Efficient variational inference for the dynamic harmonic model

Author

Taylan, A. ; Godsill, Simon J.

Author_Institution

Dept. of Eng., Cambridge Univ.

fYear

2005

fDate

16-16 Oct. 2005

Firstpage

271

Lastpage

274

Abstract

In this paper, we develop a class of probability models that are potentially useful for various music applications such as polyphonic transcription, source separation, restoration or denoising. This class unifies and extends several models such as sinusoidal and harmonic models, additive synthesis model, Gabor regression and probabilistic phase vocoder. We overcome computational intractability issues by introducing structured variational (mean-field) approximations that lead to efficient local message passing algorithms

Keywords

Bayes methods; acoustic signal processing; regression analysis; vocoders; Gabor regression; additive synthesis model; computational intractability; denoising applications; dynamic harmonic model; message passing algorithms; music applications; polyphonic transcription; probabilistic phase vocoder; restoration applications; source separation; structured variational approximations; variational inference; Covariance matrix; Frequency estimation; Message passing; Multiple signal classification; Signal analysis; Signal processing; Signal processing algorithms; Source separation; Speech analysis; Vocoders;

fLanguage

English

Publisher

ieee

Conference_Titel

Applications of Signal Processing to Audio and Acoustics, 2005. IEEE Workshop on

Conference_Location

New Paltz, NY

Print_ISBN

0-7803-9154-3

Type

conf

DOI

10.1109/ASPAA.2005.1540222

Filename

1540222