Associative rhythm learning for music Interaction

Interactions between men and machines increasingly tend towards interactive, or even collaborative, experiences. We explore the emergence of creativity in man-machine musical interaction. Participants in collaborative musical ensembles should exhibit their own unique musical individuality. We propose a system that acquires musical individuality by listening to rhythmic interaction between human performers and learns bidirectional rhythm association. This association mechanism allows the system to not only generate a suitable rhythm that compliments a performer's rhythm, but also suggest to the performer an alternative rhythm that it deems preferential. The system's musical individuality originates from its listening experience and is generally different to that of the performer's. In this manner, creative and harmonious rhythms can emerge from musical interactions between the performer and the system. The system was implemented so as to interact with a human partner using typical percussion instruments, in real time. Experimentation shows that the system can identify rhythm, play along with human performers, and inject dynamic rhythm changes such that musical creativity emerges.