Séparation de sources audio en milieu réverbérant : Factorisation en matrices non-négatives et représentation temporelle du mélange convolutif

Abstract : This paper addresses the problem of multichannel audio source separation in under-determined reverberant mixtures. We target a semi-blind scenario assuming that the mixing filters are known. The proposed method consists in working directly with the time-domain mixture signals. This approach makes it possible to accurately represent the convolutive mixing process, it is therefore suitable for the separation of highly reverberant mixtures. The source signals are represented in the modified discrete cosine transform domain with a Gaussian model based on non-negative matrix factorization (NMF). Source inference is based on a variational expectation-maximization algorithm. We experimentally show the advantage of using a time-domain representation of the convolutive mixture and a source model based on NMF.
Document type :
Conference papers
Complete list of metadatas

Cited literature [15 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01540481
Contributor : Roland Badeau <>
Submitted on : Thursday, June 29, 2017 - 3:58:30 PM
Last modification on : Monday, March 25, 2019 - 4:16:06 PM
Long-term archiving on : Thursday, January 18, 2018 - 2:28:06 AM

File

LeglaiveBadeauRichard_final.pd...
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01540481, version 1

Collections

Citation

Simon Leglaive, Roland Badeau, Gaël Richard. Séparation de sources audio en milieu réverbérant : Factorisation en matrices non-négatives et représentation temporelle du mélange convolutif. Colloque GRETSI, Sep 2017, Juan-Les-Pins, France. ⟨hal-01540481⟩

Share

Metrics

Record views

79

Files downloads

115