Multichannel Source Separation Using Time-Deconvolutive CNMF

Abstract

This paper addresses the separation of audio sources from convolutive mixtures captured by a microphone array. We approach the problem using complex-valued non-negative matrix factorization (CNMF), and extend previous works by tailoring advanced (single-channel) NMF models, such as the deconvolutive NMF, to the multichannel factorization setup. Further, a sparsitypromoting scheme is proposed so that the underlying estimated parameters better fit the time-frequency properties inherent in some audio sources. The proposed parameter estimation framework is compatible with previous related works, and can be thought of as a step toward a more general method. We evaluate the resulting separation accuracy using a simulated acoustic scenario, and the tests confirm that the proposed algorithm provides superior separation quality when compared to a stateof-the-art benchmark. Finally, an analysis of the effects of the introduced regularization term shows that the solution is in fact steered toward a sparser representation.

Music and Mathematics in Latin America: Major Developments in the Last 25 Years

“…Covering as main topics audio restoration [56] and enhancement, spatial sound and acoustic localization [93], audio quality assessment [49], audio coding, and audio analysis and synthesis [58], since its origins the group is characterized by a strong link with Music. In addition to its ubiquity in applications from restoration to spatial sound and coding, Music has always motivated a considerable part of the Group's work, in the form of regular contributions in the field of Music Information Retrieval (MIR) involving alternative time-frequency representations for music [57], pitch tracking in polyphony [47], rhythm analysis [79], separation of musical sources [76], audioto-score alignment and expressiveness in music, among others.…”