Affiliations: Fred Hutchinson Cancer Research Center, Seattle, WA 98109, USA. E-mail: [email protected] | Department of Computer Science, University of North Carolina, Charlotte, NC 28223, USA | Institute of Computer Science, Warsaw University of Technology, 00-665 Warsaw, Poland. E-mail: [email protected]
Note: [] Corresponding author.
Abstract: Recently, numerous successful approaches have been developed for instrument recognition in monophonic sounds. Unfortunately, none of them can be successfully applied to polyphonic sounds. Identification of music instruments in polyphonic sounds is still difficult and challenging. This has stimulated a number of research projects on music sound separation, new features development, and more recently on hierarchically structured classifiers used in content-based music recommender systems. This paper introduces a hierarchically structured cascade classification system to estimate multiple timbre information from the polyphonic sound by classification which is based on acoustic features and short-term power spectrum matching. This cascade system makes a first estimate on the higher level decision attribute which stands for the musical instrument family. Then, the further estimation is done within that specific family range. Experiments showed better performance of a hierarchical system than the traditional flat classification method which directly estimates the instrument without higher level of family information analysis. Traditional hierarchical structures were constructed in human semantics, which are meaningful from human perspective but not appropriate for a cascade system. We introduce a new hierarchical instrument schema according to the clustering results of the acoustic features. This new schema better describes the similarity among different instruments or among different playing techniques of the same instrument. The classification results show the higher accuracy of cascade system with the new schema compared to the traditional schemas.
Keywords: Music information retrieval, automatic indexing, knowledge discovery, classifiers, clustering
