Effects of audio feedback interventions with the Disklavier on the performance of piano students

Abstract

Objective: At an advanced level, musicians need to consider specific strategies of self-observation and self-evaluation to improve performances. Thus, the use of self-recordings is a suitable method. The purpose of this study was to investigate effects of short feedback interventions with audio playbacks on the performance of piano students. Methods: For that, 25 advanced piano students were asked to play a prepared piece. The students performed on a Disklavier, a grand piano with digital sensors and electromagnets to record and reproduce MIDI data. In the first intervention group, eight students were listening to a replay of their own performance on the Disklavier. A second intervention group of nine students was additionally able to manipulate this playback in tempo and volume by using a remote control. The third group was the control group with eight students, who did not listen to a playback but read a book instead. After the intervention, the students played the performance a second time and rated the performance compared to the first performance. Furthermore, all performances were acoustically recorded and rated by a group of professional pianists in an online survey. Results: The results showed significant differences in the self-ratings of the students between the groups in certain musical parameters such as rhythm, agogic, interpretation and musical expression. The intervention groups rated the second performance better in these parameters. The intervention groups did not differ from each other, but did differ from the control group. The external ratings did not show any significant effects. Conclusion: Overall, listening to one's own recording seems to have an influence on the process of improving one's own playing. In addition, the Disklavier showed great potential as a feedback instrument for professional piano students.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 5 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.