Automatic Cymbal Classification using Non-Negative Matrix Factorization

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Several musical instrument classifiers have been proposed.While many approaches in sound-feature extractionand in sound classification have been successfully used, mostfocus on distinguishing different harmonic instruments suchas the violin and the flute, whose sounds have very differentcharacteristics. On the other hand, much less attention hasbeen given to percussion instruments, especially if we considerthe discrimination of instruments of the same type, likethe cymbals in a drum kit.Here, we propose a classifier that is able to distinguishthis latter type of instruments. The classifier is able to distinguishsounds with very similar properties, like sounds producedby instruments with similar geometry that differ in materialor size. In particular it is able to distinguish soundsfrom the cymbals in a drum kit. Instead of using a set of predefinedfeatures, the classifier learns spectral features fromthe data using non-negative matrix factorization. This work isimportant to fill the gap on percussion instrument classificationand transcription (since most music transcribers focus onharmonic instruments).
Original languageUnknown
Title of host publicationProceedings of the IEEE International Conference on Systems, Signals and Image Processing
Pages468-472
Publication statusPublished - 1 Jan 2012
EventIEEE International Conference on Systems, Signals and Image Processing -
Duration: 1 Jan 2012 → …

Conference

ConferenceIEEE International Conference on Systems, Signals and Image Processing
Period1/01/12 → …

Cite this