Exploiting the Intermittency of Speech for Joint Separation and Diarization

Dionyssos Kounades-Bastian 1 Laurent Girin 2, 1 Xavier Alameda-Pineda 1 Radu Horaud 1 Sharon Gannot 3
1 PERCEPTION - Interpretation and Modelling of Images and Videos
Inria Grenoble - Rhône-Alpes, LJK - Laboratoire Jean Kuntzmann, INPG - Institut National Polytechnique de Grenoble
2 GIPSA-CRISSP - CRISSP
GIPSA-DPC - Département Parole et Cognition
Abstract : Natural conversations are spontaneous exchanges involving two or more people speaking in an intermittent manner. Therefore one expects such conversation to have intervals where some of the speakers are silent. Yet, most (multichannel) audio source separation (MASS) methods consider the sound sources to be continuously emitting on the total duration of the processed mixture. In this paper we propose a probabilistic model for MASS where the sources may have pauses. The activity of the sources is modeled as a hidden state, the diarization state, enabling us to activate/de-activate the sound sources at time frame resolution. We plug the diarization model within the spatial covariance matrix model proposed for MASS, and obtain an improvement in performance over the state of the art when separating mixtures with intermittent speakers.
Type de document :
Communication dans un congrès
IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, Oct 2017, New Paltz, NY, United States
Liste complète des métadonnées


https://hal.inria.fr/hal-01568813
Contributeur : Team Perception <>
Soumis le : mardi 25 juillet 2017 - 18:07:17
Dernière modification le : vendredi 28 juillet 2017 - 17:14:17

Fichier

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

Identifiants

  • HAL Id : hal-01568813, version 1

Citation

Dionyssos Kounades-Bastian, Laurent Girin, Xavier Alameda-Pineda, Radu Horaud, Sharon Gannot. Exploiting the Intermittency of Speech for Joint Separation and Diarization. IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, Oct 2017, New Paltz, NY, United States. <hal-01568813>

Partager

Métriques

Consultations de
la notice

148

Téléchargements du document

49