A music search system for expressive music performance learning

Abstract

In this paper, we present a music search system that focuses on performance style to cultivate a pupil's expressive performance of music. The system allows pupils to learn the performance style to be mastered by obtaining both model and non-model content. By browsing non-model content that is similar to the quality of a pupil's performance, the pupil can quickly identify his/her areas that require improvement. In addition, the pupil can improve his/her performance skill by repeatedly imitating the models. We evaluate the capabilities of our music search system regarding the extraction of performance style from a classical music source and the precision of the music search results for performance style. © 2014 Springer International Publishing.