Abstract

We examine in some detail Mel Frequency Cepstral Coefficients (MFCCs) - the dominant features used for speech recognition - and investigate their applicability to modeling music. In particular, we examine two of the main assumptions of the process of forming MFCCs: the use of the Mel frequency scale to model the spectra; and the use of the Discrete Cosine Transform (DCT) to decorrelate the Mel-spectral vectors.

Citations

573	A maximization technique occurring in the statistical analysis of probabilistic functions of Markov chains - Baum, Petrie, et al. - 1970
230	Construction and evaluation of a robust multifeature speech/music discriminator - Scheirer, Slaney - 1997
166	Stochastic Finite Elements: A Spectral Approach - Ghanem, Spanos - 1991
125	Content-Based Retrieval of Music and Audio - Foote - 1997
86	DARPA TIMIT acoustic-phonetic continuous speech corpus CD-ROM - Garofolo, Lamel, et al. - 1993
75	A signal subspace approach for speech enhancement - Ephraim, Trees - 1995
66	Music summarization using key phrases - Logan, Chu - 2000
24	Musical content analysis through models of audition - Martin, Scheirer, et al. - 1998
21	Method and article of manufacture for content-based analysis, storage, retrieval, and segmentation of audio information - Blum, Keislar, et al. - 1999
21	Toward automatic sound source recognition: identifying musical instruments - Martin - 1998
3	Spectral Basis Functions from Discriminant Analysis - Hermansky, Malayath - 1998
2	On the asymptotic statistical behavior of empirical cepstral coefficients - Marhav - 1993