Skip to Main content Skip to Navigation
New interface
Conference papers

An investigation of discrete-state discriminant approaches to single-sensor source separation

Valentin Emiya 1 Emmanuel Vincent 1 Rémi Gribonval 1 
1 METISS - Speech and sound data modeling and processing
IRISA - Institut de Recherche en Informatique et Systèmes Aléatoires, Inria Rennes – Bretagne Atlantique
Abstract : This paper investigated a new scheme for single-sensor audio source separation. This framework is introduced comparatively to the existing Gaussian mixture model generative approach and is focusing on the mixture states rather than on the source states, resulting in a discrete, joint state discriminant approach. The study establishes the theoretical performance bounds of the proposed scheme and an actual source separation system is designed. The performance is computed on a set of musical recordings and a discussion is proposed, including the question of the source correlation and the possible drawbacks of the method.
Complete list of metadata

Cited literature [10 references]  Display  Hide  Download
Contributor : Valentin Emiya Connect in order to contact the contributor
Submitted on : Tuesday, February 2, 2010 - 5:20:19 PM
Last modification on : Friday, February 4, 2022 - 3:15:35 AM
Long-term archiving on: : Friday, June 18, 2010 - 6:28:34 PM


Publisher files allowed on an open archive



Valentin Emiya, Emmanuel Vincent, Rémi Gribonval. An investigation of discrete-state discriminant approaches to single-sensor source separation. Proc. IEEE Work. Appli. Sig. Proces. Audio and Acous. (WASPAA), Oct 2009, New Paltz, NY, United States. pp.97-100, ⟨10.1109/ASPAA.2009.5346515⟩. ⟨inria-00452636⟩



Record views


Files downloads