.
Acknowledgement
I would like to thank the staff at
BRE, acoustics department for their
assistance with the sound insulation
measurements.

door

Fig. 8. Sound intensity levels on party walls in dB re 7 pW m~

s o u n d Insulation Meas u r e m e n t on p a r t y w a l l and

flanking w a l l
The sound insulation measurements
were performed in two adjacent terraced houses. The party wall separating the two houses had an area of
2
14 m on the receiving room side, of
2
which only 10 m was common to both
the transmitting and receiving rooms.
The common area was made of
230 mm concrete. The remaining area,

C a s e

II:

made of lighter materials (wood clad

breezeblock) extended beyond the
outer wall of the transmitting house
and faced out into the garden. The
reverberation time in the empty receiving room was approximately
1,5 sec. Placement of absorptive bales
of Rockwool decreased the reverberation time to about 0,5 sec, which made
the measurement condition better by
decreasing the reactivity Index LK (see
also Appendix B).

Intensity Measurements
The sound intensity measurements
on the party wall were measured using
the sweep technique in 12 sub-areas
with an averaging time of 32 s (Fig. 8).
The sweep rate was about 0,5 m/s.
The intensity levels at that end of the
wall closest to the garden were marketedly higher in certain octaves. The
situation is detailed for the 1 kHz oc
tave band in Fig. 9.
The sound pressure level in the gar
den near the breeze block wall was too
low to generate significant airborne
sound transmission into the receiving
room, so the high intensity levels on
this part of the wall were due to flank
ing transmission.
S w e e p and Point M e a s u r e m e n t
lecnnique
Comparison was made between
sweep and point measurements of in
tensity over the party wall in 12 subareas using the portable octave Sound
Intensity Analyzer. The resulting
sound reduction indices are shown in
Table 1. The sweep speed was about
0,5 m/s.

Hz
^***^^^^^^ta^^^M*^^*^^^^ta^^********^**

125
125
250
250
500
500
1k
1k
2k
2k
A

R point
point
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^m^^^^^^^^^^^m

47,2
47,2
52,1
52,1
58,4
58,4
63,9
63,9
74,2
74,2
61,4

61 4

p^ g souna< transmission in the 1 kHz octave band (0 dB corresponds to 35 dB Sound

Intensity level

R sweep
i w ^ a ^ F ^ i f * n i i v ^ ^ ~ ^ V B B n i i ^ H ^

47,9
47,9
52,8
52,8
58,8
58,8
63,6
63,6
73,9
73,9
61,4

g-| 4

T01200GBO
T01200GB0

Table 1. Sound reduction indices, point

and sweep measurement techmques over 12 subareas on the
party wall

Conclusion
Using intensity measurements the
sound power injected into the receiving room from a party wall and an
adjacent flanking wall have been determined. It has been shown that the
2
power/m (the intensity) produced by
the flanking wall is higher than the
intensity produced by the party wall,
and the flanking wall is excited by
structure-borne transmission.
The portable intensity system is a
very convenient tool for in situ investigations in building acoustics. The
analyzer, being battery operated, is silent, which can be of crucial importance in measurement in well insulated houses where the sound level in the
receiving room can be very low.

Fig. 10. The measured Residual Intensity Index for the 4433/Probe for a microphone
spacing of 12 mm
A T^r-*G-nr1iv A
.rVJJjJtJilLllA r\
Measurement Accuracy: The rea c t i v i t y i n d e x LK and the r e s i d u al i n t e n s i t y i n d e x , L K 0
An intensity system's ability to
measure in sound fields is mainly determined by the phase mismatch between the two channels. This phase
mismatch is conveniently expressed as
the Residual Intensity
Index LK0
(Ref.[9]), which determines the lowest
intensity level which can be detected
by the system for a given sound pressure level. This is an important parameter when measuring sound transmission through walls, as very often
the intensity level which the system is
required to detect lies much lower
than the pressure level. The measured
Residual Intensity Indices LKt0 for the
4433 and the V2" microphone pair
used in the measurements are shown
in Fig. 10.

For an accuracy of l d B in the

measured intensity, the difference between the measured intensity and
pressure levels (termed Reactivity Index, LK) should be numerically 7 dB
smaller than LK0 (Ref.[9]). This defines the dynamic capability of a
sound intensity analysing system. For
example if LK{) for the analyzer is
- 2 0 dB, then for an accuracy of
1 dB, measurements can be made in
a sound field where the sound intensity level is no lower than 13 dB under
the
sound
pressure
level
(LK > LK>0 + 7 dB, i.e. LK > - 1 3 dB).
The residual intensity index LKt0 for
an intensity system may be determined from the calibration chart of
the probe and a simple measurement
of LK 0 of the analyzer.

Appendix B
M e a s u r e m e n t Accuracy: Influence of absorption coefficient a
of the radiating w a l l and influence of the r e v e r b e r a n t field
When using the intensity approach
to measure sound insulation between
two rooms it is desireable that the re
verberant sound field in the receiving
room should be as low as possible.
This is required for two reasons. The
first concerns the fact that the analyz
er will measure the net power coming
from the wall, that is the power emit
ted by the wall minus the power ab
sorbed by the wall from the reverber
ant sound field in the receiving room.
In these circumstances there is an un
derestimate of the emitted power. The
magnitude of this error, em can be
estimated using a simple formula,
Fig. 11 (Ref. [6]). If the error is unacceptably large, it can be reduced by
distributing absorptive material in the
receiving room to reduce the reverberant field.
The second reason for desiring a low
level reverberant field in the receiving
room is that a sound intensity analyz
er may have difficulty in detecting the
low intensity levels in the presence of
a high level reverberant sound field.
The Reactivity Index LK therefore
needs to be estimated or measured to
check that the dynamic capability of
the sound intensity analyzer is not exceeded. LK can be estimated using the
formula in Fig. 11 (see also Ref.[6]),
and it can be measured directly. The
magnitude of LK can be reduced if
necessary by introducing absorbing
materials into the room which act to
lower the reverberant sound field.

References
[1] ISO 140 "Measurement of sound
insulation in buildings and of
building elements ". Part IV: Field
Measurements of airborne sound
insulation between rooms

f(g

factors which couid affect the accuracy of sound intensity measurements: (a)
absorption of sound power at the wall from the reverberant field in the receiving
room, (b) Reactivity Index LK of the sound field (Intensity Level minus Pressure
Level). See also Ref. [6]

TwQ

[2] MACADAM, J.A. "The measuremerit of sound powers radiated by

individual room surfaces in lightweight buildings", Building Research Establishment Current Paper CP 33/74, 1974
[3] KIHLMAN, T. "Measurements of
th
sound radiation into rooms" 11
ICA, LYON, 1983
[4] COPS, A., MINTEN, M., "Comparative study between the sound
intensity method and the conventional two room method to calculate sound transmission loss of
wall constructions " Noise Control
Engineering Journal, May - June

1984

[5] CAUBERG, J.J.M., " Determination in situation of the transmission loss of various components of
a building facade with the aid of
sound intensity
measurements",
FASE Proceedings, 1984

[6] ROLAND, J., "Room to room

transmission: What is really measured by intensity" Proceedings of
nd
2 congress on acoustic intensity,
CETIM, Senlis, 1985
[7] Sound Intensity Analyzer Type
4433, Product Data
[8] Sound Intensity Probe Type 3520,
Product Data
[9] GADE, S., "Validity of Intensity
measurements",
1985
B &K
Technical Review, No. 4
[10] Sound Intensity, B & K booklet