Melodic Similarity: Looking for a Good Abstraction Level

Computing melodic similarity is a very general problem with diverse musical applications ranging from music analysis to content-based retrieval. Choosing the appropriate level of representation is a crucial issue and depends on the type of application. Our research interest concerns the development of a CBR system for expressive music processing. In that context, a well chosen distance measure for melodies is a crucial issue. In this paper we propose a new melodic similarity measure based on the I/R model for melodic structure and compare it with other existing measures. The experimentation shows that the proposed measure provides a good compromise between discriminatory power and the level of abstraction of melody representation.