Fundamental limitation of frequency domain blind source separation for convolutive mixture of speech

By Shoko Araki, R. Mukai, Shoji Makino, T. Nishikawa, Hiroshi Saruwatari. 2001 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.01CH37221), 2001

398 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Geometric source separation: merging convolutive source separation with geometric beamforming (2001, background): “Note also that independent signals will have vanishing empirical cross-power only for sufficiently long estimation times [26], which may go much beyond the stationarity time of the signals.”
Blind Source Separation and Independent Component Analysis : A Review (2004, methodology): “In order to perform ICA for convolutive mixtures, most methods try to apply their algorithms to the mixtures in the time domain [139, 112, 6] and the frequency domain [130, 115, 12].”
A Consolidated Perspective on Multimicrophone Speech Enhancement and Source Separation (2017, neutral): “The ATFs can be either time-varying or time-invariant.”
A two-stage algorithm for one-microphone reverberant speech enhancement (2006, methodology): “Some methods consider unmixing systems as FIR filters, while others convert the problem into the frequency domain and solve an instantaneous BSS for every frequency channel.”