Return to Research

Variational EM for the Separation of Time-Varying Convolutive Mixtures

IEEE/ACM Transactions on Audio, Speech and Language Processing

D. Kounades-Bastian, L. Girin, X. Alameda-Pineda, S. Gannot, R. Horaud


This paper addresses the problem of separating audio sources from time-varying convolutive mixtures. We propose a probabilistic framework based on the local complex-Gaussian model combined with non-negative matrix factorization. The time-varying mixing filters are modeled by a continuous temporal stochastic process. We present a variational expectationmaximization (VEM) algorithm that employs a Kalman smoother to estimate the time-varying mixing matrix, and that jointly estimate the source parameters. The sound sources are then separated by Wiener filters constructed with the estimators provided by the VEM algorithm. Extensive experiments on simulated data show that the proposed method outperforms a block-wise version of a state-of-the-art baseline method.


Journal version (TASLP 2016) [IEEEXplore] [pdf].
A short version [IEEEXplore] [pdf] of the paper received the best student paper award (link) at IEEE WASPAA 2015.


Presentation from Waspaa 2015.


Code in MATLAB.


Derivation of the Variational Kalman Smoother.