Technical Program

Paper Detail

Paper:	SPTM-L3.3
Session:	Applications to Speech and Audio
Time:	Tuesday, May 16, 17:10 - 17:30
Presentation:	Lecture
Topic:	Signal Processing Theory and Methods: Signal Restoration, Reconstruction, and Enhancement
Title:	Sparse regression with structured priors: application to audio denoising
Authors:	Cédric Févotte, University of Cambridge, United Kingdom; Laurent Daudet, Université Pierre et Marie Curie, France; Simon Godsill, University of Cambridge, United Kingdom; Bruno Torrésani, Université de Provence, France
Abstract:	We describe in this paper a fully Bayesian approach for sparse audio signal regression in an union of two bases, with application to audio denoising. One basis aims at modeling tonal parts and the other at modeling transients. The noisy signal is decomposed as a linear combination of atoms from the two basis, plus a residual part containing the noise. Conditionally upon an indicator variable which is either 0 or 1, one source coefficient is set to zero or given a hierarchical prior. Various priors can be considered for the indicator variables. In addition to non-structured Bernoulli priors we study the performance of structured priors which favor horizontal time-frequency structures for tonals and vertical structures for transients. A Gibbs sampler (a standard Markov chain Monte Carlo method) is used to sample from the parameters of the model. We present results over denoising of a piano sequence using a MDCT basis with long time resolution to model the tonals and a MDCT with short time resolution to model the transients.