Studies Addressing Piano Voodoo of Tone Production

In July 2010, I wrote about The Piano Voodoo of Tone Production, which demonstrates that pianists cannot control timbre of a single note independently of volume because of the physics of piano escapement/letoff.  There are those who claim we can control the tone of a note without changing its volume, and some of these claims are made in a scientific manner.  The studies I discuss below are not necessarily for or against these claims, but rather investigations of any possible science behind it.  There are many other related studies that could be included here, but I have chosen the ones that would seem to be the most relevant to the topic.

Identifying Stroke Type By Listening to Tones With and Without Pre-Tone Sounds

Goebl, Bresin and Galembo did a study published in 2004, and the presentation (given at the International Symposium on Musical Acoustics) was titled Once Again: The Perception of Piano Touch and Tone. Can Touch Audibly Change Piano Sound Independently of Intensity?

Testers sought to find out if listeners could identify “struck” notes (attacked from above) from “pressed” notes (finger resting on key before stroke begins) with and without precursor noise.  When precursor noises were removed from recordings, listeners could not identify struck tones any better than random chance – they tended to give “struck” ratings for louder tones.  When attack noises were intact, even with the help of intense listening concentration (listeners found the test to be “quite difficult throughout”), only half of the listeners could identify struck tones most of the time, while the other half could not identify struck tones more than random chance.

While the 50% where were able to identify the tones might seem promising at first, there are two problems with this:  1) these finger and fingernail attack noises cannot be heard unless a listener is sufficiently close to the piano and there is sufficient silence preceding the tone, and most importantly, 2) these noises are “context clues” that precede a tone, rather than being part of the tone itself.  We cannot consider extraneous noises from the pianist that play no part in the acoustical sound envelope (noises that do not resonate in any humanly discernable way inside the piano) to be part of “tone production” of the instrument.  Those who conducted this study also concluded that the pure aural effect of touch noises is “a rather small one.”

Interestingly, the precursor noises were described by the testers to be “salient” in the recording despite the microphone being located 10 centimeters away from the strings, and it surprised testers that more listeners could not identify stroke types using precursor noise.  They “speculate” that visual cues would greatly improve the listener’s ability to discern the performer’s stroke type.  Unfortunately, this would only show that harshness can be seen rather than heard.  It would be similarly obvious to “speculate” that subjects of a telepathy study could more effectively identify what a subject is thinking if the subject draws pictures of their thoughts and speaks their thoughts out loud.

Hammer Motion After Escapement

A. Askenfelt conducted a study titled “Measuring the motion of the piano hammer during string contact” (KTH Dept. for Speech, Music and Hearing Quarterly Progress and Status Report, Vol. 32 No. 4, 1991).

In this study, Askenfelt took precise measurements of the hammer’s motion and acceleration after it passed the point of escapement, using various touches (legato, staccato with relaxed finger, and staccato with strained finger).  Each of these touches produced a different vibration frequency on the hammer shank (the “stick” the hammer is attached to):

  • Legato:  50 Hz
  • Staccato with relaxed finger:  250 Hz
  • Staccato with strained finger:  600 Hz

The hammer shank “softness” was also varied.  Wood was removed from the original hammer shank so that the wood held on to its original attributes (same length, same density), decreasing its vibration frequencies by about 50%.

In the grand piano, these vibrations caused movement in the hammer due to (1) vertical vibration (vibration toward and away from the string), and (2) horizontal vibration (vibration parallel to the string).  Presumably, the vertical vibrations could affect the speed at which the hammer strikes the string even when hammer velocity at the point of escapement is held constant.  As for the horizontal vibrations, they could affect a number of things:

  1. The striking location of the hammer on the string;
  2. The direction the hammer is pointing when it strikes the string (affecting the part of the hammer that hits the string);
  3. The distance the hammer rubs across the string itself during the time it is in contact with the string.

While all of these mechanisms indeed take place, the study found that with exception to legato touch using a soft hammer shank, the differences these factors made were not differences that are even discernable by the human ear.  The study also discloses that it was difficult to control for exact precision in keystroke since measurements of actual keystrokes were not made (only measurements in hammer trajectory).

What the study does not discuss, however, is the likelihood that a pianist might be able to actually use (consciously or subconsciously) these phenomenons to their advantage.  Even if there is indeed an audible difference using a legato touch and soft hammer:

  1. Increasing or decreasing hammer velocity (increasing or decreasing volume of the legato stroke) absolutely cannot produce a predictable result in hammer shank vibration.  For example, suppose a note is played soft, and when the hammer strikes the string, the hammer is moving slightly faster since the hammer shank was vibrating toward the string at the instant it strikes the string.  If one plays mezzo-piano instead of piano, the time it takes the hammer to travel to the string is shortened, and the note is louder.  But because of the decreased time it takes for the hammer to travel to the string, who knows what direction the hammer shank vibration is moving at the contact point (remember, we’re dealing with around 25 vibrations per second, at the slowest).  The hammer shank vibration may actually be moving the hammer away from the string instead of toward the string, counteracting some of the pianist’s effort to make the note louder.  And if the pianist again increases the volume by playing mezzo-forte, the opposite may happen again:  the hammer shank vibration may be moving the hammer away from the string at the point of contact.  For this reason, if this effect is indeed audible, it will not enhance a pianist’s performance.  It will have a randomizing effect on the pianist’s performance, which actually undermines the pianist’s efforts to control tone.
  2. The study does not address whether loudness of a legato touch (which is the same as quickness of a legato touch) could change the frequency of the hammer shank vibration.  I think it would.  If so, this would further randomize tone production, again undermining the pianist’s efforts to control tone.  A 1990 lecture by A. Askenfelt and E. Jansson presents a theorem that “string resonances with period times essentially shorter than the contact duration will be only weakly excited by the blow” (“String contact duration and dynamic level” section).  The theorem makes perfect sense, and I think a similar theorem would apply to the resonance of the hammer shank, affected by the speed of acceleration of the key.  The quickness of acceleration of the key would be analogous to the shortness of time the hammer is in contact with the string.
  3. Hammer shanks from one piano to another vary in length, circumference and density, meaning that the frequency of hammer shank vibration will vary from one piano to another.  A pianist will not be able to rely upon any one technique for their tone production if the only purpose of the technique is to optimize timing of hammer shank vibration.
  4. Even without nos. 1-3 above, non-legato touches produce frequencies in the realm of 250 Hz or more.  If a pianist attempts to strike a key staccato (mezzo-forte) twice with the exact same sound each time, no pianist exerts the amount of precision necessary to be able to control the outcome of something vibrating 250 times per second.  Even the outcome of the soft hammer shank legato touch (25 or 50 Hz) would be difficult to rely upon.

The limited-scope conclusions in the study are reasonable and appropriate, but if we are to expand the scope of the study to include usefulness in performance, I can only conclude that pianists are best served realizing that the creation of great variety in color at the piano is indeed possible by mere manipulation of hammer velocity, timing, duration and pedals – nothing more.  Certain hand positions and “tone-controlling” techniques are useful, but not for the reasons that some pianists think.

Also consider that even good artists cannot discern between a keystroke that is produced “hitting” the key from above the key vs. a stroke that is produced with the finger already touching the key, as long as the “thud” of the fingernail against the key is removed from the recording (see Stroke Types study above).  Identifying a tone via fingernail clicks on the keyboard is not judging tone, it’s judging peripheral noise that takes place before the tone even begins.  If highly trained pianists can’t discern the difference between these two different strokes which are about as drastically different keystrokes as pianists can possibly produce (without actually changing the “volume” of the note), how would they be able to discern between strokes and effects that are more subtle?  By design, the study utilized extremely different strokes that everyone must agree would be the easiest to differentiate if the science behind tone control (independent of velocity control) worked.  When trained musicians fail to identify these extremely different strokes by their sound, the logical conclusion is that more subtle differences wouldn’t be discernable either.

“The Pianist and the touch”

In the Askenfelt/Jansson lecture referenced above, one section of the lecture discusses the thumping sound that a key makes when it hits the bottom of the keybed (“The pianist and the touch“).  An audio recording on this page features a note played four times with the thumping sound removed, then four more times with the thumping sound intact.  The note without the thumping sound was very different:  it made the piano almost sound like a guitar, giving the note a “plucked” sound.  Piano manufacturers are very aware of this phenomenon, even being very careful to select the right wood for the bottom of the keybed in order to achieve the most desirable thumping sound.

While I agree with the conclusion that Askenfelt and Jansson reach, namely that thumping of the key against the keybed is perhaps the biggest candidate for manipulating tone independent of velocity, there are some important things to note that they don’t mention in their lecture.  Keep in mind during my comments that thumping sound will be more minimal and perhaps even inaudible at softer volumes, so I don’t believe changes in thumping at soft dynamic levels would have much impact on the “tone” of the piano, if any at all.  So, my notes that follow assume dynamic levels of mezzo-forte and above.

  1. I don’t believe keybed thumping can be controlled independently of loudness.  I am extremely skeptical that any pianist would be capable of playing a note fortissimo (or even mezzo-forte) without the key hitting the bottom in a “percussive” kind of way1.  A loud sound can only be achieved with a certain hammer velocity, and this minimum hammer velocity can only be achieved with a minimum key velocity.  This minimum key velocity can only be achieved with a minimum finger acceleration.  This acceleration occurs in a matter of 30 milliseconds, or 0.03 seconds.  That is the time it takes for a key to depress all the way during a normal keystroke.  I do not believe any human is capable of perceiving this time interval with such clarity that they could stop or even slow down their finger immediately once the desired velocity is achieved.  Initially, I didn’t believe it purely based on reason, and my disbelief was hugely reinforced after I fussed at the piano with every pianistic technique in the book to no avail, always unable to change the vigor with which my finger stopped at the bottom of its descent.  I don’t think this last-minute finger slowing can take place to even the slightest degree:  the only way I was able to get my finger to not hit the bottom of its descent as vigorously was to simply play softer.
  2. Even if the pianist could slow down or stop their finger, I believe the key itself will continue downward with only minimal slowing due to key resistance.  By the time the key is accelerated to the required velocity to achieve a loud sound, most of the 0.03 seconds has already elapsed.  What little time is left for the hammer to slow down will surely not produce a change in key velocity – and therefore a change in keybed thumping – that is humanly perceptible.
  3. Even if a pianist could exert great control over keybed thumping independently of loudness, this discussion does not concern manipulation of piano timbre by virtue of how the hammer hits the strings.  This contradicts what has been (in my experience) the primary assertion of those who believe piano tone can be manipulated independently of loudness:  that it is due to some kind of control that pianists have over hammers after hammers have crossed the escapement point.
  4. Having said all this, I would still be very interested to see a theoretical study that looks at keys’ hitting the keybed the same way another study might look at hammers hitting strings.  If a pianist were able to propel a key downward at a forte volume without following the key to its bottom (if that’s even possible, which I doubt, other than using a “key dusting” technique necessary for repeated notes or leggiero playing, which doesn’t happen too often in piano playing), this would allow the key to bounce back up at its maximum possible speed.  The key would be in contact with the bottom of the keybed for a minimum amount of time, especially when played extremely loud.  I propose that higher frequencies in the keybed itself would be aroused, compared to when the pianist’s fingers follow the key all the way to the bottom.  I believe these higher frequencies would also resonate throughout the piano quicker than the lower frequencies since higher frequencies travel through material faster than lower frequencies (although the quickness of this resonance would be difficult or impossible for a listener to perceive).  More weight at key bottom would cause the key to be in contact with key bottom for a longer duration, which would have the effect of dampening or silencing higher frequencies of keybed resonance.

1 This could be possible if one resorts to what I call “key-dusting” technique, where one pulls the fingers toward the palm to achieve extremely rapid repeating notes or very staccato notes in a leggiero passage.  This kind of technique is not used very frequently in musical performance, and certainly this is never what pianists are referring to when they make the sweeping claim that timbre can be controlled independently of volume.  I’ve never heard of a single pianist anywhere who uses this key-dusting technique for any purpose other desire for a certain staccatissimo articulation:  it is used for articulation, not for some kind of change in timbre.  Furthermore, if my proposition in #4 is correct, this technique would produce a keybed thumping sound that is more, not less, obnoxious, and yet pianists utilize this technique most commonly in the context of playing leggiero, achieving a “light” effect.  Any change in keybed thumping sound due to this changed technique is apparently of little or no concern to pianists since pianists are still achieving light, non-percussive sounds in leggeiro passages even when keybed thumping is at its worst.


Askenfelt, A. “Measuring the motion of the piano hammer during string contact.” STL-QPSR (Deptartment for Speech, Music and Hearing KTH Sweden: Quarterly Progress and Status Report), 1991.

Askenfelt, Anders, and Erik Jansson. “From touch to string vibration.” Five Lectures on the Acoustics of the Piano. Östen Häggmark, ed., May 27, 1988. Web. Accessed May 8, 2014. <>

Goebl, Werner, Roberto Bresin, and Alexander Galembo. “Once again: The perception of piano touch and tone. Can touch audibly change piano sound independently of intensity.” Proceedings of the International Symposium on Musical Acoustics. 2004.

Twedt, Chad. “Studies Addressing Piano Voodoo of Tone Production.” Cerebroom, Oct. 18, 2010. Web. Accessed May 8, 2014. <>

(c) 2010 Cerebroom

About Chad

Chad is a pianist, composer, piano teacher and blogger with a Masters Degree in Piano Performance. He received the 2005 Nevada Arts Council Fellowship Grant for the composing and performing on his Ostinato CD.
Science ,

Leave a Reply