On the relation between gesture, tone production and perception in classical cello performance

Abstract

On bowed string instruments such as violin or cello, the quality of sound depends mostly on the performer's bowing technique, which determines the interaction between the bow hair and the string. An accomplished string player has numerous ways of shaping the spectrum of a desired sound. This research investigates the combination of bowing gestures necessary for production of a rich tone. In particular, bowing control parameters such as bow force, bow velocity and bow-bridge distance captured by a dedicated sensing system are analysed and compared against audio features. Using audio and gesture measurements of six advanced cellists recorded on two different instruments of a luthier class, we characterise a sound palette and respective bowing control patterns of each player in performed scales. We especially focus on how performers adjust their bowing technique to control the timbre of an instrument on which they have never practised before. Observed differences between the players on the measured audio features show consistency with the bowing parameters adapted for balancing the timbral changes due to instrument, string and fingering position.

Timbra: An online tool for feature extraction, comparative analysis and visualization of timbre

“…In terms of brightness (spectral roll-off), the largest difference was observed between perf2 and perf5 and the smallest between perf4 and perf5. Similar results for the three cellists were obtained in [15] for spectral centroid identified as the parameter most strongly correlated with the brightness dimension when a larger set of audio samples from various music excerpts was analyzed and compared. While different methods of feature calculations and advanced statistical analyses were employed in the former work, this example shows the usefulness of the tool in gaining preliminary insights into the studied data and helping to formulate initial research hypotheses.…”

“…The aim of the study was to reveal whether classical musicians can be discriminated by their tone quality using extracted acoustic parameters. The major experiments and research results were reported in [15].…”

“…The audio signal was captured by a high class ambient mono microphone (model Neumann U87Ai P48) located in the near-field to the cello. From the recorded repertoire (more details in [15]), a fragment of the 20th century Cello Sonata by Shostakovich: Movement I (bars 3-5: notes 5-15) was selected for this study.…”

“…It has been often correlated with acoustic features such as spectral centroid and spectral rolloff that are indicators of high frequency energy content in the audio spectrum. As an acoustical correlate of proposed timbre spaces, spectral centroid was proofed to be a good discriminator between timbres of different musical instruments [11], [12] as well as between timbres of the same instrument class [13]- [15].…”