The Relationship between Biomechanics and Sound Quality during Piano Playing

Pozzo, Renzo
PhD in Medicine, Professor, Department of Medical and Biological Sciences, University of Medicine of Udine, Udine, Italy

Gaertner, Henriette
Professor, University of Music of Trossingen, Trossingen, Germany


Background: During last decades, we have conducted some investigations about the relationship between essential biomechanical parameters and the quality of piano interpretations. The main approach was to investigate the force acting on the key by the finger and on the pedal by the foot associated to the sound generated. This is crucial because it makes the difference between interpretation and pure mechanical played notes. In some cases, we also assessed the kinematics of the fingers and the muscular activation of essential muscles.

Method: To record finger forces a complete set of 2 octaves of pressure sensors (24 sensors, 300 Hz, Pliance-Novel, Germany) was used. For pedal force, two load-pads (100Hz, Novel) assured the acquisition of the forces applied on the heel and on the toe. Sound tracks were recorded by means of MD-MZ-R50 digital Sony recorder at 44,1 kHz sampling frequency. Time structure of force curves, peak force and impulse were calculated for each sensor and for each repetition. Video records were taken via normal CCD-cameras (50-300 Hz). In order to quantify the quality of the sound (interpretation) specific questionnaires were developed with a 4 level scores (very good-by far not so gut) concerning specific criteria as legato, crescendo, voice differentiation, sound quality, total impression and were submitted to professional piano teachers.

1st study: Fifteen pianists (8 females and 6 male) with different level of expertise (from IP: winning international prices to TL: top level) played the first 2 measures of Schubert’s Wanderer Phantasie (4 times in ff and 4 times in pp) on a grand piano. They were asked to play with theirs best interpretation and constancy; the finger sequences (fingering) were identical for all subjects. Very high differences were noted in the total impulses. Overall average values for TL subjects were 73% (0,91 Ns vs 0,24 Ns in ff) and 69% (0,33 1 Ns vs 0,11 Ns in pp) higher than for IP. Significant difference was also found in the CV of impulse structure (11%-18% for IP vs 33%-35% for TL). Moreover, an efficiency index was calculated by computing the proportion of the impulse up to the peak force to the total impulse. This index reveals the capacity of the pianist to produce the desired loudness without exert unnecessary force and the IP subjects showed higher values.
2nd study: Five males piano student and a professional soloist played the first 4 bars of the Schumann’s Scenes from Childhood. Students were under the supervision of a teacher using the traditional verbal information method (VIN) and a visual feedback method (FBIN) based on force parameters obtained by the pressure sensors. They had to reach the sound quality expressed by the professional interpret. Time structure of overlap-time phase for musical intervals and the force quotients at 150 ms after the force onset for each key stroke seem to be relevant for the optimization of the learning strategy of the participants. These improvements are significant greater using the FBIN method in comparison with the traditional one (VIN). Again, the questionnaire score accounted for e better estimation of the performance with FBIN method.
3thd study: 13 pianists (4 professional PL and 9 students SL) with different level of expertise played the first 5 measures of Galuppi’s Sonata V Cmajor 1st movement (4 times in p and 4 times in f) on a grand piano. They were asked to play with theirs best interpretation and constancy; the finger sequences (fingering) were identical for all subjects. PL pianists showed clear differentiation of finger forces with respect to music dynamic, i.e. with higher values in f than in p. For the efficient force (first 150 ms) peak values ranged 11,3 N - 5,51 N for PL and 7,17 N - 4,25 N for SL. This is quantified via f/p ratio with values ranged from 2,05-1,59 for PL and 1,41-1,08 SL subjects. To define the pedal-finger coordination the time delay between the peak force of pedal and the efficient force of the finger was calculated. PL subjects show almost positive values (0,30 s), whereas SL reach positive (0,50 s) but also negative values (-0,20 s), i.e. they anticipate with the pedal. In the analyzed piece the quality of the sound depends also on the relationship between the dominant scores (right hand) and the accompanying scores (left hand). Therefore, the quotient between right and left notes was calculated. Significant differences were observed between the groups, with PL showing higher values in p dynamic (1,84-2,10) as well in f dynamic (1,89-2,98), while SL reached lower values 0,82-1,78 and 1,74-2,41 corresponding

Conclusions: the results of these investigation are relevant for the understanding the relationship between movement’s parameters and the quality of sound. Moreover, they demonstrate a new way to improve the didactics and methods of piano playing using biomechanical methods.
Using visual or haptic information can make the pianist more confident with the neuromuscular mechanism of force perception and modulation.

Keywords: Biomechanics, piano playing, sound quality
Made on