Application of Musical Expression Generation System to Learning Support of Musical Representation

This paper proposes a learning support system of musical representation by piano using teacher’s example of musical expression that is generated based on impression expressed by an adjective with our musical expression generation system. The system evaluates learner’s performance comparing with teacher’s example using a Kansei space and fuzzy rules expressing the relationship between musical expression and impression. The system presents good points of learner’s performance and advice by text to a learner for improvement of learner’s musical representation. A learner tries to improve his/her own performance based on system’s advice, and the system presents other advice again. From the experimental results, it is show that the proposed system is useful to learn musical representation and an approach of the proposed system is suitable because the affirmative evaluation is obtained from the participants who have taken piano lesson. On the other hand, it is found that to learn musical representation is difficult using the proposed system for learner of low-performance skills.