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

net/publication/2472699

The piano action as the performer's interface: Timing properties, dynamic

behaviour and the performer's possibilities.

Conference Paper · July 2003

Source: CiteSeer

CITATIONS READS

4 50

3 authors, including:

Werner Goebl Roberto Bresin

Universität für Musik und darstellende Kunst Wien KTH Royal Institute of Technology
105 PUBLICATIONS   1,420 CITATIONS    134 PUBLICATIONS   2,396 CITATIONS   

SEE PROFILE SEE PROFILE

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

Ljudskogen/Sound Forest View project

Interactive Sonification for High-Speed Trains View project

All content following this page was uploaded by Roberto Bresin on 12 February 2015.

The user has requested enhancement of the downloaded file.

 Proceedings of the Stockholm Music Acoustics Conference, August 6–9, 2003 (SMAC 03), Stockholm, Sweden
THE PIANO ACTION AS THE PERFORMER’S INTERFACE: TIMING PROPERTIES,
DYNAMIC BEHAVIOUR AND THE PERFORMER’S POSSIBILITIES
Werner Goebl1,2 , Roberto Bresin2 , and Alexander Galembo2
1
Austrian Research Institute for Artificial Intelligence (ÖFAI), Vienna
2
Department of Speech, Music, and Hearing (TMH)
Royal Institute of Technology (KTH), Stockholm
werner.goebl@oefai.at, {roberto,alex}@speech.kth.se
ABSTRACT
A concert pianist is able to produce a wide range of imaginable nu-
ances of musical expression by actuating the 88 keys on a piano,
none of which travel through a distance greater than one centime-
ter. In this study, we investigated the temporal behaviour of grand
piano actions from different manufacturers using different types
of touch (‘legato’ versus ‘staccato’). An experimental setup con-
sisting of accelerometers and a calibrated microphone was used to
capture key and hammer movements, as well as the acoustic sig-
nal. Five selected keys were played by pianists with the two types
of touch. The analysis of the three-channel data was automated
by computer software. Discrete measurements (e.g., finger–key,
hammer–string, and key bottom contact times, hammer velocity)
were extracted for each of the over 4000 recorded tones in order to
study several temporal relations. Travel times of the hammer (from
finger–key to hammer–string) as a function of hammer velocity
varied clearly between the two types of touch, but only slightly
between pianos. A travel time versus hammer velocity function
found in earlier work [1] derived from a computer-controlled pi-
ano was replicated. Key bottom contact times exhibited larger vari-
ability between types of touch and pianos. However, no effect of
touch type was found in the peak sound level (in dB as a function
of hammer velocity).
1. INTRODUCTION
The grand piano action is a highly elaborate and complex mechan-
ical interface, whereby the time and the speed of the hammer hit-
ting the strings is controlled by varying the manner and the force
of striking the keys. Its temporal parameters have been studied in
detail [2, 3]. However, only exemplary data was reported so far.
The present study aims to collect a large amount of measure-
ment data from different pianos, different types of touch, and dif-
ferent keys, in order to determine benchmark functions for perfor-
mance research. The measurement setup with accelerometers was
the same as used by [4], but the data processing procedure was
automated with specially developed computer software in order to
obtain a large and reliable data set. With the present setup, vari-
ous temporal properties (travel time, key bottom time, time of free
flight) and acoustic properties (peak sound level, rise time) were
determined, only a few of which can be discussed here.
In a performance study on melody lead [1], finger–key onset
times were inferred from the hammer–string onset times through
an approximation of the travel times of the hammer (from finger–
key to hammer–string contact) as a function of hammer velocity.
159
This travel time function was approximated from data of an inter-
nal chip of a Bösendorfer SE290 reproducing system. The present
study also aims to reconsider that approximation.
2. METHOD
2.1. Material
Three grand pianos by different manufacturers were measured in
this study: A Steinway grand piano, model C, 225 cm1 , a Yama-
ha Disklavier grand piano DC2IIXG, 173 cm2 , and a Bösendor-
fer computer-controlled grand piano SE290, 290 cm3 . Imme-
diately before the experiments, all three instruments were tuned,
and the piano action and—in the case of the computer-controlled
pianos—the reproduction unit serviced.
2.2. Equipment
The tested keys were equipped with two accelerometers: one fas-
tened on the key4 and one on the bottom side of the hammer shank.5
Each of the accelerometers was connected with an amplifier6 with
a hardware integrator inside. Thus, their output was velocity in
terms of voltage change. A sound level meter (Ono Sokki LA–
210) placed above the strings of that particular key (approximately
10 cm distance) picked up the sound. The velocities of the key
and the hammer as well as the sound were recorded on a multi-
channel digital audio tape (DAT) recorder (TEAC RD–200 PCM
data recorder) with a sampling rate of 10 kHz (16 bit). The DAT
recordings were transferred onto computer hard disk into multi-
channel WAV files (with a sampling frequency of 16 kHz). Further
evaluation of the recorded data was done in Matlab programming
environment with routines developed for this purpose (by the first
author). The recording sessions were preceded by a calibration
1 Situated at KTH-TMH in Stockholm (#516000, built in Hamburg, ap-
prox. 1992; this particular grand piano was already used in [5]).
2 Situated at the Dept. of Psychology at Uppsala University (#5516392,
built in Japan, approx. 1999; this series issued 1997 by Yamaha).
3 Situated at the Bösendorfer Company in Vienna (#290–3, built in Vi-
enna in 2000). The Stahnke Electronics (SE) system dates back to 1983 [6],
but this particular grand piano was built in 2000. The same system used
to be installed in an older grand piano (internal number 19–8974, built in
1986), but was put into a newer one for reasons of instrumental quality.
4 Brüel & Kjær Accelerometer type 4393; 2.4 g; #1190913.
5 ENDEVCO Accelerometer Model 22; 0.14 g; #20845.
6 Brüel & Kjær Charge Amplifier Type 2635.
 Proceedings of the Stockholm Music Acoustics Conference, August 6–9, 2003 (SMAC 03), Stockholm, Sweden

STEINWAY C YAMAHA DISKLAVIER BÖSENDORFER SE290

−0.7147 −0.5959 lg st rp
250 lg = 98.57 × HV 250 lg = 89.41 × HV 250 G6 (91)
st = 65.19 × HV −0.7268 st = 57.43 × HV −0.7748
C5 (72)
C4 (60)
200 200 200 G2 (43)
Travel time (ms)

C1 (24)

lg = 89.96 × HV −0.5595
150 150 150 st = 58.39 × HV −0.7377

100 100 100

50 50 50

0 0 0
0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7
Maximum hammer velocity (m/s)

Figure 1: Travel time (from finger–key to hammer–string contact) against maximum hammer velocity for the three grand pianos (three
panels), different types of touch (legato, staccato, and reproduction by the piano), and different keys (from C1 to G6, see legend). The two
types of touch (‘lg’–‘st’) were approximated by power functions (see legends). The travel time functions by the reproducing devices were
similar to those produced by staccato touch.

STEINWAY C YAMAHA DISKLAVIER BÖSENDORFER SE290

40 40 40
−0.3936 −0.4158 lg st rp
lg = 19.09 × HV −12.30 lg = 14.63 × HV −11.05
35 35 35 G6 (91)
st = 59.57 × HV −0.1131 −51.19 st = 10.15 × HV −0.6825 −3.74
C5 (72)
Key bottom times (ms)

−0.1639
30 30 rp = 16.20 × HV −12.47 30 C4 (60)
G2 (43)
25 25 25 C1 (24)

20 20 20 lg = 11.59 × HV −0.4497 −9.98

st = 13.96 × HV −0.3559 −10.15
15 15 15
rp = 10.09 × HV −1.1080 −2.09
10 10 10

5 5 5

0 0 0

−5 −5 −5

−10 −10 −10

0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7
Maximum hammer velocity (m/s)

Figure 2: Key bottom time relative to hammer–string contact against maximum hammer velocity. Negative key bottom values denote
instants precedinghammer–string contact. Legends list power curve fits of the data separately for legato (‘lg’), staccato (‘st’), and repro-
duction (‘rp’).

STEINWAY C YAMAHA DISKLAVIER BÖSENDORFER SE290

90 90 90
lg = 3.73 × HV −1.6850 lg = 3.72 × HV −1.7850 lg st rp
80 st = 6.76 × HV −1.7710 80 sthigh = 18.08 × HV −1.2880 80 G6 (91)
stlow = 6.71 × HV −2.1490 C5 (72)
Time of free flight (ms)

70 70 −1.1410 70 C4 (60)
rp = 4.04 × HV G2 (43)
C1 (24)
60 60 60
lg = 5.04 × HV −1.1870
50 50 50
st = 7.64 × HV −1.8300
rp = 17.60 × HV −1.8660
40 40 40

30 30 30

20 20 20

10 10 10

0 0 0
0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7
Maximum hammer velocity (m/s)

Figure 3: Time of free flight of the hammer. Time intervals between the point of maximum hammer velocity and hammer–string contact
are plotted against maximum hammer velocity. Power functions were approximated for each type of touch (‘lg’, ‘st’, ‘rp’). Two functions
were fitted for the staccato data of the Yamaha.

160
 Proceedings of the Stockholm Music Acoustics Conference, August 6–9, 2003 (SMAC 03), Stockholm, Sweden
STEINWAY C
120 120
110 110
100 100
Peak SPL (dB)
90 90
80 80
70 70
60 60
50 50
40 40
0 2 4 6
Maximum hammer velocity (m/s)
Figure 4: Peak dB–SPL values as a function of maximum hammer velocity for the three grand pianos (three panels), different types of
touch (legato, staccato, and reproduction by the piano), and different keys.
procedure which allowed conversion from voltage changes into re-
quired units like meters per second or dB–SPL.
2.3. Procedure
Five keys distributed over the whole range of the keyboard were
tested: C1 (MIDI note number 24), G2 (43), C4 (60), C5 (72),
and G6 (91)7 . The first two authors served as pianists to perform
the recorded test tones. Each key was hit in as many different
dynamic levels (hammer velocities) as possible, in two different
kinds of touch: once with the finger resting on the surface of the
key (‘legato touch’), once hitting the key from a certain distance
above (‘staccato touch’), striking the key already with a certain
speed.
Parallel to the accelerometer setting, the two reproducing pi-
anos recorded these test tones with their internal device on com-
puter hard disk (Bösendorfer) or floppy disk (Disklavier). For
each of the five keys, each player played in each type of touch
from 30 to 110 individual tones, so that a sufficient amount of
data was recorded. Immediately after each recording of a particu-
lar key, the recorded file was reproduced by the grand piano, and
the accelerometer data was recorded again onto the multi-channel
DAT recorder. For the Steinway, 595 individual sound events were
recorded, for the Yamaha Disklavier 1992, and for the Bösendorfer
1512 (including the reproduced keystrokes).
2.4. Data analysis
For each keystroke, several instants in time were defined as listed
below and obtained automatically from the recorded data with the
help of Matlab scripts.
The hammer–string contact time was defined as the moment
of maximum deceleration (minimum acceleration) of the hammer
shank (hammer accelerometer) which corresponded well to the
physical onset of the sound, and conceptually with the ‘note on’
command in the MIDI file. In mathematical terms, the hammer–
string contact is the minimum of the first derivative of the mea-
sured hammer velocity.
The finger–key contact time was defined to be the moment
when the key started to move. It was obtained by a simple thresh-
7 Only three keys were tested at the Steinway piano (C1, C5, G6).
YAMAHA DISKLAVIER BÖSENDORFER SE290
120
110
100
90
80
70
lg st rp
60 G6 (91)
C5 (72)
C4 (60)
50
G2 (43)
C1 (24)
40
0 2 4 6 0 2 4 6
old procedure applied to the key velocity track. In mathematical
terms, it was the moment when the (slightly smoothed) key accel-
eration exceeded a certain threshold which varied relative to the
maximum hammer velocity. Finding the correct finger–key point
was not difficult for staccato tones (they showed typically a very
abrupt initial acceleration). However, automatically determining
the correct moment for soft legato tones was sometimes more dif-
ficult and needed manual correction. When this procedure failed,
it failed by several tens of milliseconds—an error easy to discover
in explorative data plots.
The key bottom contact time was the instant when the down-
wards travel of the key was stopped by the keybed. This point was
defined as the maximum deceleration of the key (MKD). In some
keystrokes, the MKD was not the actual keybed contact, but a re-
bound of the key after the first key bottom contact. For this reason,
the time window of searching MKD was restricted to 7 ms before
and 50 ms after hammer–string contact. The time window was it-
eratively modified depending on the maximum hammer velocity
until the correct instant was found. MKD as indicator was espe-
cially clear and non-ambiguous when the key was depressed in a
range of medium intensity.
The maximum hammer velocity (in meters per second) was
the maximum value in the hammer velocity track before hammer–
string contact. An intensity value was derived by taking the max-
imum energy (RMS) of the audio signal immediately after ham-
mer–string contact, using a RMS window of 10 milliseconds. It
was calibrated in order to obtain dB–SPL.8 Data were controlled
and inspected for errors with the help of an interactive tool that
displayed simultaneously key and hammer velocity and the sound
signal for each keystroke.
3. RESULTS & DISCUSSION
Travel times are plotted against the maximum hammer velocity
in Figure 1 together with power curve approximations separately
for the pianos and type of touch. It can be seen that the travel
time functions varied considerably between the types of touch (the
difference was of the order of 20 ms in the middle range), but
only marginally between pianos. The travel time function of the
8 B&K Sound Level Calibrator Type 4230; test tone 94 dB, 1 kHz.
161
 Proceedings of the Stockholm Music Acoustics Conference, August 6–9, 2003 (SMAC 03), Stockholm, Sweden

7 Steinway C single display. It becomes evident that type of touch affected the
Yamaha Disklavier Legato
6 Bösendorfer SE290 shape of the curves more than the instrument.
5 Key bottom
Maximum hammer velocity (m/s)

4 4. CONCLUSIONS
3 Finger−key
2 Max. HV This contribution provided benchmark data for the temporal prop-
1 erties of different grand pianos under two touch conditions. The
temporal properties varied considerably between type of touch,
−200 −175 −150 −125 −100 −75 −50 −25 0 25
only marginally between pianos, and not at all between the dif-
ferent tested keys. The temporal differences between a pp and a ff
tone can become as large as 200 ms, between types of touch about
7
Staccato 20–30 ms. A pianist must not only adapt to the intended dynamic
6
level, but also to the kind of touch in order to control the precise
5 timing of the produced tone.
4 Generally, a keystroke starts for the pianist kinesthetically with
3 finger–key contact (the acceleration impulse by the finger) and
2 ends at key bottom, but for the audience it starts with the hammer–
1 string contact. Although these three points in time can be apart
from each other as far as 200 ms, they might not be perceived as
−200 −175 −150 −125 −100 −75 −50 −25 0 25
separate events. The intrinsic timing properties of a particular pi-
Time relative to hammer−string (ms)
ano might be an important factor pianists have to get acquainted
with when they ‘warm up’ on a yet unknown instrument.
Figure 5: Temporal properties of three grand piano actions. Power
curve approximations for the three pianos (line style), the two
types of touch (panels), and for finger–key (left), maximum ham- 5. ACKNOWLEDGEMENTS
mer velocity (max. HV, middle), and key bottom times (right).
This study was supported by the European Union (Marie Curie Fellow-
ship, HPMT-GH-00-00119-02, the Sounding Object project (SOb), IST-
2000-25287, http://www.soundobject.org, and the MOSART IHP network,
Bösendorfer replicated the curve used in [1]. Legato keystrokes HPRN-CT-2000-00115) and the START program from the Austrian Fed-
needed longer time to produce a sound than staccato keystrokes eral Ministry for Education, Science, and Culture (Grant No. Y99–INF)
that produced equally fast attacks. Moreover, the softest notes and a visiting grant from the Swedish Institute for the third author. The
were performed with legato touch, the loudest with staccato. Ac- Austrian Research Institute for Artificial Intelligence acknowledges basic
celerating the keys in legato manner did not allow to generate ham- financial support from the Austrian Federal Ministry for Education, Sci-
mer velocities far beyond 4 m/s. ence, and Culture.
Key bottom times are plotted relative to hammer–string con-
tact in Figure 2 (negative values denote key bottom contact pre-
ceding the hammer–string contact). The data for legato, staccato, 6. REFERENCES
and reproduction were approximated by power functions. The data
clusters were clearly distinct for the two types of touch on the [1] Werner Goebl, “Melody lead in piano performance: Expres-
Yamaha and the Bösendorfer, but almost overlapping on the Stein- sive device or artifact?,” Journal of the Acoustical Society of
way. America, vol. 110, no. 1, pp. 563–572, 2001.
The instant of the escapement was calculated as the maximum [2] Anders Askenfelt and Erik V. Jansson, “From touch to string
of the hammer velocity. This was a good estimation for strong and vibrations,” in Five Lectures on the Acoustics of the Piano,
medium keystrokes but not for keystrokes when the escapement Anders Askenfelt, Ed., vol. 64, pp. 39–57. Publications issued
can occur before the hammer maximum velocity. The time inter- by the Royal Swedish Academy of Music, Stockholm, 1990.
val between the escapement and the hammer hammer–string con- [3] Anders Askenfelt and Erik V. Jansson, “From touch to string
tact was calculated (Figure 3). With escapement, the pianist looses vibrations. I. Timing in grand piano action,” Journal of the
control over the tone. The point of maximum hammer velocity Acoustical Society of America, vol. 88, no. 1, pp. 52–63, 1990.
coincided well with the escapement point for medium and fast
keystrokes, but was earlier for slow keystrokes. The time of free [4] Anders Askenfelt and Erik V. Jansson, “From touch to string
flight was almost zero beyond hammer velocities of 1.5 m/s for the vibrations. II. The motion of the key and hammer,” Journal
Steinway, and beyond 3 m/s for the Yamaha and the Bösendorfer. of the Acoustical Society of America, vol. 90, no. 5, pp. 2383–
On the Yamaha, the staccato data splits into two groups at medium 2393, 1991.
intensities: maximum hammer velocity occurred at two different [5] Anders Askenfelt and Erik V. Jansson, “From touch to string
instants (Figure 3, middle panel). vibrations. III. String motion and spectra,” Journal of the
In Figure 4, peak SPL values are plotted against maximum Acoustical Society of America, vol. 93, no. 4, pp. 2181–2196,
hammer velocity. The data formed different groups for different 1992.
tones, but overlapped almost entirely between legato and staccato [6] Robert A. Moog and Thomas L. Rhea, “Evolution of the
keystrokes. This indicates that the sound level of the tones does keyboard interface: The Bösendorfer 290 SE recording piano
not depend on the way the tone was produced, but exclusively on and the Moog multiply-touch-sensitive keyboards,” Computer
the hammer velocity. Music Journal, vol. 14, no. 2, pp. 52–60, 1990.
In Figure 5, all approximations reported above are plotted in a

162

View publication stats