See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/52001722

Wood for sound

Article  in  American Journal of Botany · October 2006

DOI: 10.3732/ajb.93.10.1439

CITATIONS READS
74 3,091

1 author:

Ulrike Wegst
Dartmouth College
64 PUBLICATIONS   4,131 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Materials in Musical Instruments View project

All content following this page was uploaded by Ulrike Wegst on 07 June 2016.

The user has requested enhancement of the downloaded file.

American Journal of Botany 93(10): 1439–1448. 2006.

WOOD FOR SOUND1

ULRIKE G. K. WEGST2
Max-Planck-Institut für Metallforschung, Heisenbergstr. 3, D-70569 Stuttgart, Germany; and Lawrence Berkeley National
Laboratory, Materials Sciences Division, Berkeley, California 94720 USA

The unique mechanical and acoustical properties of wood and its aesthetic appeal still make it the material of choice for musical
instruments and the interior of concert halls. Worldwide, several hundred wood species are available for making wind, string, or
percussion instruments. Over generations, first by trial and error and more recently by scientific approach, the most appropriate
species were found for each instrument and application. Using material property charts on which acoustic properties such as the
speed of sound, the characteristic impedance, the sound radiation coefficient, and the loss coefficient are plotted against one
another for woods. We analyze and explain why spruce is the preferred choice for soundboards, why tropical species are favored
for xylophone bars and woodwind instruments, why violinists still prefer pernambuco over other species as a bow material, and
why hornbeam and birch are used in piano actions.

Key words: wood; acoustical properties; materials selection; musical instruments.

‘‘[I]t appears probable that the progenitors of man, either the plants have mechanical properties that would have made it
males or females or both sexes, before acquiring the power of even easier for our ancestors to shape them with relatively
expressing their mutual love in articulate language, endeav- simple stone, bone, or antler tools.
oured to charm each other with musical notes and rhythm. . .’’, The earliest surviving flutes were found in Europe.
wrote Darwin in 1871 (p. 880). So far we can only speculate, However, finds of similar and other musical instruments on
when our early ancestors started to perform music, on the role all continents show that sound and rhythm have been an
of music in our biology and evolution (Cross, 2001), and when intricate part of human nature, independent of region and
the first musical instruments were made and from which culture.
materials. What we know for certain is that about 35 000 years
ago, when the oldest surviving sculptures and cave paintings
were created, flutes were played in the Geißenklösterle Cave in MUSICAL INSTRUMENTS
the southwestern part of Germany (Conard, 2004). The flutes
heard then are universally accepted as the oldest musical Over the millennia since the swan-bone and mammoth-ivory
instruments found so far. Of the three flutes found in the flutes, the range and color of sounds that can be produced by
Geißenklösterle Cave, two are made of hollow swan wing humans have increased immensely. The development of new
bones, and the third of solid mammoth ivory that was first and improved musical instruments has played an important part
carefully separated into halves for hollowing, then glued in this, and for many of them, the human voice has been the
together along a perfectly prepared, air-tight seam (Conard, benchmark. Today, musicologists group the vast number and
2004). All three flutes show not only early artistry in their kinds of past and current musical instruments into five families
manufacture, but experiments by Seeberger (2003) on modern based on a system devised by von Hornbostel and Sachs
reproductions of these flutes conclusively demonstrate that they (1914). The initial four classes were (1) idiophones—instru-
are intricate musical instruments with which complex and ments that make sound primarily by vibrating themselves,
aesthetically pleasing music can be produced. without the use of membranes or strings (e.g., xylophones); (2)
The oldest surviving examples of musical instruments are membranophones—instruments that use a stretched membrane
to create the sound e.g., drums); (3) chordophones—instru-
those made from bone and ivory—probably not because they
ments that rely on a stretched string (e.g., violins, guitars,
were the preferred material for flutes, but rather because purely
pianos); and (4) aerophones—instruments that rely on
organic materials decay much more quickly than their
a vibrating air column for sound creation (e.g., flutes, clarinets
mineralized counterparts. Flutes of a similar design and other and didgeridoos). More recently, a fifth class has been added to
musical instruments were likely made from leaves, grasses, the Sachs–Hornbostel system: (5) electrophones—instruments
wood or fruits at the same time or even earlier than these oldest that produce sound by electronic means (e.g., keyboard
surviving ones. Plants, as do bird wing bones, have shapes synthesizers).
suitable for musical instrument making, and additionally, Even though the range and sophistication of musical
instruments has increased significantly, particularly during the
1
Manuscript received 8 May 2006; revision accepted 17 July 2006. last four centuries or so, the range of materials from which the
The author thanks J. Woodhouse and C. Y. Barlow, Cambridge instruments of all these classes are manufactured has changed
University Engineering Department; R. Pitteroff, Bosch, Stuttgart; W. remarkably little. Whenever music is made by hand, whatever
Mauer, Louis Renner GmbH & Co. KG; A. P. Tomsia, Lawrence Berkeley
National Laboratory; and E. Arzt, MPI für Metallforschung, Stuttgart for the location and culture, from folk to classical, from jazz to rock
very stimulating and educational discussions; and two anonymous and pop, the vital parts in most musical instruments are still
reviewers for valuable comments and suggestions. This paper was written made from natural materials and primarily from wood, despite
during the author’s stay at the Lawrence Berkeley National Laboratory, the arrival of sophisticated alloys, polymers, and composites.
whose hospitality and support are gratefully acknowledged. There are a number of good reasons why this is so, as will be
2
E-mail: wegst@mf.mpg.de illustrated. We begin with a brief description of the composition
1439
1440 A MERICAN J OURNAL OF B OTANY [Vol. 93

and structure of wood that give it its exceptional mechanical and

acoustical properties.

WOOD AS AN ENGINEERING MATERIAL

The unique and desirable spectrum of physical and

mechanical properties of wood that so far can only in
exceptional cases truly be matched by manmade materials,
make it the material of choice for a multitude of applications
ranging from construction to sports equipment and musical
instruments even today. Wood is advantageous in its
comparative abundance and in being relatively easy to shape
with simple tools. One feature that sets wood apart from most
manmade materials is that it is an orthotropic material, meaning
that it has unique and independent mechanical properties in the
directions of three mutually perpendicular axes: longitudinal,
radial, and tangential. The longitudinal axis (L) is defined as
parallel to the fiber (grain), thus along the length of a tree trunk;
the radial axis (R) is perpendicular to the growth rings; and the
tangential axis (T) is perpendicular to the grain but tangent to
the growth rings. This orthotropy is due to the cellular structure
of wood. Wood is primarily composed of hollow, slender,
spindle-like cells, that are arranged parallel to each other along Fig. 1. A material property chart for woods, plotting Young’s
the trunk of a tree. The microscopic properties of the individual modulus, E, against density, q, for woods parallel to the grain. It
cells such as their composition and structure, their physical and illustrates that Young’s modulus and density are almost linearly correlated.
mechanical properties, and their shape and connectivity Figure created using the Natural Materials Selector (Wegst, 2004).
determine the overall performance of wood.
Wood is a hierarchically structured composite. The cell walls perpendicular to the grain of this orthotropic material are
consist of cellulose microfibrils embedded in a lignin and among these and have been shown to be important for musical
hemicellulose matrix in which minor amounts (5–10%) of plate vibration. Lacking complete sets of measurements of
extraneous extractives (e.g., oils) are contained (Wood these moduli for woods used in musical instruments, we
Handbook, 1999; Dinwoodie, 2000). Variations in the volume concentrate on the Young’s modulus parallel to the grain
and chemistry of these ingredients, combined with differences because it has been determined for a large number of wood
in the amount and distribution of porosity, determine the species and because the Young’s modulus, together with the
structure and thus the density and mechanical properties of wood’s density, determines most acoustical properties of
a wood. While the properties of a single wood species are a material. Additionally, the side hardness is important
constant within limits, the range of properties among species
whenever wood carries contact or impact loads, as is the case
can be large. Worldwide, the density of wood ranges from
in xylophones, for example. For clear, straight-grained soft-
about 100 kg/m3 for balsa (Ochroma pyramidale) to about
woods and hardwoods, the Young’s modulus parallel to the
1400 kg/m3 for lignum vitae (Guaicum officinale) and
snakewood (Brosimum guianense) (Wagenführ and Schreiber, grain and side hardness may be estimated from the density, q,
1989), a value close to that of carbon-fiber-reinforced polymers in kg/m3 according to the correlations in Table 1 (Wood
(CFRP) (Fig. 1). However, even in Europe, which has a small Handbook, 1999). Figure 1, a plot of Young’s modulus parallel
diversity of tree species due to a relatively recent ice age, the to the grain vs. wood density, illustrates the modulus–density
density ranges from about 400 kg/m3 for willow (Salix alba) to correlation; the diagonal line has a slope of 1.
about 800 kg/m3 for hornbeam (Carpinus betulus) and 950 kg/ Another important feature peculiar to wood and important
m3 for boxwood (Buxus sempervirens) (Sell, 1989). From for musical instruments is that it reacts and adapts to the
these, our ancestors found, over the generations and by trial environmental conditions to which it is exposed, particularly
and error, the best wood species available for a given function. that it exchanges moisture with air. Material properties that are
As today, their choices were not made on purely technical and critical for the acoustical performance of a wood such as
mechanical grounds, such as the wood’s decay resistance and density, Young’s modulus, damping, and shrinkage are highly
dimensional stability and the ease with which it could be dependent upon the wood’s moisture content. Thus, important
shaped and joined. Very important always were additional criteria during the material selection process are also how much
ones, such as the wood’s appearance—its texture, grain pattern, and how quickly a wood exchanges moisture with the
and color. We will see later that these last, seemingly ‘‘soft’’
criteria are often the critical ones that determine whether TABLE 1. Correlations of Young’s modulus, hardness and density (q) in
a material will be accepted or not in cases of material kg/m3 for soft and hardwoods.
substitution, even when the technical criteria are clearly met.
Property correlations Softwoods Hardwoods

Physical and mechanical properties of wood—Many

Young’s modulus [GPa] 56.3
q0.84 121.1
q0.7
physical and mechanical properties of wood are correlated Side hardness [N] 0.229
q1.5 6.48
103
q2.09
with density. The Young’s and shear moduli parallel and
October 2006] W EGST —W OOD FOR SOUND 1441

environment and how the moisture affects its dimensional longitudinal value. Generally, the speed of sound in wood
stability and mechanical properties. In general, the speed of decreases with an increase in temperature or moisture content
moisture sorption decreases with increasing density and and proportionally to the influence of these variables on
content of extractives (Sell, 1989). The rate and amount of Young’s modulus and density. It decreases slightly with
water uptake along with the dimensional stability of wood can increasing frequency and amplitude of vibration (Wood
further be controlled through treatments with waxes or oils. Handbook, 1999).

Acoustical properties of wood—The acoustical properties of Characteristic impedance—Like the speed of sound, the
wood, such as the volume, quality, and color of the sound of characteristic impedance is directly related to the modulus of
xylophone bars and soundboards are determined by the elasticity and density of a material. This quantity is important
mechanical properties of the material from which they are when vibratory energy is transmitted from one medium with
made because the sound is produced by vibrations of the impedance z1 to another with impedance z2. The first medium
material itself. The properties on which the acoustical could be a string and the second the soundboard of a musical
performance of a material depends are primarily its density, instrument. The ratio of the reflected sound intensity, Ir, to the
Young’s modulus, and loss coefficient. They determine the incident intensity, I0, can be expressed as a function of the
speed of sound in a material, the eigenfrequencies of a wooden impedances of the two media:
bar, and the intensity of the radiated sound.  
The most important acoustical properties for selecting Ir z2  z1 2
¼ : ð5Þ
materials for sound applications, such as musical instruments I0 z2 þ z1
and building interiors, are the speed of sound within the
material, the characteristic impedance, the sound radiation And the ratio of the transmitted sound intensity, It, to the
coefficient, and the loss coefficient: incident intensity is
It 4z2 z1
 The speed, c, with which sound travels through a material, is ¼ : ð6Þ
I0 ðz2 þ z1 Þ2
defined as the root of the material’s Young’s modulus, E,
divided by the material’s density, q: From these equations, we see that the transmitted intensity goes
sffiffiffi to zero if there is a large mismatch between z1 and z2, thus
E either z1 ,, z2 or z2 ,, z1 (Fletcher and Rossing, 1991).
c¼ : ð1Þ The soundboard’s impedance is proportional not only to the
q
characteristic impedance of the material from which it is made,
 Incidentally, this ratio, which describes the speed of but also to the square of the soundboards thickness. As a result,
soundboards with considerable thickness, such as that in
longitudinal waves in a material, also characterizes the pianos, for example, have an impedance significantly larger
transverse vibrational frequencies of a bar. than that of the strings. To achieve a high sound quality, the
 The impedance, z, of a material, is defined as the product of impedances of the strings and the soundboard must thus be
the material’s speed of sound, c, and its density, q: controlled very carefully. This is not a trivial undertaking,
pffiffiffiffiffiffi because two conflicting requirements must be met: sufficient
z ¼ cq ¼ Eq : ð2Þ vibratory energy must be transmitted from the string to the
soundboard to make the strings vibrate audibly, while the
 The sound radiation coefficient, R, of a material, is defined as energy should not be transmitted too readily or too rapidly,
the ratio of the material’s speed of sound, c, to its density, q: causing the vibrations of the string to die down quickly and
sffiffiffiffiffi their sound to resemble that of a thud (Benade, 1990).
c E
R¼ ¼ ð3Þ Sound radiation coefficient—The sound radiation coefficient
q q3
describes how much the vibration of a body is damped due to
 The loss coefficient, g, measures the degree to which sound radiation. Particularly in the case of idiophones, such as
a material dissipates vibrational energy by internal friction. xylophones and soundboards, a large sound radiation co-
Other measures of damping include the quality factor, Q, the efficient of the material is desirable if we wish to produce
a loud sound. To maximize loudness, we need to maximize the
logarithmic decrement, d, and the loss angle, w. For amplitude of the vibrational response of the soundboard for
excitation near resonance and small damping, these quanti- a given force, a quantity that is described by the frequency
ties are related as (Newland, 1989): response function (Barlow, 1997). According to Skudrzyk’s
(1980) mean value theorem, the mean value of the amplitude is
1 d
g¼ ¼ ¼ tanw : ð4Þ equal to the driving-point admittance, which for an infinite
Q p isotropic plate is
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Speed of sound—As we saw in Eq. 1, the speed of sound is 1 3ð1  m2 Þ
directly related to the modulus of elasticity and density. It is Y¼ 2 ; ð7Þ
4h Eq
roughly independent of wood species, but varies with grain
direction. The transverse Young’s modulus of wood is only where h is the thickness of the soundboard or the bar and m is
between 1/20 to 1/10 of the longitudinal; consequently, the the Poisson’s ratio of the material from which it is made.
speed of sound across the grain is only c. 20 to 30% that of the Assuming that we do not wish to change the timbre of the
1442 A MERICAN J OURNAL OF B OTANY [Vol. 93

instrument, we need to ensure that the modal density of the whether it is hit by a soft or a hard mallet and whether the
soundboard has the correct value (Manning, 1997). For an vibration is caused by a plucked or a bowed string. The
isotropic plate, the modal density (in frequency space) is harmonics also depend on the shape of the body and on the
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi material from which the body is made, as is explained next.
A q
nðxÞ ¼ 3ð1  m2 Þ ; ð8Þ
h E
where A is the area of the soundboard. Rearranging Eq. 8 and MATERIAL SELECTION FOR MUSICAL INSTRUMENTS
substituting it for h in Eq. 7 yields an expression that describes
how the mean amplitude for a soundboard of a given area and When we plot the various physical and mechanical
the modal density depend on the soundboard’s material properties and acoustical quantities described earlier against
properties: one another for the woods commonly used for different types
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi of instruments, we can illustrate the design requirements for
nðxÞ E 1 these instruments and analyze why certain species are
Y¼ : ð9Þ especially suited for particular sound applications and therefore
4A2 q3 3ð1  m2 Þ
traditionally chosen by musical instrument makers. The scales
Thus, if we wish to maximize the average amplitude or the of the charts are logarithmic to accommodate a large range of
average loudness of a violin for a given modal density and materials and to be able to represent the acoustical criteria with
prescribed soundboard dimensions, assuming the term in straight lines. The materials are represented as bubbles and
parentheses equals 1, we need to maximize the combination grouped by colors, which indicate the traditional use for
of materials properties which was earlier defined as the sound a material. Data sources for information on materials and their
radiation coefficient, R: properties are Haines (1979), Richter (1988), Finscher (1994),
sffiffiffiffiffi Yano et al. (1997), Sadie and Tyrell (2001), Oberhoff (2003),
E Rujinirun et al. (2005) and Bucur (2006).
R¼ : ð10Þ Figure 1 shows Young’s modulus, E, plotted against density,
q3
q, for woods parallel to the grain. From the chart, we read that
If we wish to maximize the peak response of a soundboard or soundboards are made from low density woods, which have
a bar, rather than the average response, we need to maximize a relatively high Young’s modulus, that woods for wind
the ratio of the sound radiation coefficient to the loss instruments and xylophone bars have a high density, and that
coefficient, g (Barlow, 1997). woods for violin bows have both an exceptionally high density
Incidentally, the quantity that we need to maximize if and a high Young’s modulus. The reasons for these choices are
maximum stiffness per unit weight is sought is the same as that described in detail later.
represented by R. In soundboard design, this means that the one Figure 2 shows the speed of sound, c, plotted against
that radiates the loudest sound also is the stiffest per unit mass, density, q, allowing two additional acoustical properties to be
thereby ensuring that the thin top plates of violins, which read from this chart. Lines with slope 1 represent the ratio c/q,
typically are only 2–3 mm thick, can support the 70 to 90 N (c. the sound radiation coefficient, R. Lines with slope 1
7 to 9 kg) load of the strings with minimal deflection. represent the product of the two, cq, the characteristic
impedance, z. All materials on an individual line with a slope
Loss coefficient—When a solid material vibrates, it is of 1, or 1, have the same value for this property, thus they
strained and some of its mechanical energy is dissipated as radiate the same amount of sound, or have the same
heat by internal friction. The mechanism by which this occurs impedance, respectively. Materials above a line have a higher
in wood is complex and depends on the temperature and value for the respective property than those below. Woods for
moisture content within a sample and on the type and amount soundboards stand out. They have both an exceptionally high
of extractives characteristic for the wood species. The value of speed of sound and a remarkably high sound radiation
the loss coefficient ranges from about 0.1 for hot, moist wood coefficient.
to about 0.002 for air-dry wood at room temperature. Unlike Figure 3 shows the sound radiation coefficient, R, plotted
the three acoustical properties described earlier, the loss against the loss coefficient, g. Lines with slope 1 represent the
coefficient, which quantifies the damping of vibration due to ratio of the two, the peak response of a sounding body, rather
internal friction, is independent of density and Young’s than its average response represented by R. All materials on an
modulus, as Fig. 4 illustrates. individual line have the same value for this property. Materials
above a line produce a louder peak response than those below.
Pitch and timbre of sound—The loudness or intensity of The chart reveals that woods for soundboards have both a high
a sound depends on the square of the amplitude of the average and a high peak response and that soundboards,
vibration, as described earlier. The pitch of sound of a musical xylophone bars, and violin bows all have an exceptionally low
instrument is determined by the spectrum of frequencies it loss coefficient.
radiates and transmits into the air. Each body has its own Figure 4 shows the Young’s modulus plotted against the loss
particular set of eigenfrequencies defined by the size of the coefficient, g, illustrating that these properties are not at all
vibrating body, the material from which it is made, and in the correlated. The chart also shows that woods for violin bows are
case of strings, on its tension. The timbre and quality of the unique in their combination of an exceptionally high Young’s
sound that a vibrating body produces is due to the presence of modulus with a very low loss coefficient.
eigenfrequencies, also termed overtones or upper partials, and In the next sections, these four material property charts are
their relative strengths. Which overtones of a sounding body used to explore which of the depicted properties dominate the
are excited depends on what causes the body’s vibrations: material choice and quality of sound produced by woodwind
October 2006] W EGST —W OOD FOR SOUND 1443

instruments, xylophone bars, soundboards, piano actions, and moisture is frequently further improved by oiling or by
violin bows. impregnating with paraffin.
Because African blackwood is an endangered species, plain
Woods for wind instruments (aerophones)—Wind instru- and wood-filled polymers have recently been introduced as
ments made from wood are, for example, the recorder, the a substitute material for the manufacture of clarinets. Makers
flute, the oboe, the clarinet and the bassoon. We concentrate on and players tend to claim that they can hear differences in
these, since the principle of sound production applies to all sound quality between instruments made from different woods
other instruments which, like these, use an air column as and alternative materials. Such differences can be due to
a resonating body to produce and radiate the sound. The sound differences in surface and bore properties of the instrument or
characteristics of these instruments that allow us to distinguish due to the wall material’s thermal properties, because sound
one from the other are due to the mode of air column excitation waves involve temperature fluctuations, which, due to the
(whether by blowing over the column for the flute or through thermal inertia in the walls, produce a damping effect.
a single reed for the clarinet or a double reed for the oboe), on Appropriate material selection and the optimization of the
the shape of the air column (whether it is cylindrical or manufacturing process should make it possible, however, to
conical), on the length of the air column (controlled by opening manufacture polymer instruments that can produce a sound
and closing the finger holes on the instrument), and very similar in quality and timbre to that of wooden ones. Current
importantly, on the exact shape and design of the finger holes reservations of musicians toward polymers use for top quality
(Benade, 1990; Fletcher and Rossing, 1991). instruments might have an additional, more subtle reason: the
Traditionally, the woods from which these instruments are ‘‘feel’’ towards an instrument made from synthetic ‘‘plastics’’ as
made are dense, have a fine structure, and a high dimensional opposed to the natural material wood.
stability, particularly when exposed to high levels of moisture.
Before the arrival of tropical wood species in Europe, boxwood Woods for xylophones (idiophones)—The xylophone
(Buxus sempervirens) and fruit woods such as pear (Pyrus consists of wooden bars mounted horizontally and supported
communis) and plum (Prunus domestica), and in exceptional by a soft material at the two nodal points of its lowest
cases the material of the oldest surviving flute, ivory, were eigenfrequency. The fundamental frequency of vibration of an
used. Since the arrival of the denser tropical species, African individual xylophone bar is determined by its length. Bars of
blackwood or grenadill (Dalbergia melanoxylon), Brazilian identical length and shape have vibrational frequencies that
rosewood (Dalbergia nigra), and Macassar ebony (Dyospyros scale with the speed of sound from the chosen material. An
celebica) have been favored for oboe and clarinet, while the arch is cut into the underside of xylophone bars for two
large bassoon has always been and still is made of maple (Acer reasons: to reduce the bar length required for low pitches and to
tune the overtones (Fletcher and Rossing, 1991).
platanoides) (Table 2).
Xylophone bars are impulsively excited instruments. The
Why were these woods chosen? Their mechanical properties
energy that causes the bar to vibrate is transferred to it in a time
differ vastly, by almost a factor of two between the least and
span that is very short in comparison to the decay time of the
the most dense. The reason for their use is their structure. All bar’s vibration. As a result, the choice of the mallet used to
these woods can be turned and drilled with great accuracy, and excite the bar has an enormous effect on the color and quality
they are sufficiently dimensionally stable under the influence of of the sound produced (in a way similar to the quality of the felt
moisture. These characteristics are critical because for these on the hammers in a piano). A hard-headed mallet produces
instruments, the material determines the sound quality of the a bright and penetrating sound, while a softer mallet produces
instrument not by vibrating itself—the acoustic pressure of the a more mellow sound that is often preferred for the lower notes.
standing wave in the air column is by far too weak to excite and This effect is due to the frequencies excited on impact. The
couple to vibrations of the thick wooden tube of wooden harder mallet has a shorter contact time upon impact and
instruments to produce audible vibrations—but the sound excites a spectrum rich in overtones characteristic for a given
quality is determined by the interaction of the material with the material, while the softer excites only the harmonically tuned
enclosed column of vibrating air. lower partials and dampens the higher partials due to its longer
The tube material influences the sound of the instrument and contact time.
its playability significantly and mainly by two mechanisms: by If a significant loudness and reverberation time of an
vibrational damping due to air friction at the tube walls and by impulsively excited instrument is to be achieved, the best
turbulence in the vibrating air at the edges (Benade, 1990). materials are those that radiate sound well. Additionally, bars
Vibrational damping is lower in tubes with a smooth finish. with a low loss coefficient will result in a brighter sound
Turbulences, which also dampen the vibration and affects the because the higher partials are less damped than in a material
tonal quality of an instrument, are reduced when the edges can with a high loss coefficient (Richter, 1988; Bork, 1995).
be cut precisely and finished slightly rounded as for finger Figures 3 and 4 show that xylophone bar woods (Table 3)
holes, for example. The best wood for a wind instrument thus have a low loss coefficient and high value of peak sound
has a high density and a fine grain to obtain an optimal finish of radiation, but not of the average sound radiation coefficient,
the tube walls and finger holes. Because the breath of the and that woods for soundboards (Table 4) have much higher
musician introduces significant amounts of water into the air values for both. What are the consequences of these differences
column, the material should further absorb a little moisture and why are xylophone bars not made from soundboard
since water droplets on the tube wall spoil the sound. At the woods? For two reasons: the first is that soundboard woods
same time, the material should be dimensionally stable when have a very low density (Fig. 1), and as a result a low side
exposed to moisture, so that the instrument remains tuned over hardness (Table 1). Low-density species hit by a mallet would
a significant amount of time. All the woods named earlier have dent easily or even split, and the bars’ tuning and damping
these qualities to a larger extent than others. Their resistance to properties would suffer as a result. Thus we seek a wood with
1444 A MERICAN J OURNAL OF B OTANY [Vol. 93

Fig. 2. A material property chart for woods, plotting the speed of Fig. 4. A material property chart for woods, plotting Young’s modulus
sound, c, against density, q, allowing two additional acoustical properties against the loss coefficient, g, revealing that these two properties are not at
to be read from this chart. Lines of slope 1 represent the sound radiation all correlated. Figure created using the Natural Materials Selector (Wegst,
coefficient, R ¼ c/q. Lines of slope 1 represent the product of the two, the 2004).
characteristic impedance, z ¼ cq. Figure created using the Natural
Materials Selector (Wegst, 2004).
that, due to their higher density, woods for xylophone bars are
a high side hardness, a design requirement best met by high- much less damped by sound radiation than soundboard woods,
density tropical species (Holz, 1996). The second reason is also resulting in slower sound decay. Additionally, the higher
related to density: the sound of xylophone bars should decay impedance characteristic of these woods (Fig. 2) also means
sufficiently slowly, a criterion of less importance in a contin- that they lose less energy through the supports because the
uously excited string instrument, for example. Figure 2 shows impedance mismatch is greater. Consequently, tropical species
with high density, high peak response, and low loss coefficient
such as Amazon rosewood (Dalbergia spruceana), Honduras
rosewood (D. stevensonii), African padauk (Pterocarpus

TABLE 2. Woods traditionally used for wind instruments.

Common name Taxon

Norway maple Acer platanoides

Sycamore (curly) maple Acer pseudoplatanus
Ma-had Artocarpus lakoocha
Cocuswood Brya ebenus
European boxwood Buxus sempervirens
Castella boxwood, castelo Calycophyllum multiflorum
West Indian boxwood, zapatero Casearia praecox
Brazilian kingwood Dalbergia cearensis
Bahia rosewood Dalbergia decipularis
Indian rosewood Dalbergia latifolia
African blackwood, grenadill Dalbergia melanoxylon
Brazilian rosewood Dalbergia nigra
Ching-chan Dalbergia oliveri
Brazilian tulipwood Dalbergia variabilis
Cocobolo Dalbergia retusa
Macassar ebony Dyospyros celebica
Bubinga Guibourtia tessmannii
Red chacate Guibourtia schliebenii
Pear Pyrus communis
Fig. 3. A material property chart for woods, plotting the sound Plum Prunus domestica
radiation coefficient, R, against the loss coefficient, g. Lines of slope 1 Olive Olea europea
represent the ratio of the two, R/g, the peak response of a sounding body. Indian kamba Stephegyne parviflora
Figure created using the Natural Materials Selector (Wegst, 2004).
October 2006] W EGST —W OOD FOR SOUND 1445

TABLE 4. Woods traditionally used for soundboards, sides, and backs.

TABLE 3. Woods traditionally used for xylophone bars. Common name Taxon

Common name Taxon Silver fir Abies alba

King William pine Athrotaxis selaginoides
Granadillo Dalbergia cubilquitzensis Norway spruce Picea abies ¼ P. excelsis
Amazon rosewood Dalbergia spruceana Sitka spruce Picea sitchensis
Honduras rosewood Dalbergia stevensonii Pine Pinus sylvestris
Merbau Intsia bijuga Douglas fir Pseudotsuga menziesii
Wenge Millettia laurentii
African padauk Pterocarpus soyauxii Woods traditionally used for sides and backs of violins
Cristobal Platymiscium pinnatum Silver maple Acer saccharinum
Pau rosa Swartzia fistuloides Sycamore (curly) maple Acer pseudoplatanus

soyauxii), Ching-Chan (D. oliveri) and Ma-Had (Artocarpus new ones need to be ‘‘played-in.’’ Humidity and creep are
lakoocha) perform well by these criteria. believed to play an important role in this. Hunt and Balsan
(1996) show experimentally that regular playing at intermedi-
Woods for soundboards (chordophones)—The sound that ate or high humidities leads to an increase in stiffness and
a single plucked or bowed string produces is barely audible a decrease in loss coefficient. Beavitt (1996) presents
because one string sets only a small volume of air into motion. experimental evidence to support his hypothesis that creep
To produce sounds with satisfactory volume for our ears, the facilitated by humidity cycling results in changes in the
string must be coupled to a resonator, which has a better overtone spectrum of the instrument, making it sound more
coupling to air to transmit the vibratory energy of the string and sonorous and resonant. Segerman (1996, 2001) claims that
radiate the sound. In the violin family, the string is coupled via creep in newly strung instruments affects the sound as it
the bridge to the top plate of the instrument, the soundboard, absorbs sound vibrations and that vibrations accelerate creep
which usually is a piece of softwood with the grain running and thus help a newly strung instrument to settle in faster.
parallel to the strings. The bridge transmits the vibrations of the Other research shows that the gradual decomposition and loss
string to the soundboard, which is connected to the back plate of hemicellulose with time lowers a wood’s density without
by the sound post and the ribs. The back plate is part of the affecting its Young’s modulus (Bucur, 2006). This avenue is
vibrating structure and as such also contributes to radiating the being pursued further in current research to ‘‘age’’ soundboard
sound. The shape and material of the body strongly influence wood by infecting it with a carefully selected fungus to lower
the sound quality and the way in which it is radiated into the the density at a constant Young’s modulus and thereby
room. The f-holes in the top-plate, for example, have not only improving the sound radiation coefficient and quality of the
a direct influence on the vibrations of this soundboard, but also soundboards (Zierl, 2005).
provide a passageway through which the air enclosed in the Further research on the ‘‘playing-in’’ and ageing of musical
instrument’s body communicates its oscillations to the outside. instruments is required to explain conclusively the various
Figures 2 and 3 illustrate this. Soundboard-quality woods are observed phenomena in terms of the chemical, structural, and
not only the best sound radiators of all, but they also have a low mechanical properties of wood and the environment to which it
characteristic impedance. A low impedance is beneficial for is exposed.
sound transmission into the air. The impedance of spruce
(Picea sp.) is very similar to that of maple, the traditional Woods for piano actions—So far we have considered the
material for the bridge, back plate, and ribs. Maple also radiates transmission of vibratory energy from the string to our ear via
sound well; its characteristic impedance is sufficiently high to a soundboard. This leaves questions about how strings are
act as a reflector for the air oscillations within the corpus of the excited and which role wood plays here. To produce sounds on
instrument and to help radiate them into the room through the f- a piano, the pianist presses down a key, which by a complicated
holes in the top plate (Table 4). mechanism, the piano action, sets a hammer in motion. The
A closer look not only at the holes in the top plates of string hammer strikes the string at a suitably chosen point and thereby
instruments, but also at their overall shape reveals a feature excites it to vibrate. Over their lifetime, the various piano
common to almost all wooden musical instruments: both the action parts, which consist of hundreds of small levers, are
instrument itself and the holes cut into it are either round or moved millions of times. The best materials for such an
composed of arcs. Such instrument design is not only application are those that provide a reliable structure by having
aesthetically pleasing, but it is also prescribed by the a high wear resistance in the moving parts, a long fatigue life,
orthotropic nature of wood. The softwoods commonly used a high toughness to resist repetitive impact loading and, most
for soundboards, such as spruce, very readily split parallel to importantly, that are dimensionally stable and shapable with
the grain, particularly when they have the shape of a plate and great precision. For over 300 years, woods such as hornbeam
a modest thickness of 2–3 mm, as is typical for violin and (Carpinus betulus), beech (Fagus sylvatica) and maple (Acer
guitar soundboards. By cutting curves and circles, the platanoides) have been the prime choice for these applications
instrument maker avoids creating the stress concentrations because they fulfill all these requirements (Table 5).
associated with sharp corners. However, not only mechanical requirements have to be
Common beliefs are that regular playing and aging of wood fulfilled by the components of piano actions; some components
improve the acoustical properties of musical instruments, that are also chosen for their acoustical properties as described by
instruments that are exhibited in museums rather than played Holz (2000). The hammershank, a stick 120 mm in length and
lose their quality, that ‘‘old fiddles sound sweeter,’’ and that c. 5.5 mm in diameter, holds the felt-covered hammer and is
1446 A MERICAN J OURNAL OF B OTANY [Vol. 93

TABLE 5. Woods traditionally used for piano actions. quality sound can be achieved on a violin by an experienced
violinist. Factors influencing the playability of a bow are partly
Common name Taxon
structural (shape, point of balance) and partly material
Norway maple Acer platanoides dependent (mass, stiffness, mechanical damping) (Woodhouse,
Silver birch Betula pendula 1993a, b).
Hornbeam Carpinus betulus Particularly, the damping performance of the bow material is
Beech Fagus sylvatica thought critical for the bows playability. Additionally, the
damping properties of the bow stick are thought to have
a significant effect on the violin’s sound. Of great mechanical
traditionally made from birch (Betula pendula). First, its importance is the bow’s bending stiffness, which for a given
elasticity is tested by hand bending. Only those which return to geometry, depends on the Young’s modulus of the bow stick
their original shape are chosen, and these are then dropped onto material. Stiffness is important for three reasons. First, if a bow
a hard surface to excite their eigenfrequencies. Variations in is too flexible, the desired horse-hair tension is reached only
density and modulus between sticks lead to variations in the when the bow is almost straight. Second, if the spring stiffness
eigenfrequencies by as much as a fifth for these fixed of the stick-hair system is low, the control of the bow dynamics
dimensions. The sticks are grouped as ‘‘dark,’’ ‘‘medium,’’ or is reduced—this is particularly important for techniques in
‘‘bright,’’ depending on the sound produced and used in which the bow leaves the string, such as spiccato. And third,
respective sections in the piano. Birch with its higher speed of a low lateral stiffness leads to reduced bow control in this
sound is often preferred over hornbeam and maple for this direction. Finally, the mass of the bow affects the bow’s
application. How does the material for the hammershanks playability because it is considered advantageous if the lowest
affect the sound quality of the piano, and why is a high speed bow forces can be achieved by the self-weight of the bow and
of sound preferred, particularly for hammershanks in the a small additional load applied by the player.
descant registers of a piano? Might the higher speed of sound Pernambuco fulfills all these requirements. It has the desired
be an indicator of stiffness, and might the stiffness of the density, bending stiffness, and damping behavior for the
hammershank influence the contact time of the hammer and current violin bow design. However, like African blackwood,
with it the sound produced? Lacking rigorous analysis, we can the most revered wood for clarinets, this tropical wood
only speculate. So far, the only evidence in support of the currently faces a high risk of extinction in the wild, which
hypothesis that hammershank ‘‘tuning’’ improves the sound raises the question whether alternative materials for violin
quality of a piano is the long, continuing, and sometimes costly bows exist.
tradition of labor-intensive materials selection by an acoustic Figures 1 and 4 show that pernambuco is a wood with an
criterion. exceptionally high Young’s modulus for its density and
a remarkably low loss coefficient, which is thought to be due
Woods for violin bows—String instruments like the violin to an exceptionally high amount of extractives (Matsunaga et
have the ability to produce steady tones when a bow is used to al., 1996, 2000). Figure 4 also shows that alternative materials
excite the string’s vibrations. Originally, bows for string exist, which, on purely mechanical grounds, could make as
instruments had the same shape as musical bows and resembled good violin bows as pernambuco. Examples are the wood
those used for archery. The bow as we know it today is thought Swartzia (Swartzia spp.) and CFRP (violin-bow quality), to
to have been developed by the French bow-maker François name but two. In the case of alternative woods, however,
Tourte in the second half of the 18th century. He changed the further experimentation is required to explore whether they
shape of the bow from a concave to a convex curvature and is fulfill the other important design requirements, namely, that
additionally thought to have been the first to recognize the they are straight and fine-grained, free from defects such as
knots and splits, and whether they can be worked easily,
qualities of the still preferred material for bows, the tropical
whether they can be bent over heat, and whether they retain
wood pernambuco (Guilandia echinata, syn. Caesalpinia
their curvature well. Trials by bow makers and musicians are
echinata), a wood that first reached Europe as a dyewood for
needed for a thorough evaluation of alternative materials.
red cloth in the 16th century (Table 6). Other critical criteria are again those of ‘‘feel’’ and
Mechanically speaking, the wooden bow stick acts as a leaf appearance. The alternative wood or material has to be
spring. It tightens the horse hair sufficiently so that by pulling it acceptable to the musical community. As in the case of
across a string the player can excite and control its transverse woodwind instruments, aesthetics plays an important role: new
oscillations (Pitteroff, 1995; Wegst, 1996; Wegst and Ashby, materials for musical instruments are generally not only chosen
1996; Oberhoff, 2003). This interaction of string and bow is for mechanico-acoustical considerations, but also for optical
highly complex and still not fully understood. Neither are the ones. A wood which has the wrong look, color or feel may be
complete requirements for the bow stick material. Generally, rejected purely for these reasons—even if their performance
the quality of a bow is judged by the ease with which a top- was similar or even superior to other materials. CFRP initially
suffered from such reservations and the fact that it is a synthetic
material. Now it is more and more accepted by musicians,
TABLE 6. Woods traditionally used for bows for stringed instruments.
particularly for pieces that require a more forceful style of play.
The CFRP bow’s great advantage is that it is less prone to
Common name Taxon fracture than an old French pernambuco bow and that it can be
replaced.
Snakewood Brosimum guianense
Pernambuco Guilandia echinata
Massaranduba Manilkara elata Wood and acoustics in buildings—Not only musical
instruments produce sounds and noises. Buildings also do.
October 2006] W EGST —W OOD FOR SOUND 1447

Everyone has heard the sounds of wooden floors and stairs that are used to illustrate and explain why spruce is the preferred
act as soundboards and amplify rather than dampen the sound choice for soundboards, why tropical species are favored for
of walking. To a building’s inhabitants, they can be both xylophone bars and woodwind instruments, why violinists still
a severe irritant or a charming reminder of the building’s prefer pernambuco over other species as a bow material, and
history. However, most often they are undesirable, unless they why hornbeam and birch are used in piano actions.
are part of a clever alarm system. Probably the best-known
example of a floor deliberately designed to produce a noise, for LITERATURE CITED
added security, is the Nightingale Floor laid in Nijo Castle,
BARLOW, C. Y. 1997. Materials selection for musical instruments.
Kyoto, Japan. Its sound is a very pleasant one, similar to the
Proceedings of the Institute of Acoustics 19: 69–78.
cheeping of a nightingale. It is produced when the nails on BEAVITT, A. 1996. Humidity cycling. Strad (November): 916–920.
which the floorboards are freely suspended above the frame rub BENADE, A. H. 1990. Fundamentals of musical acoustics, 2nd revised,
against the floorboards when walked upon (Yokochi and reprint ed. Dover, Mineola, New York, USA.
Yoshimoto, 2004). BORK, I. 1995. Practical tuning of xylophone bars and resonators. Applied
Acoustics 46: 103–127.
Wood in concert halls—At least as much as we appreciate BUCUR, V. 2006. Acoustics of wood, 2nd ed. Springer Series in Wood
a home with pleasant room acoustics, we hope for sound Science, Springer, Berlin, Heidelberg, Germany.
optimization in a concert hall. The acoustics of rooms are CONARD, N. J. 2004. Eine Mammutelfenbeinflöte aus dem Aurignacien des
complicated by the fact that the sound produced by musicians Geißenklösterle. Archäologisches Korrespondenzblatt 34: 447–462.
CROSS, I. 2001. Music cognition, culture and evolution. Annals of the New
on stage reaches the audience not only directly, but also after York Academy of Sciences 930: 28–42.
many reflections on the surfaces of the room and the objects in DARWIN, C. 1871. The descent of man and selection in relation to sex. John
it. Important quantities in room acoustics are the level of Murray, London, England.
reverberant sound and the reverberation time because they DINWOODIE, J. M. 2000. Timber, its nature and behaviour, 2nd ed. E & FN
determine the loudness, the clarity, and the liveliness of the Spon, London, UK.
sound (Rossing and Fletcher, 2004). Optimal reverberation FINSCHER, L. [ED.]. 1994. Die Musik in Geschichte und Gegenwart:
times depend on the size and function of a room and range allgemeine Enzyklopädie der Musik (MGG). 2nd revised ed.
from 0.5 s for small practice rooms of up to 135 m3 to 2.2 s in Bärenreiter, Kassel, Germany.
large concerts halls of 20 000 m3 or larger, and also on cultural FLETCHER, N. H., AND T. D. ROSSING. 1991. The physics of musical
preferences: Europeans, for example, prefer 10% longer instruments, Springer-Verlag, New York, New York, USA.
HAINES, D. 1979. On musical instrument wood. Catgut Acoustical Society
reverberation times than Americans (Knudsen, 1988). Newsletter 1: 23–32.
Both the shape of the room and the acoustical properties of HOLZ, D. 1996. Acoustically important properties of xylophone-bar
the materials and objects in it determine the reverberation time. materials: can tropical woods be replaced by European species?
Materials with a highly porous surface, such as carpets, Acustica/Acta acustica 82: 878–884.
drapery, and upholstery absorb well at high frequencies, while HOLZ, D. 2000. Weshalb Hammerstiele aus Birkenholz? Europiano 40:
wooden panels act as reflectors and as resonators to absorb low 35–36, 39–40.
frequencies. The use of wood for flooring, seating, reflecting HUNT, D. G., AND E. BALSAN. 1996. Why old fiddles sound sweeter. Nature
panels, and wall panelling in concert halls in combination with 379: 681.
the aforementioned textiles is thus not only aesthetically KNUDSEN, V. O. 1988. Raumakustik. In Winkler [ed.], Die Physik der
pleasing, but also necessary for the optimization of the sound Musikinstrumente, 136–149. Spektrum der Wissenschaft, Heidelberg,
Germany.
field, the reverberation time, and the brightness of a concert MANNING, J. E. 1997. Statistical modeling of vibrating systems. In M. J.
hall for a wide range of frequencies. Crocker [ed.], Encyclopaedia of acoustics, 757–768. Wiley, New
York, New York, USA.
MATSUNAGA, M., E. OBATAYA, K. MINATO, AND F. NAKATSUBO. 2000.
CONCLUSIONS Working mechanism of adsorbed water on the vibrational properties
of wood impregnated with extractives of pernambuco (Guilandina
Over the millennia, we have learned to use wood to its best echinata Spreng.). Journal of Wood Science 46: 122–129.
advantage in musical applications. The musical instruments we MATSUNAGA, M., M. SUGIYAMA, K. MINATO, AND M. NORIMOTO. 1996.
know today are the result of the simultaneous optimization of Physical and mechanical properties required for violin bow materials.
Holzforschung 50: 511–517.
material and shape for the expectations of musicians and
NEWLAND, D. E. 1989. Mechanical vibration analysis and computation,
audiences at a given time and in a given culture. Despite much Longman, Harlow, UK.
scientific effort to illuminate the properties of universally OBERHOFF, S. 2003. Bögen für Streichinstrumente: Das elastische und
accepted perfect instruments and their reproduction, we still anelastische Verhalten von Fernambukholz und alternativen Materi-
rely mainly on the art, knowledge, and experience passed on alien. Diploma thesis. Institut für Metallkunde, Universität Stuttgart,
from one generation of skilled instrument makers to the next. Stuttgart, Germany.
They have the expertise to judge the quality of the material for PITTEROFF, R. 1995. Contact mechanics of the bowed string. Ph.D. thesis.
an instrument using eye, ear, and touch—and this often when it Department of Engineering, University of Cambridge, Cambridge,
is still hidden in the trunk of a tree or in wooden planks. UK.
The aim of this contribution is to illustrate the unique range RICHTER, H. G. 1988. Holz als Rohstoff für den Musikinstrumentenbau.
Moeck Verlag, Celle, Germany.
and combination of mechanical and acoustical properties of ROSSING, T. D., AND N. H. FLETCHER. 2004. Principles of vibration and
wood, which still make it the material of choice for musical sound, 2nd ed. Springer-Verlag, New York, New York, USA.
instruments and the lining of concert halls. Material property RUJINIRUN, C., P. PHINYOCHEEP, W. PRACHYABRUED, AND N. LAEMSAK. 2005.
charts that plot acoustic properties such as the speed of sound, Chemical treatment of wood for musical instruments. Part I.
the characteristic impedance, the sound radiation coefficient, Acoustically important properties of wood for the ranad (Thai
and the loss coefficient against one another for various woods traditional xylophone). Wood Science and Technology 39: 77–85.
1448 A MERICAN J OURNAL OF B OTANY [Vol. 93

SADIE, S., AND J. TYRELL [EDS.]. 2001. The new Grove dictionary of music WEGST, U. G. K. 2004. Natural Materials Selector, created using the CES
and musicians, 2nd ed. Oxford University Press, New York, New Constructor Software, Granta Design, Cambridge, UK.
York, USA. WEGST, U. G. K., AND M. F. ASHBY. 1996. Alternative woods for violin
SEEBERGER, F. 2003. Klangwelten der Altsteinzeit [music CD]. Urge- bows. Newsletter of the British Violin Making Association 5: 7–14.
schichtliches Museum Blaubeuren, Blaubeuren, Germany. WOOD HANDBOOK. 1999. Wood handbook: wood as an engineering
SEGERMAN, E. 1996. Wood structure and what happened in the Hunt & material. General Technical Report 113. U.S. Department of
Balsan experiment. Fellowship of Makers and Researchers of Agriculture, Forest Service, Forest Products Laboratory, Madison,
Historical Instruments Quaterly 84, Communication 1471: 53–55. Wisconsin, USA.
SEGERMAN, E. 2001. Some aspects of wood structure and function. Journal WOODHOUSE, J. 1993a. On the playability of violins. Part 1. Reflection
of the Catgut Acoustical Society 4: 5–9. functions. Acoustica 78: 125–136.
SELL, J. 1989. Eigenschaften und Kenngrössen von Holzarten. LIGNUM,
WOODHOUSE, J. 1993b. On the playability of violins. Part 2. Minimum bow
Schweizerische Arbeitsgemeinschaft für das Holz, Zürich, Switzer-
land. force and transients. Acoustica 78: 137–153.
SKUDRZYK, E. 1980. The mean-value method of predicting the dynamic YANO, H., Y. FURUTA, AND H. NAKAGAWA. 1997. Materials for guitar back
response of complex vibrators. Journal of the Acoustical Society of plates made from sustainable forest resources. Journal of the
America 67: 1105–1135. Acoustical Society of America 101: 1112–1119.
VON HORNBOSTEL, E. M., AND C. SACHS. 1914. Systematik der Musikin- YOKOCHI, K., AND A. YOSHIMOTO. 2004. Nijo Castle. Website http://www.
strumente. Ein Versuch. Zeitschrift für Ethnologie 46: 553–590. kyopro.kufs.ac.jp/dp/dp01.nsf/ecfa8fdd6a53a7fc4925700e00303ed8/
WAGENFÜHR, R., AND C. SCHREIBER. 1989. Holzatlas, 3rd ed. VEB Fach- [Accessed 28 February 2006].
buchverlag, Leipzig, Germany. ZIERL, B. 2005. Obtaining the perfect violin sound: with fungi. Website
WEGST, U. G. K. 1996. The mechanical performance of natural materials. http://www.empa.ch/plugin/template/empa/981/40307/—/l¼2 [Ac-
Ph.D. thesis. University of Cambridge, Cambridge, UK. cessed 28 February 2006].

View publication stats