Content-based methods for the management of digital music

7 November 2002

proceedings article
Published by Institute of Electrical and Electronics Engineers (IEEE)

Vol. 4, 2437-2440
https://doi.org/10.1109/icassp.2000.859334

Abstract

The literature on content-based music retrieval has largely finessed acoustic issues by using MIDI format music. This paper however considers content-based classification and retrieval of a typical (MPEG layer III) digital music archive. Two statistical techniques are investigated and appraised. Gaussian mixture modelling performs well with an accuracy of 92% on a music classification task. A tree-based vector quantization scheme offers marginally worse performance in a faster, scalable framework. Good results are also reported for music retrieval-by-similarity using the same techniques. Mel-frequency cepstral coefficients parameterize the audio well, though are slow to compute from the compressed domain. A new parameterization (MP3CEP), based on a partial decompression of MPEG layer III audio, is therefore proposed to facilitate music processing at user-interactive speeds. Overall, the techniques described provide useful tools in the management of a typical digital music library.

Keywords

This publication has 4 references indexed in Scilit:

Content-based retrieval of music and audio
Published by SPIE-Intl Soc Optical Eng ,1997
Speaker identification and verification using Gaussian mixture speaker models
Speech Communication, 1995
A tutorial on MPEG/audio compression
IEEE MultiMedia, 1995
A vector space model for automatic indexing
Communications of the ACM, 1975