Nonnegative matrix factorization and spatial covariance model for under-determined reverberant audio source separation

Simon Arberet 1 Alexey Ozerov 2 Ngoc Duong 2 Emmanuel Vincent 2 Rémi Gribonval 2 Frédéric Bimbot 2 Pierre Vandergheynst 1
2 METISS - Speech and sound data modeling and processing
IRISA - Institut de Recherche en Informatique et Systèmes Aléatoires, Inria Rennes – Bretagne Atlantique
Abstract : We address the problem of blind audio source separation in the under-determined and convolutive case. The contribution of each source to the mixture channels in the time-frequency domain is modeled by a zero-mean Gaussian random vector with a full rank covariance matrix composed of two terms: a variance which represents the spectral properties of the source and which is modeled by a nonnegative matrix factorization (NMF) model and another full rank covariance matrix which encodes the spatial properties of the source contribution in the mixture. We address the estimation of these parameters by maximizing the likelihood of the mixture using an expectation-maximization (EM) algorithm. Theoretical propositions are corroborated by experimental studies on stereo reverberant music mixtures.
Type de document :
Communication dans un congrès
Information Sciences Signal Processing and their Applications (ISSPA), 2010 10th International Conference on, May 2010, Kuala Lumpur, Malaysia. IEEE, pp.1--4, 2010, 〈10.1109/ISSPA.2010.5605570〉
Liste complète des métadonnées

Littérature citée [15 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/inria-00541436
Contributeur : Rémi Gribonval <>
Soumis le : samedi 5 février 2011 - 12:53:44
Dernière modification le : mercredi 16 mai 2018 - 11:23:03
Document(s) archivé(s) le : vendredi 6 mai 2011 - 02:23:36

Fichier

2010_ISSPA_ArberetEtAl.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

Citation

Simon Arberet, Alexey Ozerov, Ngoc Duong, Emmanuel Vincent, Rémi Gribonval, et al.. Nonnegative matrix factorization and spatial covariance model for under-determined reverberant audio source separation. Information Sciences Signal Processing and their Applications (ISSPA), 2010 10th International Conference on, May 2010, Kuala Lumpur, Malaysia. IEEE, pp.1--4, 2010, 〈10.1109/ISSPA.2010.5605570〉. 〈inria-00541436〉

Partager

Métriques

Consultations de la notice

586

Téléchargements de fichiers

422