Sound Isolation by Harmonic Peak Partition For Music Instrument Recognition

Identification of music instruments in polyphonic sounds is difficult and challenging, especially where heterogeneous harmonic partials are overlapping with each other. This has stimulated the research on sound separation for content-based automatic music information retrieval. Numerous successful approaches on musical data feature extraction and selection have been proposed for instrument recognition in monophonic sounds. Unfortunately, none of those algorithms can be successfully applied to polyphonic sounds. Based on recent successful researches in sound classification of monophonic sounds and studies in speech recognition, Moving Picture Experts Group (MPEG) standardized a set of features of the digital audio content data for the purpose of interpretation of the information meaning. Most of them are in a form of large matrix or vector of large size, which are not suitable for traditional data mining algorithms; while other features in smaller size are not sufficient for instrument recognition in polyphonic sounds. Therefore, these acoustical features themselves alone cannot be successfully applied to classification of polyphonic sounds. However, these features contains critical information, which implies music instruments’ signatures. We proposed a novel music information retrieval system with MPEG-7-based descriptors and we built classifiers which can retrieve the important time-frequency timbre information and isolate sound sources in polyphonic musical objects, where two instruments are playing at the same time, by energy clustering between heterogeneous harmonic peaks.