[Translated from French by Mattia Valente, mvalente@xs4all.

nl, original document at

http://www.lutherie-guitare.org/]

Daniel Friedrich Emile Leipp

June 1977 N° 92

GAM

The Guitar
Function

Newsletter of the Musical Acoustics Group

Université PARIS VI Tour 66

4 place Jussieu
75005 Paris
PRINCIPLES OF CLASSICAL GUITAR FUNCTION

Two parts may be identified:

1)The excitation system – the strings
2)The amplifier and resonator system, consisting of the instrument's sound body

The strings provide a rich signal containing numerous frequencies that will be amplified by the
soundboard, assuming it is constructed properly, as well as by other parts of the sound body. All of
these pieces and parts have their own natural frequency and act as resonators. As they are secured,
attached, glued together, they act as coupled resonators.

The major challenge is amplifying all of the frequencies emitted by the strings, without emphasising
or restricting some of them, specifically without super resonance or holes in the instruments
frequency response, which must be even for all notes and strive for this goal.
If the soundboard amplifies one frequency, reinforced by the vibration of one of its constituent
parts, this is called a super resonance. The opposite is also possible.

It is therefore possible to modify the constituent parts, all of the pieces in order to obtain an even
response without damaging the qualities already acquired.
However, a new problem soon arises. The energy communicated by the thumb to the string may be
reconstituted by the guitar in three seconds or in six seconds, depending on construction. This
leaves us with two extreme choices:
1)Build a lightweight guitar, with a maximum weight of 1600 grams with a thin soundboard,
similar to a drum skin. This will lead to large amplitudes but swift decay of sound, in part
because the back and sides will also be very thin. This will deliver a strong bass and a
generally rounded, dull and brief sound with a prominent attack.
2)Build a heavier instrument, with thicker plates, with a more heavily braced soundboard
(properly reinforced) which resists, which “springs back”, which is not blocked. With a well
thought-out neck, the instrument as a whole will be balanced, allowing good oscillating
potential. This leads to long sustain, maximum power and excellent output.

It must be noted that no part of the guitar can be modified without changing some aspect of its
response, the guitar's sound, for better or for worse. Extremely careful observation is necessary in
order to advance. Finally, it is the builder's sole responsibility to make the choices. He must be
gifted with the ability to synthesise and a strong artistic sensibility.

There is no 'secret key' to conceiving of, to creating a guitar; there are thousands of things to know.
There are no miraculous 'tricks', only experience, a great deal of experience, from which
observations, teachings and solutions may be drawn.

Amongst the numerous problems faced by the luthier, some present realities that are difficult to
comprehend and partially evade scientific attempts to decipher them, placing them in the field of
artistic endeavours (which itself is broad enough). These experiments will focus on:
1. Power (from afar, close by)
2. Sustain
3. The evenness of the sound level
4. The timbre
5. The balance between treble and bass
6. Easy or difficult playability of the instrument
7. The homogeneity of sounds
8. The spontaneity and responsiveness of the instrument's response
9. Attack – both audible and inaudible
 10. Contrast (playing with a harpsichord or harp)
11. Sympathetic resonances – present or absent
12. Clarity or darkness of chords
All of these elements constitute the primary characteristics of the instrument, but we will only
examine some of them in detail in this article. Power and sustain in particular. In fact, the
investigator will also assess other elements in the course of his research into power and sustain.

Which component parts of the instrument should be prioritised for study?

If we experiment on multiple parts at the same time, interpretation of results will be unreliable and
difficult. In descending order of importance, we must examine:
1)The soundboard and its bracing; the choice of wood, thickness
2)The neck; the type of wood, thickness
3)The body; the sides, the back
4)The bridge; its volume and shape, arching/radius
5)Glues and finishes

The investigations presented below focus primarily on the system of 'Soundboard – Neck – Strings'.
Also described are an experiment examining the roles of the soundboard and the back, and finally a
historic perspective and the mechanics of guitar bridges.

SOUNDBOARD – NECK – STRINGS SYSTEM

Researching Power – testing a theory of functioning

(it is assumed that the majority of guitarists desire an instrument with good sustain with a less
explosive initial attack, but which retains body in the subsequent seconds). Upon release of the
string – let’s take the sixth string, Low E – after having created excess tension, meaning
simultaneous movement of both soundboard and neck, via finger pressure on the string, we can
observe a complex vibrational excitation of the instrument.
1)Direct excitation The soundboard begins to oscillate upon release, vibrating in response to the
string's excursions. The neck does the same, particularly the headstock (easy to verify by
holding it between two fingers) at the other extreme end of the string.
If the neck wood and heel are hard and rigid enough, and the back is flexible and lightly
braced, the back will also move with the pull of the heel (figure 25). Movement which
cannot exist without the presence of an arched, domed back.
The neck
The part of the soundboard (which is domed) behind the bridge is also subject to additional
traction with each vibration of the string (which collapse into a complex vibrational mode).
These vibrations are instantly transmitted to the sides and the back by the materials used,
which will reinforced certain frequency, leading to particular sound characteristics for a
series of notes.
2)Indirect excitation: air resonances
The vibrating air volume, created by soundboard displacement, will act on the back and
sides at their fundamental frequencies, and project sound waves through the soundhole.
All of these vibrational modes combine and superimpose themselves to yield complex
sounds with specific characteristics which the luthier tries to master.

Closer observation of the neck reveals three potential points of deflection/bending that are the site
of vibrational motion upon release of the string, combining instantaneously, fig. 26

Let us examine two extremes:

The neck may be constructed out of a hard, heavy and rigid wood. Or made of a flexible,
lightweight wood, or simply be too thin, or too thick. (A neck may be considered too thin if the
amount it bends under string tension, the relief, exceeds 4/10 of a mm. Past this value, phenomena
such as string rattle in the basses and 'snapping' in the trebles once past the sixth or seventh fret,
requiring raising of the saddle).

A heavy, rigid neck

It is easy to imagine, therefore, that a neck that is too heavy and too rigid will inhibit the movement
of the string upon its sudden release. It dominates the system and will prevent the swift
development of lower frequencies, for example. Bass notes will be dry and lacking in mellowness.
The neck acts as a filter. This neck, with its anarchistic response, will also inhibit the soundboard,
leading to difficult coupling, dulled response and finally, a loss of power.

A flexible, light or thin neck

The opposite situation, a very flexible or too thin neck, is equally problematic. When the finger
applying pressure to the string creates additional tension, the weakest, most flexible point will
move. In this specific case, the neck itself will pull forward rather than the soundboard, which will
only move a minimal amount; this is not the goal strived for in luthiery.

The notion that creating movement in the soundboard is the only important thing, ensuring string
action only manifests in the soundboard, is seductive. We see some luthiers use very robust necks
on their guitars, or add internal stiffeners to the neck. The solution to this type of problem is often a
compromise between the advantages and disadvantages.

The soundboard
The problem is complicated by the soundboard, situated at the other end of the string.
It too can be described conceptually in two extreme, differing ways
1)Soundboard that is too stiff. There is minimal movement, leading to a small sound which is
dry and articulated, with minimal amplitude and is difficult to play.
2)Soundboard that is too flexible. Large amplitudes favouring bass notes are obtained, with
strong second order harmonics, easy playability, mellow bass, but generally short sustain.

This oscillating system can be visualised and illustrated at a low cost, by placing a 5mm x 5mm
stick of spruce on two small dowels spaced forty cm apart (fig. 27). Attaching a 500g weight at the
centre the stick of spruce pulls it down slightly, and when released, a fairly rapid, sustained
movement may be observed (the weight represents the effect of the strings on the soundboard).

Soundboard and Neck

Next, by placing thinner stick (3mm thick) made of the same wood onto the dowels and repeating
the experiment, we can observe a much slower, shorter movement with greater amplitude which
closely mirrors the behaviour of a soundboard that is too flexible.

Combining Soundboards and Necks yields two extreme scenarios:

1)Soundboard and Neck are too stiff – only the string stretches, nothing moves voluntarily,
volume is pathetic, the sound is small, dry, with a poor, metallic timbre and lacking in bass
2)Soundboard and Neck are too flexible: A lot of bass, the guitar will have a soft touch (in a
word, mushy). The sound will be generally short and dull, and attack will be characterised
by string rattle and snap.

One could also combine a flexible neck and a stiff soundboard, or vice versa, which we have
already described above. As with all aspects of luthiery, this is an artistic choice representing the
builder's personality. The luthier's choices define his unique signature.
The quest for power – for the equilibrium point in the 'Soundboard – Neck – Strings' system – may
be facilitated by measuring the flexibility of the Neck and the Soundboard separately and
subsequently, with the instrument completed, repeating these measures with and without string
tension as outlined in the paragraph 'Controlling deflection'.
Comments: Luthiers of days past tensioned the strings of their instruments until they 'Rang out' and
functioned properly. Modern day demands appear greater; obtaining a sound with good sustain and
timbre, without weaknesses in any of the three and a half octaves requires in-depth research while
maintaining fixed tuning to A 440 Hz.
In order to achieve optimal tuning and balance for soundboard, neck and strings, an adjustable neck
may be employed, with an adjustable joint and two screws that rest on the soundboard. We
personally built such a design, but found the variation available to be lacking.
Second solution: adjust the soundboard after testing the instrument (only possible if bracing
has been designed with this approach in mind).
Third solution: the system can be optimised by changing string pull.

A final observation relating to the wood that has been used in high-quality classical guitar necks for
over a century: Spanish Cedar (Cedrella Odorata) appears to have a rare property in addition to its
stability and light weight. Its damping property (and sensitivity to impact) allows it to stop, to filter
certain vibrations coming from the headstock which could be a nuisance if they travelled all the way
to the soundbox, overlapping vibrating frequencies emitted from the bridge. This wood therefore
delivers a clean, neat, homogeneic character to the sound, as well as contributing to a spontaneous
sound due to its light weight.

WOOD SELECTION – TESTING

General considerations
Once a luthier achieves an excellent result following a string of successive attempts, a major
problem arises: “How to precisely reproduce this instrument in terms of Power, timbre,
homogeneity, touch, etc...” (This relates to luthiery of the highest level). Woods are most often
extremely variable, even within the same tree or within one board from said tree. There is a number
of different ways to address this issue
a) The optimist says to himself: “Chance will bring me pleasant surprises”
b)The careful builder will think: “I like to control as many aspects as possible, experiment in
order to improve areas in which my understanding is lacking, and be able to implement
diverse and desirable changes.”
c) Alternatively, one might say: “Using the knowledge I have gained, I will attempt to construct
homogenous guitars of very similar quality, with woods that often have very different
characteristics”.

Comments: The first experiment which comes to mind is obtaining soundboards that come from
the same board (difficult to realise, as most commercially available soundboards tend to be mixed
up). Over the course of several years, one notes that the soundboards rom the top of the tree trunk
are tighter, stiffer. During the next resawing session, one takes care to number the soundboards,
dividing the tree trunk into three or four sections of increasing rigidity and fineness of grain.
Despite these efforts, results remain uneven. The same observations can be made for necks, which
when taken from the opposite sides of a board (in width), can vary significantly in terms of both
weight and rigidity.

To move beyond a luthiery of chance, one must go further and systematically control and
weigh every part, choosing exactly the same species of wood which yielded an interesting
result.
Thankfully, guitar making, with its multiple thin plates of wood used to construct the guitar, is well
suited to tests and measures that are far more difficult for violin builders. We present a number of
quick observations on the resinous woods used for soundboards (Spruce in the example below)
before providing more precise ideas about deflection and calibration of various constituent parts, as
well as examining the deflections that the instrument may present during the course of its
construction and in its finished state.

EXAMINATION OF A PIECE OF RESINOUS WOOD

THE STRUCTURE OF SPRUCE

These woods consist primarily of axial cells (lengthwise) which act as minuscule 'juxtaposed pipes'
1.5 to 4 mm in length. They often account for 90% of the wood's total volume. There are also radial
cells (radiating from the centre to the outside), which create the medullary rays. There are two types
of axial cells (tracheids) (fig. 27)
1)Cells formed at the beginning of the growing season (earlywood), with thin walls, which are
elements that conduct sap
2)Cells formed late in the growing season (latewood), thicker walled and flattened, which are
the support elements and lend the wood rigidity and resistance. These close-packed cells
define the annual rings and form the grain.

External factors (the meristem) will greatly influence structure and determine the proportion of
earlywood and latewood cells. These factors include altitude, the tree's immediate environment, the
terrain, rainfall. We know that altitude (around 1,000 meters) often yields thin layered spruce, as the
growing season is short. This generally leads to a heavy, strong grained wood.

The cells in the latewood are the spruce's natural reinforcement in a longitudinal direction. Trees
that have experienced regular growth over the two to three centuries required to create a beautiful
tonewood are rare. Growth rings are irregular due to greater rainfall, or milder temperatures for one
or more years. Felling a neighbouring tree immediately provokes enlargement of cellular layers.
This is visible as areas of varying density across a board.

Along its radial axis, the tree has a multitude of medullary rays with a lenticular or extremely
flattened cross-section, creating the 'cross frame' for the wood, its 'solid' reinforcement (figure 28).
The medullary rays are not very tall, between one and three tenths of a millimetre.
They consist of cells with moderately thick walls. The rays, depending on their number and
thickness, appear to contribute to radial rigidity, ie 'cross grain stiffness' (1).
(1) When asked about this hypothesis, the Technical Wood Office replied it was 'very likely'

An important comment arises: if one takes an off-axis board of spruce, one will no longer encounter
whole medullary rays, but rather rays that have been broken, lending the board a certain lateral
'floppiness', a significant difference compared with a perfectly on-axis board taken from another
part of the tree.

This leads to the following deduction: when buying woods, the presence of fine chatoyance when
held up to the light indicates the presence of cleanly split medullary rays along with the 'fibres'
surrounding them, indicating the board was taken from the centre of the tree and will present
maximal 'cross grain' stiffness.

One notorious cause of irregularity in the mechanical properties of wood in general, and of spruce
in particular, is that the wood fibres often don't grow perfectly parallel and vertically, resulting in
constantly shifting directions of wood fibres throughout the board. In these cases, the wood fibre is
considered 'cut', leading to extremely diverse mechanical properties from one piece of wood to the
next.

For all of the reasons enumerated above, when one is selecting pieces of wood for bracing, for
example (from a single board), one may note that pieces taken from a particularly hard, resinous
part of the board will exhibit deflection numbers, mechanical properties, that vary by a factor of up
to three, meaning they can be up to three times as stiff as a piece taken from a softer area.

This lack of homogeneity poses a problem that can be resolved by controlling, measuring and
sometimes weighing almost every component, allowing the luthier to choose those that exhibit the
desired qualities.

These controls – deflection of components – will relate to their weight, and for soundboards to their
cross-grain compressibility. Compressibility that can be tested using a sample a few centimetres
wide over the entire width of the joined soundboard. Compressibility replaces cross-grain
deflection, a measurement which is always difficult to obtain.

A soundboard that is highly compressible is a supple piece of wood, with low cross-grain stiffness,
which will lead to a dulled response and a risk of rapid wood fatigue under string tension. Precise
measurement of this value is interesting.

Device for measuring compression.

The goal is to slide the cross-grain sample vertically into an case with walls that keep it squared
Next, a weight of 10 kg per square centimetre of sample is applied (more or less). The amount of
compression is read using a dial gauge which must be integrated into the measurement apparatus.
The soundboards will also be subjected to tap tuning by tapping the centre of gravity, potentially
followed by frequency analysis of this impact in a laboratory (described elsewhere in this
newsletter).

CONCEPTS OF DEFLECTION
and assorted tests

A board flexed within its limits is subject to constrained forces of compression for the outer face,
and traction for the inner face. These forces act most strongly on the surface fibres. We observe a
neutral fibre in the centre.
− The deformation, the curvature of the piece can be calculated
− We also encounter the notion of the modulus of elasticity, or Young's modulus. Simply put,
this represents the degree of rigidity for the wood, which will determine greater or lesser
easing.
When building the guitar, we encounter many potential deflections that can be measured with a
selection of simple devices created by the investigator.
Longitudinal deflection
I. Take one of the two halves that will constitute the soundboard, dimensioned uniformly to a
thickness of 3mm, for example, and place it in the device (fig. 30) so that the folding axis
will be situated precisely where the saddle would be. By applying a weight of about 500
grams at an adequate distance, the difference in deflection can be measured either by using
two metal rulers (one on each side, attached to a magnet, or using a dial gauge mounted to a
support.
We can call this longitudinal deflection
Combined deflection
II. Once the soundboard is jointed and glued a more complete measurement can be obtained,
providing a more realistic indication of flexibility or stiffness using a device (fig 31)
consisting of two 'external molds' of thick, rigid, non-deformable wood, with a cutout
 precisely matching the guitar's shape.
By placing these two 'molds' on top of one another and clamping them firmly in place (with
the soundboard between them), it is possible to apply a weight of about eight kilograms to
the bridge location, and measure deflection using a dial gauge The calliper will be attached
to a bridge on one of the two transverse slats, between which the 8 kg weight will be
inserted to rest on a bridge, positioned where it will be on the completed instrument. This
deflection figure represents the resultant of the combined transversal and longitudinal
rigidity, hence combined deflection.
III. After the soundboard, we turn our attention to the various bracing components (fig 32). The
deflection of each brace can be measured easily by placing it on two small dowels spaced
about 40 cm apart. A 500 gramm weight can be hung from the brace and a dial gauge can be
used to measure the maximum deflection (you will be surprised by the significant variations
between different pieces of wood).
IV. The same setup can be used to measure deflection for transverse braces, but deformation
will be so minimal, a gauge measuring tenths of a millimetre should be used.
V. Bridge deflection (the bridge is essentially a transverse brace) is also interesting. This allows
testing a batch of items, randomly varying dimensions, or varying them systematically, one
by one. Immobilise the bridge by clamping one of the wings firmly in a modified vise, and
hang a 1,000 gram weight from the end of the other wing. Deflection is measured using a
dial gauge measuring tenths of millimetres.
Bridges can also be evaluated for their tendency for torsion, an interesting notion if one tries
to allow the bass and treble sides of the guitar to work independently. A bridge that is too
thick, or made of wood that is too hard will only move as a single block, without any
flexibility.
VI. Deflection of the roughly shaped neck (fig. 33). This important measurement should be
made when a neck is rough cut, meaning rough shaped with headstock and heel block glued
in place. By clamping the neck by the spanish foot and placing a 10 kg weight on the
headstock face, deflection can be measured using a decent calliper The calliper can be
attached to a magnetic base.
The overall gross weight of the neck should also be recorded; a systematic record for all
woods used for this purpose should be created.
VII. Deflection of the ebony fretboard. The method to use is similar to the one described under I
for measuring soundboard deflection (longitudinal deflection). After immobilising one end
of the fingerboard, a 1,000 gram weight is placed on a defined reference point, and
deflection is measured using a metal ruler or a dial gauge It goes without saying that all
pieces of wood to be compared must be planed down to the same thickness.
VIII. Dimensional stability. In humid environments in tropical countries, or by the seaside, some
guitars seem 'muffled'. One may assume certain woods used for soundboards are particularly
hygroscopic. It is a well-known fact that certain extremely dense and resinous spruces resist
impregnation with conservation products through immersion, and their saturation point is
significantly lower than 28% for water – the average for this wood species.
To gain better insight into the effects of water, measure samples before and after submerging
them in water for ten minutes and until they reach maximum saturation. The investigator
will note extremely wide variations, and based on findings will be able to select appropriate
woods when building for Japan or Brazil
The system is simple: a sliding stop (like a sliding rule) is applied to the sample. A dual
magnet calliper is placed behind this stop.
IX. Deflection of the neck with and without string tension (fig 34)
Once the instrument has been completed, and before varnishing, the behaviour of the neck
when combined with the soundbox can be checked by placing the guitar in an apparatus
where it rests on three points, situated at the end of the headstock and at the two widest
points of the lower bout (the contact point the headstock rests on will need to be flexible in
 order not to falsify the measurement). It is then easy to place a 10 kg weight between the
12th and 19th frets, and measure displacement using a calliper between the 11th and 12th frets.
Once the guitar is complete, this operation can be repeated with the strings tuned to pitch,
and one will be surprised to note that some instruments display greater displacement with
the strings tuned to pitch, raising questions of why.
X. Soundboard deflection – strings tuned to pitch – strings slack
Without changing the guitar's position on the apparatus, following the previous
measurement, move on to the measurement concerning the soundboard is flexibility once
the guitar is complete. The 8 kg weight is inserted between the transverse slats (as described
above) and applies weight to the bridge, positioned temporarily before varnishing.
The operation will be repeated once the bridge is glued in place, and again with the strings
tensioned, allowing the luthier to observe the most significant movements when the strings
are tensioned on certain instruments. In general, the guitar will have the lightest touch if the
deflection of the top exceeds 2 millimetres and if the deflection of the neck also reaches the
figures found under the previous section (number IX in this section).

The wise luthier will learn from these different measurements. He will have to weigh the rough-
sawn and then dimensioned soundboards, as well as the necks, backs, fingerboards and sides in
order to build an instrument whose weight matches his wishes and plans.

Attack, the sound of fingers moving along the string is far less audible if the instrument is heavy.
However, there is also a limit beyond with responsiveness, spontaneity and suppleness of sound
find themselves altered (the margin is narrow).

At the end of the twentieth century, we do not feel it is not at all useless to spend two or three hours
taking measurements and performing tests (and recording and archiving the findings) considering
the overall build time of at least 100 hours required to create a quality guitar.

For the Artistically inclined luthier, this careful approach and understanding of these mechanical
elements will allow a greater flexibility and variety within his building; they are a weapon against
the 'industrial aspects of the guitar' which attempt to thwart change. It is not wrong to think that past
masters had already created a system of personal references that allowed them to secretly move
forward on sure footing. (We personally hope that these testing procedures constitute the most
significant technical contribution possible within the context of this essay).

A final note should be writ large:

It is impossible to build two comparable instruments without using the

same woods
SOUNDBOARDS AND BRACING

In the 17th and 18th centuries, guitar soundboards were reinforced with nothing more than a brace
placed at either side of the rosette (fig 35). In exceptional cases, imitating the lute, a light brace was
placed between the rosette and the bridge.

Towards the end of the 18th century, a number of experiments began appearing in France. Marchai
placed an angled brace between the rosette and the bridge (fig 36), but the breakthrough discovery
for the birth of the modern guitar appears to have been made in Andalusia, in Cadix : Bénédit in
1788 and Pages in 1792 were already employing a five brace fan (fig 37).

With this pattern, they discovered a revolutionary bracing system that would further differentiate
and separate the guitar's tonal character from that of the lute, giving it the character we are now
familiar with. Antonio de Torres (1817 – 1892) must have been familiar with the guitars built by
Pages, renowned in their time, and he accented the system by adding two further braces (fig 38),
and increasing the size from 35 to 36 cm for the maximum width of his guitars dated around 1860.

A little bit of research will show that this bracing pattern allows equivalent longitudinal rigidity and
solidity while maximising transversal flexibility with less thickness. All that is required is to glue
three braces on a softwood board forty centimetres long and 2.6 millimetres thick (fig. 39), and it
will become apparent that a board 5.2 millimetres thick is required to provide the same amount of
deflection, the same resistance, and will weigh 73 grams instead of 46 grams (1)
This significant weight gain will allow a more spontaneous response
− greater responsiveness
− a greater maximum amplitude (more bass)
− longer sustain, because internal friction is less significant in a thinner section of wood

Antonio de Torres's methods created a solid bass sound and a taste for the instrument that has lasted
for over a century. In recent years, a slight shift in the opposite direction may be noted. (It must be
remembered that the guitar has gained bass response since the end of the 16th century). All luthiers
are capable of building instruments with a generous bass response... the difficulty lies in building a
guitar with a bass, midrange and treble response of equal quality.

THE ROLES OF WELL-DESIGNED SOUNDBOARD BRACING

1)Allows the implementation of a large number of vibrational modes and the amplification of
many of the frequencies delivered by the rich string signal (the subdivision of the vibrating
surface into sectors and areas by a bracing pattern allows the creation of complex vibrational
modes not possible if the soundboard were a simple plank of wood)
2)Creating a pleasing, interesting, characteristic sound
3)Bracing must hold up mechanically for the long term, withstanding the violent effects of
torsion applied by the strings.
4)Must not be too rigid. This will quickly lead to a heavy touch and difficult playability. The
sound will be dry, devoid of smoothness, mellowness and bass.
5)Must not be too flexible (3): this will yield a large, short sound, bassy and dull with a very
easy touch, but strings that 'slap' under the fingers, with a mushy sound in rapid runs; the
instrument will easily be overdriven when played heavily.
6)Bracing must not lead to anarchistic responses within the system of 'Neck – Soundboard –
Strings'.
7)If necessary, minor touch-ups can be performed, to correct mistakes or make adjustments after
testing.
8)The bracing pattern may limit the appearance of shrinkage cracking. This is an important
function for reinforcing a piece of wood such as the soundboard, which is exposed to
tension caused by drying out or swelling due to humidity.

The importance of the soundboard bracing pattern takes a long time to understand. It is well-
established in terms of woods used and goal strived for. The best results are often found at the limits
of resistance for these woods, making the longevity of the system extremely precarious. Modern
guitars have limited solidity reserves in terms of bracing, and are not designed to last 100 years with
the same sound.
(1) The same results are obtained if the braces are part of the plank's mass (not glued)
(2) See the Savart memoirs collected in the Roret Encyclopaedia, recently re-edited. Title: “Le
Luthier” (The Luthier). The author speaks on experiences with vibrating plates.
(3) The deflection of the soundboard can be measured – see the chapter on deflection
Comments:
Woods used
It is not only the bracing layout that defines the quality of the sound emitted. The properties of the
soundboard wood itself, of the vibrating plate, are extremely important and will provide an initial
colour to the sound; a particular, specific quality.

Each piece of wood, even within the same species, has a property that must be discovered. It takes
skill to create a very warm, mellow sound full of character, but with enough darkness while
providing opacity in chords, with multiple superimposed notes.

Another tree may deliver different qualities; sustain, clarity, biting brilliance -characteristics that
will make the guitar more suitable for playing the work of Bach over that of Turina or Villa-Lobos.
The guitar that can do everything does not exist (one cannot ask for the qualities of both a 'brunette'
and a 'blonde' in a single instrument), and the luthier will have to either accept a compromise, of sell
multiple guitars to his wealthier clients...this is often the case for renowned guitarists. Using bracing
adjustments, these different qualities can be balanced and adjusted sensibly (while exhibiting
mastery, should one have the ambition).
In any case, we note that string vibration dissipates more or less swiftly for three reasons:
a) Opposition to the resistance of air to movement
b)Internal friction in the moving plate (which transforms the communicated energy into heat)
c) Internal friction in the string.

Two species of resinous softwoods are particularly interesting to luthiers:

1)European spruce (Picea Excelsa, Picea Abies), growing at an altitude of about 1,000 metres,
provides wood with a solid texture
2)Canadian Western Red Cedar (Thuja Plicata), growing in British Columbia, close to the
Pacific Ocean, is a softer wood, more fragile and sensitive to shocks (a wood with low
resilience) but which has the appealing quality of always being more flexible, more elastic
in the longitudinal direction than spruce. It is also a wood with low damping.

These two resinous softwoods, extremely variable from one tree to the next, can be:
1)Damped or not (this means when a board that is considered 'damped' is lightly tapped, the
sound of the tap is dull and rapidly dissipates, absorbed by the wood. If the wood is
resonant, 'non-damped', the sound or hock of the tap yields a long, clear sound).
2)Heavy or lightweight
3)Fibrous or clear grained
4)Strong textured (like iron) or weak (like copper)
5)The grain may be straight, parallel, or skew
6)The wood may be elastic or rigid

Let us look at what we can expect from a Damped, flexible and lightweight soundboard (soft and
supple). Without correction using bracing, it will be very difficult to build a guitar with a clear,
spontaneous, lively response with long sustain. The sound will be rather round and dull, dark with
short sustain. The guitar will be easy to play, but easy to overdrive. A heav and very hard
soundboard wood will easily yield a hard, dry, metallic sound, with weak, bassy harmonics. The
result will lack mellowness. The touch will be heavy, the sound quick and not very responsive.

The 'ideal' wood lies somewhere between these two extremes for most luthiers. We also note that
some resinous woods have 'fibres' that are firmly interlocked and extremely solid (like iron). After
ten years of use, we see minimal deformation, and the sound remains appealing and vigorous. Other
woods of the same species will have long left go, and the instrument no longer develops anything
other than bass tones or a worse timbre. The central part of the soundboard will have been
completely neutralized, broken down, 'brought to its knees'.

It is possible to gain an understanding of the wood's resistance to deformation by keeping a sample

board lightly tensioned at its centre, with the two ends resting on small slats. After a year, the
tension is release, and the board will straighten out – or not (fig 40). The other option is to build a
single guitar using this wood and observing what happens to it after one year.

Having wood fibres parallel to the thickness of the soundboard favour stability over time. If the
fibres are cut, if there is runout, the board will not present a great deal of longitudinal resistance,
and significant deformation will occur if uncorrected by bracing.

The difficulty for the builder who wants to be a creator is controlling all of the data we have
described, and progressively integrating this knowledge. The artistic design of an instrument is the
most delicate aspect, given the level of mastery required and the numerous contradictions implied.
It is a perpetual series of choices that can engender some anxiety. With each change, the instrument
improves as the builder's understanding of the woods he has increases...but there are still many
obstacles.

Here is an irritating contradiction: for soundboard, if one wants a rich sound with a strong
personality, a wood must be selected that is hard enough and stiff enough in both directions, which
will deliver well-defined resonant peaks. However, a damped wood with opposite characteristics
will deliver a homogenous response, an even sound level. A whole book would be necessary to
address the problems posed by bracing and soundboards, so to end this chapter, let us define a
golden rule:

“Obtaining a complex, rich sound with personality and character is impossible without a
'bracing' that is complex and personal and applied to a 'selected' top”

This is one of the secrets.

THE BRIDGE
The bridge transmits string vibrations to the soundboard. It is the final component to be glued
during the construction process. For over one hundred years, its shape and structure have been
refined to suit guitars played with fingers.

Historical bridges from the 16th , 17th and 18th centuries were similar to those of the lute – a simple
piece of wood, a small glued-down bar the strings were attached to, which should properly have
been called 'tie block', as the strings did not pass over a bridge, as was the case in the 13th century.
The modern bridge was developed following numerous experiments throughout Europe. A few
sketches will provide an idea of the main stages in its evolution (fig 41).

French luthier La Prévotte appears well-placed in this creative processes, as he was using a bridge
featuring modern characteristics before 1856, namely the implementation of two distinct parts for
its structure (fig 42)
Part A: a new addition, called the 'bridge' with its saddle
Part B: which can be called the 'tie block' to make things easy, which is the traditional
simple bar common to both lutes and guitars.

Part A will precisely define the end of the vibrating string length, and the height of the saddle will
insert itself under the string, raising it upwards, letting the string tension apply pressure and torsion
when combining parts A and B. One has only to combine the two parts on a single base to
understand the practical reality of their function.

One element becomes immediately apparent. The more closed angle CDE (fig 43) is, the greater the
torsion and the more vigorous the string pressure is on D. This will generate strong bass notes with
large amplitude, deep tone and good timbre if the soundboard is not too rigid. This was one of the
characteristics of guitars built in the 1850's; with their larger bodies, they had the novel ability to
deliver truly deep, robust bass notes. The luthier can adjust this angle if he decides to use a 'thick'
soundboard, which requires a brute force approach to work.

The modern bridge is made of rosewood, and weighs about 18 grams. Being a transverse brace
located more or less in the middle of the soundboard, if it is too rigid and too heavy – for example if
made of ebony – the sonic results with poor. The inertia and rigidity combined suppress a certain
independence between the bass and treble sides of the instrument, canceling out certain possible
vibrational modes, thereby cutting a series of frequencies and impoverishing the instrument's
timbre and evenness of sound.

In order to objectively determine the effects of a specific type of bridge, the apparatus depicted in
fig. 44 must be built. Using 10 kg of traction acting on the string via a weight, the deformation of
the board can be measured in front of and behind the bridge.
The modern classical bridge leads to greater distortion towards the back of the soundboard than
towards the front. In reality, this action may be more strongly emphasized if a transverse brace is
glued to the soundboard close to the front edge of the bridge.

In this case, the back of the soundboard is subjected to movements of a greater amplitude and worsk
harder than the front, because the bridge rests on the stiffer front part of the soundboard in order to
gently oscillate and pivot.

When the guitar is completed, this can be verified using two dial guages placed to either side of the
bridge. With the strings tensioned, the part of the soundboard doing the most work can be identified,
providing additional information aobut funciton (differences can vary with a factor of up to three or
more). Using the same apparatus described in Fig 44, it is also possible to measure and trace the
effects a less wide bride (17mm) with the same CDE angle and the same overall saddle height (in
effect, the actual bridge without the tie block seciton) has on deformation. Though the local effect is
stronger, the overall effect is the same, or slightly greater, as the force necessary to achieve similar
deformation in a rigid plank is greater for a smaller lever than for a normal bridge 30 mm wide. One
aspect compensates the other (1)

The only negative aspect of the narrow bridge is that it is at risk of delaminating if used. On the
other hand, its light weight may be interesting for the instruments response, and the establishment
of the higher frequencies generated by the string.

The modern bridge has certainly not reached the pinnacle of its evolution. Asymmetrical bridges
can be constructed, with a larger gluing area for the bass side, making it heavier and wider at the
bass side, theoretically allowing more solid basses, with a narrower, thinner part at the treble side
favouring the latter. Guitarists still need to be convinced to allow this modification.

In the United States, Kasha concieved of the use of seperate bridges for bass and treble strings, and
is building his guitars this way. This raises the question – with some reservations – of whether this
attentuates sympathetic frequencies, leading to difficulties in the midrange frequencies. The
evolution continues...
(1) Other viewpoints on the problems of bridge design may be found in Newsletter nr 72 on
Lutes, by Charles Besnainou.
THE BACK – THE SIDES

General considerations

The overall weight of an instrument can be said to influence its sound. If the guitar is heavy, the
sound of string attack (cause by finger action) is less easily perceived than on a guitar that is
extremely light; the sound is born differently, with a matte, extremely attenuated sound that seems
more crystalline. The guitarist's phrasing becomes more precise in rapid runs.

The heavier instrument has a longer sustain and generally has a better response in the upper register
– the overall effect is more brilliant. In order to achieve this end, modern luthiers have used heavy
woods, or thicker component parts for their instruments.

Where Jean Voboam was already using extremely exotic woods in 1706 – 'Violet wood' (Kingwood)
– very thin and very dense, we interpret this as being for decorative effect. However, when Pages in
Andalusia used rosewood for his backs in the late 18th century, one had to consider he may have
made certain observations concerning its tonal effects

Over time, this wood would come to be recognised as delivering the best results, and it is worth
noting that all attempts to use other woods were invariably abandoned. Antonio de Torres frequently
tried Maple in the last century, Francisco Simplicio loved Cuban Mahogany in the first third of this
century and more recently, Ignacio Fleta created a Flamed Maple guitar around 1958; woods we
have personally tried include Purpleheart, Ceylon Satinwood and Cuban Mahogany.

Whether Indian, Madagascar or Brazilian, rosewood confers a noble voice on the guitar, with a
pleasing, robust, full timbre that defines its quality, its personality as a classical guitar. These guitars
are significantly heavier now than they were 15 years ago, weighing between 1650 and 1850 grams.
Guitarists heavily influences by the lightweight creations of Antonio de Torres, Manuel Ramirez
and Francisco Simplicio found it difficult to recognise the qualities of a heavier instrument. It was
rejected, often compared to the 'old guitar' without which it never could have succeeded. This
mindset has completely disappeared, and concert guitarists playing guitars more than 10 years old
are currently extremely rare.

In terms of construction, two extremes may be identified:

1)Build with thin, lightly braced back and sides which will deliver a strong bass response,
mellowness, extremely audible attack, a short sound that is slightly dull and possibly a little
thick.
2)Build with a thicker back (heavier than 300 grams), brace more heavily, using heavy, thick
sides which will provide brilliance, a metallic timbre, with longer sustain and a more
moderate bass response.

The sides

The role of the sides is not negligible, as their vibrating surface is approximately equal to that of the
back. Additionally, their height defines the volume of the resonant air cavity of the guitar's
soundbox, potentially lending a certain 'boxiness' to the guitar. The fundamental frequency of the
sides is raised due to their curvature.

The experiment by Antonio de Torres, in which he built a guitar with cardboard sides (currently
conserved at the museum of the Barcelona Conservatory) may not have been interpreted fully by its
author, who apparently declared that “only the soundboard played an important role and the sides
served no purpose, as they could be made of cardboard”. It can be said that this instrument sounded
like a guitar made with more or less hard cardboard sides, thereby conditioning part of its response.

The back

Three well-defined functions can be attributed to the back:

1)The back and its bracing more or less vigorously stabilise the guitar's heel, preventing it from
tilting under string tension (40 to 50 kg).
2)The back braced with three or four transverse wood braces is also a multiple resonator for
numerous frequencies.
3)The back is also a sound reflector within the acoustic enclosure that is the instrument's
soundbox.

The thicker the back, the higher its fundamental frequencies and vibrational modes are. It will then
naturally favour trebles. The greater the number and height of the braces which divided it into
sectors, the more higher frequencies will be able to resonate.

Resonator

This is easily verified using a hollow glass tube roughly 25 cm long and 5 to 7 mm in diameter to
acts a frequency generator. Simply place the tube (fitted with a wooden shim) perpendicularly onto
the back of the guitar, gently holding the top of the tube between thumb and two fingers, like a
pencil. After first moistening those three fingers with a bit of vinegar and water, slide them along
the tube in a single, swift motion. Varying the amount of finger pressure will excite the 'vibrating
plate' of a defined sector of the back, emitting a tone specific to this sector. By changing the point of
contact with the back, numerous frequencies will appear.

If this experiment is repeated on a strung guitar, each sound, each frequency emitted by the tube
will elicit sympathetic resonances, high frequency harmonics from the strings, which might lead
one to deduce that the phenomenon could be reversed – if a string is excited by a finger, the
produced signal and its harmonic content will find the amplification and resonant potential of
various sectors of the back, confirming its second role as a multipurpose resonator.

It is easy to measure the vibrations of a back at a given point when excited by one of the open
strings. A dial gauge capable of measuring up to one hundredth of a millimetre will yield variations
that can be significant: 2 to 3 hundredths of a millimetre for low E and 4 to 5 hundredths for the
open D, displaying the back's preferences for certain frequencies.

Reflector

The back is also a reflector, or as some builders say, a 'pusher' of sound. In effect, the interior sound
wave generated by the soundboard upon release of the string (a compression series) will encounter a
more or less rigid, more or less polished or varnished surface when it reaches the inside of the back.
This series of waves will be reflected back – slightly damped or not – through the soundhole, which
acts like a speaker placed at the centre of the soundboard, contributing to the overall sound
produced by the instrument – all surfaces of which are actually vibrating (1).

(l) These waves emitted from the soundhole can be visualised by placing a light piece of paper on
the soundboard, almost entirely covering the soundhole opening. Plucking a low note will cause the
paper the tremble
A number of luthiers active in the first third of the 20th century even thought of placing an 'acoustic
cone' facing inwards, embedded around the rosette and under the soundboard. This device was
supposed to 'collect' all of the waves reflecting off the back, and direct them towards the exit with
greater ease. This system has not withstood the test of time, so the experiment must not have been
conclusive.

Comments
All of the guitar's parts components vibrate, but it is largely a directional instrument. A listener
facing the instrument will hear a louder sound than the listener seated at the side. We have already
seen that the back and its first brace hold the neck heel in position. This part of the back vibrates
more intensely due to the lever effect produced by the neck pulling back and forth in accordance
with the string's vibration (to simplify the concept). This can easily be verified by pressing a hand
onto this spot after strumming a chord, and noting the disappearance of one or more harmonics as
well as their return if the hand is removed again. This is more apparent the more flexible and lightly
braced the back is, and the lower the struck chord is.

These observations led Aguado to conceive of a device he named the 'Tripod' in the last century,
which allowed him not to squeeze and rest the guitar against the player, allowing it to vibrate freely
(see cover).

The lever represented by the neck and its heel applies yet another direct force onto the flexible,
lightly braced back; it acts on the back in time with the impulse of the vibrating string. Each string
excursion corresponds to maximum tension at its extremes, leading to an elongation of the back if it
is arched, flattening it out.

For a low E, at about 80 Hz, there will be 160 such movements per second, producing a vibration an
octave above the fundamental – a strong second harmonic – which is common in guitars with thin
backs. Theoretically, if the part of the soundboard in front of the bridge (1) is excited by the same
antinode and the same tension present in the vibrating string, its inward displacement will create an
inverse motion to that in the back, which move outwards. The combination of these two movements
will create a train of high amplitude sound waves, creating an even more prominent second
harmonic, generating a deep, full bass response with good timbre.

However, if the part of the soundboard behind the bridge is most active, it's outward motion will be
compensated by the back's inward motion, and the two can cancel each other out; this explains the
presence of poor bass response in lightweight guitars with thin construction that should not suffer
from this lack.

In the real world, every possible vibrational situation is possible, layering onto each other and,
depending on the frequency of the note, making better or worse use of their resonators, resulting in
a more or less homogenous, satisfyingly even response.

Given the broad array of problems facing the luthier – at the level of the neck, the soundboard, the
back, the sides and the harmonious cooperation between all of these coupled parts – numerous years
of observation are required. The situations are always complex and often contradictory, but if the
answers start forming, the investigator and creator draw satisfaction from them. Albeit limited
satisfaction, as the degree of control and advancement opens the door to other problems.

(l) The bridge, according to the internal bracing layout selected for the soundboard, can
preferentially excite the area in front of or behind where it is situation.
In order to precisely define elements of back and side response, we created and performed a series
of experiments.

Description of experiments (and luthier's comments)

To this effect, we constructed a guitar with a back that was laminated and arched in from three
pieces (three plys) – two rosewood plys surrounding a 1.8 millimetre thick centre ply. The total
thickness was 2.8 millimetres. The maximum height of the arch was 19 millimetres, with a total
weight of 272 grams. This back was unbraced, it was experiment number 1.

Experiment 1
This guitar was easy to play due to the extreme flexibility of the back. It had strong, deep basses
due to the back's low fundamental frequency as a resonator, whose movements – initiated by the
action of the neck and its heel – also contributed to the bass response. On the other hand, the mids
and trebles had poor timbre and were of little interest, with little homogeneity in terms of timbre
and sound level. The three highest notes on the fretboard were weak.

Experiment 2
On the same guitar, the sound hole was completely blocked using a foam rubber stopper. The
instrument emits only part of its normal sound, the soundboard's movement is inhibited due to the
air trapped inside, which tends to couple soundboard and back, the sound is short, extremely
damped, heavily veiled, with poor bass.

Experiment 3
Having liberated the soundhole, an opening was made in the back comparable to the size of the
soundhole (87 millimetres) at the centre of the widest point of the back. The basses lose all of their
initial timbre, the sound becomes generally poor. The vibrational mode of the back has changed due
to the opening. Part of the sound passes through the back and is dispersed (the instrument is less
directional); the air volume in the soundbox loses its role, and the soundhole appears to emit very
little sound.

Experiment 4
The soundhole is stoppered again, leaving the hole in the back in place. The guitar works better, the
'speaker' emits the series of compressed waves, the bass frequencies from the back of the soundbox
which now functions better. The trebles seem to suffer less.

Experiment 5
The back glued on initially is removed, and replaced by a very heavy back (braced with three sticks
of mahogany), consisting of two layers of different woods glued together. The first is rosewood,
weighing 280 grams, the second is mahogany, weighing 155 grams and placed on the outside,
yielding a total weight of 435 grams without bracing. The functioning guitar weighs 1920 grams.
The sound is dry but lacking in suppleness, rich but lacking in mellowness, it is difficult to play as
the neck is entirely immobilised by the thick, heavily braced back. The direct effect of the heel and
neck, which tries to stretch and relax lightly arched back is nullified due to its rigidity. As a
resonator, this back favours higher frequencies.

Experiment 6
The superficial layer of mahogany is planed off the lower bout of the back (fig 45), representing 90
grams less of wood, for a total weight of 1830 grams for the guitar. The sound is better, the
mellowness appears, the touch seems easier. The normal rosewood back, free from an additional
layer of mahogany on its lower bout, once again favours lower frequencies and yields a more
interesting, warmer timbre. The balance between basses and trebles is satisfying. This is the best of
the entire series of experiments, but the three highest notes are still weak and have poor timbre.

Experiment 7
The remaining layer of mahogany is planed off the upper bout (fig 45). The guitar now weights
1765 grams with a single layer indian rosewood back weighing 280 grams (with its three additional
mahogany braces). The sound becomes even rounder but also thicker and mushier, less clear, less
crystalline. The basses are dark, sound across the fingerboard is uneven, the three highest notes
remain dull. The back works even better for the bass frequencies, and the tone is darker by a degree.
The neck is far more free to move under string action, the thickness of the back being thinner at the
heel than it was in the previous experiment.

Experiment 8
The normal rosewood back is removed carefully, and the side depth is halved. The removed back is
then reinstalled – it is re-glued, resulting in a very thin guitar, only 6 cm deep. Overall, the sound is
fairly thin, clear and biting. Basses are weak, lacking in depth (but the three highest notes on the
fretboard are better), with the lack of bass resulting in a feeling of dryness when listened to directly.
The volume of the cavity, of the soundbox having been halved, the instrument has lost one of the
major components and is missing its 'boxiness'. The fundamental frequencies of the sides are
significantly higher, naturally vibrating at a higher pitch, reinforcing the extreme trebles. The back
is relatively passive in the bass frequencies, the action of the neck and heel is minimal due to its
size (the excursion of the 'lever' arm are much reduced).