Sie sind auf Seite 1von 9

Ru Lemer

University of Surrey - Department of Music & Sound Recording


TON201954386 - Philip Coleman
November 2017

The Acoustics of Piano Tuning


Abstract

Recent and historic research on the acoustic theories describing different aspects of
piano tuning are reviewed. A workflow of the tuning process is given, with
explanations of temperament, inharmonicity, phantom partials, the Railsback curve,
tuning unisons, string coupling and false beats. These phenomena are all relevant to
the tuner and if understood could aid enthusiasts in understanding the complex craft
of piano tuning. A brief summary of modern electronic tuning systems is given and
suggestions are made as to possible further research and continuations.

1. Introduction 2.0 The tuning process


T h e p i a n o f o r t e h a s p l a ye d a The generalised goal of the piano
prominent part in Western classical and tuner is to make the instrument sound as
popular music since its invention in the “good” as possible. This is achieved by
1700s. Contributions to its present design maximising consonance in accordance with
have been made by numerous human perception, while still making the
manufacturers over the years, often through instrument versatile enough to be played in
empirical experiments, and without all keys.
assistance from scientists [Dolge 1911 p. Whilst there are many discrepancies
107]. As such, scientific discoveries are still between piano tuners as to exact workflow,
being made as to why certain design the process can be broken down into this
decisions provide the most pleasing results to simplified order of events:
pianists and listeners.
The process of tuning a piano is 1. Choosing a pitch standard. (The
unique to the instrument, and requires a majority of modern music uses the
dedicated “piano tuner” who has typically internationally agreed pitch standard of
gone through at least 3 to 7 years of
A4 being 440Hz [Karp 1983 p. 407],
dedicated training. As with the instrument’s
design, there is a rich history of different this is known as “concert pitch”.)
techniques and workflows with regards to 2. Tuning a single note using the pitch
tuning the instrument. Only recently has the
standard as a reference.
scientific community been able to quantify
and explain some of the more intricate 3. “Setting the temperament” using the
practices. The purpose of this paper is to tuned single note (see section 4).
review historical and recent acoustic piano
theory and its relationship with existing 4. Tuning the upper and lower ranges of
piano tuning techniques. the piano to the temperament.

ARS PAPER !1
2.1 Partial theory “Setting the temperament” refers to
tuning all notes within the tuner’s chosen
All vibrating things produce a series of range to their exact theoretical pitch, so that
vibrational components. The first of these is the rest of the piano may be tuned to it.
the fundamental, and all higher components The range of the “temperament” varies
are known as overtones. If the overtones’ with the preference of the tuner, but
frequencies are integer multiples of the traditionally it is one octave between F3 and
fundamental frequency, then they follow the F4.
harmonic series and are called harmonics Before the widespread standardisation
[Young 1952]. The fundamental and the of the equally tempered scale,
overtones together are known as partials. temperaments were used that allowed
certain intervals to be pure. An interval is
3. Piano Strings classed as pure when it has a small-integer
frequency ratio. These temperaments were
In order for piano manufacturers to known as just intonation [Johnston 2006
equalise tone over the range of the piano p.xiv].
within reasonable limitations, treble strings The combination of a sine wave
are made of steel wire, and bass strings are whose fundamental is at 100Hz and a sine
made of steel wire wrapped with a copper wave whose fundamental is 200Hz will be
or iron winding. This effectively decreases pure, this is because the ratio of the two
the thickness of the bass strings whilst tones is exactly 2:1. This ratio is known as
maintaining their linear density, allowing an octave. The reason this sounds consonant
them to vibrate at low frequencies even is because the partials of sine waves are
though they are disproportionately short. exact harmonics, so the partials of each
The notes in the lower-mid register of tone will precisely coincide, resulting in a
the piano have 2 strings each (“bichord”), harmonically simple combined tone. It is
and the notes in the upper register have 3 agreed across all mainstream historic and
strings each (“trichord”). If multiple strings modern tuning systems that the octave is the
are tuned to the same pitch and struck, the purest interval of the 12-note scale and must
sound produced by one reinforces the others remain as pure as possible.
and the combined tone is louder. This is When tuning most other instruments,
known as “constructive the tuner relies on their own conception of
interference” [Reblitz 1981 p. 16]. pitch for accuracy. Counterintuitively, pitch
Once one string in a bichord or cannot be used as a reference point for
trichord note is tuned to the right pitch, the piano tuning, it is merely a rude guide. Beat
other string(s) are then tuned to the rates are the primary cue for obtaining
frequency of the original string in a process absolute accuracy in pitch placement of
known as “tuning unisons” [Maneri, 2017]. piano notes. For example, if one tone is
tuned to 100Hz and another to 102Hz, then
the resulting difference would produce an
4. Temperament and Tuning audible amplitudinal “beat” of 2Hz.
Systems Pythagorean tuning is a system of just
intonation based on the ratio of 3:2. This
Collins English Dictionary [2017] ratio is known as the pure perfect fifth. A
defines a temperament as “a system of piano tuned with this system will be very
adjustment of the intervals between the consonant for specific intervals, but will
tones of an instrument with fixed result in musically unusable dissonances in
intonation”. other intervals.

ARS PAPER !2
Since the 17th century, the standard supports at either end so that every wave
tuning system for pianos has been to use an behaves in the same way. In practise, the
equal temperament. This system puts thickness and tension of piano strings cause
priority on the octave being the purest a mechanical resistance whose effect
interval, and the 12 fixed tones within the b e c o m e s m o re p ro n o u n c e d a s t h e
octave are equidistant and impure. This wavelength of the wave diminishes and
means that any chord will be as harmonious approaches the thickness of the string. A
and consonant in one key as in another in rough descriptor of this effect is given by
spite of the general dissonance within the
harmonic structure. f! n ≈ n f1 (1 + α n 2)
Thus, the frequency ratio of 2
adjacent semitones can be represented by where α! is proportional to the Young’s
Modulus of the string, and the value of f! 1 is
fn+1 12 higher than the value f! 01 found for the ideal
! = 2 ≈ 1.05946
fn string. Whilst this equation is incomplete, it
can still be seen that there is a correlation
This ratio can also be expressed as 100 between the partial number and the
cents. deviation of said partial from its associated
Equal temperament will result in all harmonic. Fletcher [1964] gives a more in-
intervals being equally impure, allowing the depth mathematical explanation of string
pianist to modulate between keys freely behaviour.
without the limits of a just temperament. A string under ideal conditions
vibrates periodically; this means that the
oscillations follow a regular pattern.
5. Inharmonicity Inharmonicity causes strings to develop
complex vibrations that are not periodical,
Under ideal theoretical conditions; a
and this significantly impacts the tone of the
string has modes of vibration whose
instrument. Fletcher et al. [1962] stated that
frequencies form partials in a harmonic
inharmonicity is responsible for the
series. Therefore if the frequency of the
fundamental component is f! 01 ; then the nth perceived warmth of a piano’s tone, and it is
this that gives the instrument its distinctive
harmonic can be described as:
“piano quality”. Namgyal [2014] quantified
this study and found that synthesised piano
f! n = n f01
tones without inharmonicity were perceived
as less ‘warm’ than inharmonic ones by a
Where n is an integer. control group of experienced listeners.
Railsback found in 1937 that there The human ability to detect
was a marked trend in the “stretching” of mistunings has been shown to diminish
octaves by expert piano tuners. Young progressively with both higher partial
[1952] continued this research and proved numbers and increasing fundamental
that the partials in piano strings deviate frequency [Järveläinen et al. 1999]. Thus
from the harmonic series due to a number inharmonicity becomes less perceivable and
of mechanical factors. This departure of less important to the tuner towards the
partials from the harmonic series is known higher end of the piano where there is a
as inharmonicity. greater frequency gap between partials.
When a piano string is struck, a rich This is clearly visible in the spectra of
and complicated combination of waves are related piano tones. Figure 1 [Giordano,
produced that vary in wavelength and 2015] shows a comparison of treble notes
amplitude. An ideal string would be A5 and A6 on a standard equal tempered
infinitely thin and flexible, and have rigid

ARS PAPER !3
Fig. 1 These notes can be seen to produce a far
more complex cluster of partials, most of
which are of a higher intensity than the
A5 fundamental. For A1, the fundamental is
barely discernible in comparison with the
power (arb. units)

partials, and it can be seen that there are


considerably more coinciding peaks between
the two notes.
The inharmonicity of piano strings
causes complications to arise when tuning to
A6 an equal temperament. As stated earlier, the
goal of the tuning process is to make the
instrument sound as ‘good’ and as
consonant as possible in its own right. Piano
players often play with a tessitura (range) of
500 1000 2000 5000 at least 3 octaves, so consonant
f (Hz) considerations must be made towards the
overlapping partials of wide-ranging
Fig. 2 intervals if the piano is to sound in tune
with itself. To use Figure 1 as an example; if
A5 is to sound in tune with A6, then the
second partial of A5 and the fundamental
A1 of A6 must be tuned to be beatless. As the
power (arb. units)

second partial of A5 is sharpened by


inharmonicity, this will result in the
fundamental of A6 being slightly higher in
pitch than its theoretical value of 1720Hz.
A2 When considering the consonance
between complex low tones like A1 and A2,
it is much more challenging to minimise
dissonance. This is due to the large amount
of overlapping partials within a given
25 50 100 200 500 1000 2000 bandwidth that all result in individual
beating patterns. While the first overlapping
f (Hz)
partial (the second partial of A1 and the
fundamental of A2) still has to be
piano. It can be seen that the partials are considered, the proceeding partials also
distinct and that they progressively dissipate influence the ideal tuning between the notes.
in energy. In this case the crucial overlap Therefore, for lower notes, the first
that must be taken into consideration by the overlapping partial sometimes must be
piano tuner is between the second partial of mistuned in order for the overall sound to
A5 and the fundamental of A6. This be consonant. Maximum consonance
relationship will define the sensory between the notes will be achieved when
dissonance between the two notes. The some function of the beat rates of multiple
fourth partial of A5 and the second partial partials is minimised. This function was
of A6 is of a considerably lower magnitude recently quantified by Giordano in 2015.
and therefore much less relevant to the Studies have reliably shown than a
consonance of the interval. piano is more pleasing to the ear when
Contrast this with Figure 2, which tuned with octave stretching in consonance
shows a spectra of bass notes A1 and A2.

ARS PAPER !4
with its own notes, rather than tuned 5.2 Phantom Partials
precisely to an equal temperament.
Listeners expect an octave to be roughly 10 It can clearly be seen in Figure 2 that
cents greater than a 2:1 frequency ratio there are more partials than one would
[Burns and Ward, 1982] because of how expect from a pure tone. The harmonic
embedded the impacts of inharmonicity are series of A1 within the range of 25Hz to
in Western piano music. 2000Hz consists of 110Hz, 220Hz, 440Hz,
880Hz and 1720Hz. The interspersed peaks
5.1 The Railsback Stretch between these frequencies in Figure 2 are
known as “phantom partials”.
Railsback carried out a series of tests Phantom partials (also known as
in 1938 on a selection of well tuned pianos secondary partials) are the byproduct of
tuned by various expert piano tuners. He existing primary inharmonic partials, and
found that there was a universal deviation have been classified by Conklin [1997] as
from the equal tempered theoretical either “even” or “odd”. “Even” phantom
placement of notes across all tested pianos. partials appear in the harmonic series of
Figure 3 shows his results against that of a existing primary inharmonic partials, whilst
well tuned single piano. it is the sums of these partials that can
The curve shows how inharmonicity is produce “odd” phantoms. It should be
increasing as the tested notes grow more noted that the exact occurrence and
distant from the set temperament. High intensity of phantom partials is individual to
piano strings are shorter but maintain the the piano in question.
same thickness as mid-range piano strings,
and low strings are thicker but maintain a
similar length. It can be seen in both cases
how string stiffness has more of an impact
on the tuning than in the mid-range strings.

Fig. 3

ARS PAPER !5
6. Tuning unisons study shows that fine unison tuning is not so
much about setting the beats rate as it is
In the case of the trichord about regulating the amount of aftersound.
temperament octave; once the exact tuning The aftersound also is very much
has been set for the first string of a note, the affected by irregularities in the head of the
other strings must then be tuned to a hammer, this causes different strings within
unison. Typically the middle string is tuned notes to be struck slightly differently. It was
first, followed by the two outer strings Weinreich’s conjecture that expert piano
[Maneri 2017]. Hundley et al. [1956] have tuners are able to adjust unisons to
shown that the complex phasing of partials compensate for hammer irregularities.
in the piano tone envelope is mainly due to The Una Corda pedal (“soft pedal”) on
small differences in the tuning of grand pianos is also relevant to tuning
fundamentals between unison strings, so it is unisons. The basic function of this pedal is
clear that some difference is musically to shift all hammers to the right by one
desirable. Kirk [1959] quantified this by string. The result of this is that when the
performing a paired comparison preference pianist plays a trichord note, only the two
test with 123 subjects. He found that the rightmost strings will be struck. When a
most preferred tuning condition for 3 unison bichord note is played, only a single string is
string groups are 1 and 2 cents maximum struck. Weinreich showed that the Una Corda
deviation among strings, with a clear pedal had a considerable effect on the
preference against strings tuned to exactly aftersound of these notes. It would therefore
the same frequency. be logical for the tuner to be considerate of
A preference of some tuners is to tune the relationship and coupling between the
the leftmost strings of a trichord group second and third string within a trichord
slightly flat, and the rightmost string slightly group.
sharp [Yamaha n.d.]. This author could find
no further research as to whether it is more 6.2 False beats
tonally pleasing having the strings tuned in
ascending pitch from left to right in a “False beats” is a common term in
trichord note, or whether this arrangement piano tuning terminology, Capleton [2014]
is irrelevant. describes the phenomenon as an audible
beat pattern present in one or more partials
6.1 String coupling during the decay of a single string. As beat
patterns are a primary cue for determining
Martin [1977] suggested that the issue the exact tuning of a string in relation to
of unison tuning is further complicated by other strings, the interference from false
coupling between strings. He reported on beats can be problematic for the tuner, as
the sound amplitude from a piano note well as being distracting for the player.
dying away at two distinct rates, an initial False beats can be effectively “hidden”
“fast” decay turning into a secondary “slow” by utilising the other strings in a note to
decay known as “aftersound”. The length of cancel out problematic partials through very
the aftersound is dependant on the phase fine adjustments. It is possible that one of
relationship between multiple strings. If the the relevant phenomenon in aiding the
strings are in phase then the decay can last diminution of false beats is string coupling.
three times the length of a note with the
three strings out of phase. Due to slight
differences in the strings, the phase
relationship will slowly change over the
course of the note [Weinreich, 1977]. His

ARS PAPER !6
7. Modern technology theorised that with a fundamental
understanding of these acoustic
The use of electronic tuning devices phenomena, modern piano tuners could
has been increasing since their introduction apply these sciences to hastening their
in the 1970s, and now are available in app training time.
form on smart phones or tablets. This is an For example, it would be fairly
area of much contention between tuners, as straightforward to tune the temperament
it makes the act of piano tuning accessible using a standard guitar tuner, and then tune
to anyone with a smart phone. The the rest of the piano accordingly by
principle of modern apps is that they will focussing on the inharmonic partials
measure the inharmonicity present in relevant to the temperament octave. In
individual pianos and adjust their practise, there are complications to do with
calibrations accordingly. These apps and irregularities in individual pianos, as can be
devices can deliver results that are very close seen in Figure 3. On the single well tuned
to ideal, but Woerd [2010] found that they reference piano that is compared with the
still fall short in areas such as hammer Railsback Curve, there is considerable
inconsistencies and the complex optimal variation of individual notes from the
tuning of bass fundamentals when expected placement.
compared with an expert tuner working by There has always been a shortage of
ear. piano tuners compared with the amount of
Buda [1991] designed the first active pianos. Reblitz [1981] estimated that
automated system for adjusting tuning pegs. to tune every piano in the United States
His system accounted for stretch tuning and only once a year it would take about 20
the slight, optimum detuning of unison thousand technicians, and at the time there
strings. were only 5,000 qualified working
A system could be conceived that technicians. Since then the number of piano
would carry forward the research of tuning training facilities has fallen, so it is
electronic tuning devices into Buda’s appropriate for home owners to be able to
automated tuning system. Such a system perform basic maintenance and tunings on
could be fitted to concert grand pianos and their pianos with the aid of modern
keep them perpetually in tune, thereby technology.
removing the need for the piano tuning Piano tuning is much more accessible
profession and causing them all the start to the general population than it was even
working as gardeners. fifty years ago through the development of
apps and the widespread knowledge of
8. Summary and discussion theories through popular mediums like
Youtube.
For centuries piano tuners have tuned It is still uncertain as to whether the
pianos with no knowledge of some of the left-to-right arrangement of tuning trichord
acoustic topics covered above. The historic notes has an impact on the tone, or whether
expert tuner would have accounted for this is relevant when the Una Corda pedal is
inharmonicity naturally through ear in use. This is an area for further research.
training and experience, this would have Another area would be the potential
i nvo l ve d “ t u n i n g o u t ” i n h a r m o n i c relationship between string coupling and the
overlapping partials between intervals. The elimination of false beats.
next step would be compensating for false
beats, as well as adjusting unisons to balance
our irregularities in hammer heads to
maximise the perceived warmth. It could be

ARS PAPER !7
References
- Alfred, Dolge, 1911: Pianos And Their Makers (Covina, California: Covina Publishing
Company).
- Buda, Eric de, 1992: ‘Piano tuning system’, J. Acoust. Soc. Am., 91, 3596, .
- Burns, E.M. and Ward, W.D., 1982: The Psychology of Music (San Diego, California: D.
Deutsch, ed.).
- Capleton, Brian, 2004: ‘False beats in coupled piano string unisons’, J. Acoust. Soc. Am., 115,
885.
- Collins English Dictionary, 2017: ‘Temperament’ (Collins).
- Conklin, Harold A., 1997: ‘Piano strings and “phantom” partials’, J. Acoust. Soc. Am., 102,
659.
- Fletcher, Harvey, 1964: ‘Normal vibration frequencies of a stiff piano string’, J. Acoust. Soc.
Am., 36, 203, .
- Fletcher, Harvey, Blackham, E. Donnell and Stratton, Richard, 1962: ‘Quality of Piano
Tones’, J. Acoust. Soc. Am., 34, 749, .
- Giordano, N., 2015: ‘Explaining the Railsback stretch in terms of the inharmonicity of
piano tones and sensory dissonance’, J. Acoust. Soc. Am., 138, 2359.
- Hundley, T. Chase, Martin, Daniel W. and Benioff, Hugo, 1956: ‘Factors Contributing to
the Multiple Decrement of Piano Tone Envelope’, J. Acoust. Soc. Am., 28, 769.
- Järveläinen, Hanna, Välimäki, Vesa and Matti, Karjalainen, 1999: Audibility of
Inharmonicity in String Instrument Sounds, and Implications to Digital Sound Synthesis,
Proc. Int. Computer Music Conf. (ICMC’99).
- Johnston, Ben, 2006: “Maximum Clarity” and Other Writings On Music (University of Illinois
Press).
- Maneri, Sal, International Piano Tuner, 2017: Personal correspondence with the author,
Conversation, 12th January 2017.
- Martin, Daniel W., 1947: ‘Decay Rates of Piano Tones’, J. Acoust. Soc. Am., 19, 722.
- Namgyal, Tashi, 2014: The effect of inharmonicity on the perception of warmth of synthetic piano
tones. Technical Project. (University of Surrey).
- Karp, Theodore, 1983: Dictionary of Music (Northwestern University Press).
- Kirk, Roger E., 1959: ‘Tuning Preferences for Piano Unison Groups’, J. Acoust. Soc. Am., 31,
1644.
- Railsback, O.L., 1938: ‘Scale Temperament as Applied to Piano Tuning’, J. Acoust. Soc. Am.,
9, 274.

ARS PAPER !8
- Reblitz, Arthur A., 1981: Piano Servicing, Tuning, & Rebuillding For The Professional, The Student,
The Hobbyist (Vestal, NY: The Vestal Press).
- Weinreich, Gabriel, 1977: ‘Coupled piano strings’, J. Acoust. Soc. Am., 62, 1474.
- Woerd, A.C. van der, 2010: ‘New Electronic Tuning Device for Pianos’, J. Acoust. Soc.
Am., 51, 1972.
- Yamaha, n.d.: https://www.yamaha.com/en/musical_instrument_guide/piano/
mechanism/mechanism004.html, accessed 09/11/2017.
- Young, Robert, 1952: ‘Inharmonicity of Plain Wire Piano Strings’, J. Acoust. Soc. Am., 24,
267.

ARS PAPER !9

Das könnte Ihnen auch gefallen