On the Use of Latent Mixing Filters in Audio Source Separation

Laurent Girin (1, 2) , Roland Badeau (3, 4)

1 PERCEPTION - Interpretation and Modelling of Images and Videos
2 GIPSA-CRISSP - GIPSA - Cognitive Robotics, Interactive Systems, & Speech Processing
3 LTCI - Laboratoire Traitement et Communication de l'Information
4 Télécom ParisTech

In this paper, we consider the underdetermined convolutive audio source separation (UCASS) problem. In the STFT domain, we consider both source signals and mixing filters as latent random variables, and we propose to estimate each source image, i.e. each individual source-filter product, by its posterior mean. Although, this is a quite straightforward application of the Bayesian estimation theory, to our knowledge, there exist no similar study in the UCASS context. In this paper, we discuss the interest of this estimator in this context and com- pare it with the conventional Wiener filter in a semi-oracle configuration.

Mots clés: MCMC sampling; Audio source separation; source image; MMSE estimator; latent mixing filters

Domaines: Traitement du signal et de l'image [eess.SP]

Laurent Girin, Roland Badeau. On the Use of Latent Mixing Filters in Audio Source Separation. LVA/ICA 2017 - 13th International Conference on Latent Variable Analysis and Signal Separation, Feb 2017, Grenoble, France. pp.225-235, ⟨10.1007/978-3-319-53547-0_22⟩. ⟨hal-01400965⟩