Subband-Based Blind Separation for Convolutive Mixtures of Speech

Authors:
Shoko Araki;Shoji Makino;Robert Aichner;Tsuyoki Nishikawa;Hiroshi Saruwatari
Affiliations:
The authors are with the NTT Communication Science Laboratories, NTT Corporation, Kyoto-fu, 619-0237 Japan. E-mail: shoko@cslab.kecl.ntt.co.jp,;The authors are with the NTT Communication Science Laboratories, NTT Corporation, Kyoto-fu, 619-0237 Japan. E-mail: shoko@cslab.kecl.ntt.co.jp,;The author is with the University Erlangen-Nuremberg, Cauerstrasse 7, 91058 Erlangen, Germany.,;The authors are with the Graduate School of Information Science, Nara Institute of Science and Technology, Ikoma-shi, 630-0192 Japan.;The authors are with the Graduate School of Information Science, Nara Institute of Science and Technology, Ikoma-shi, 630-0192 Japan.
Venue:
IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences
Year:
2005

Citing 0
Cited 5

Wavelet packets approach to blind separation of statistically dependent sources

Neurocomputing
Online Subband Blind Source Separation for Convolutive Mixtures Using a Uniform Filter Bank with Critical Sampling

ICA '09 Proceedings of the 8th International Conference on Independent Component Analysis and Signal Separation
Time-domain blind audio source separation method producing separating filters of generalized feedforward structure

LVA/ICA'10 Proceedings of the 9th international conference on Latent variable analysis and signal separation
Subband blind audio source separation using a time-domain algorithm and tree-structured QMF filter bank

LVA/ICA'10 Proceedings of the 9th international conference on Latent variable analysis and signal separation
Efficient blind dereverberation and echo cancellation based on independent component analysis for actual acoustic signals

Neural Computation

Quantified Score

Hi-index	0.00

Visualization

Abstract

We propose utilizing subband-based blind source separation (BSS) for convolutive mixtures of speech. This is motivated by the drawback of frequency-domain BSS, i.e., when a long frame with a fixed long frame-shift is used to cover reverberation, the number of samples in each frequency decreases and the separation performance is degraded. In subband BSS, (1) by using a moderate number of subbands, a sufficient number of samples can be held in each subband, and (2) by using FIR filters in each subband, we can manage long reverberation. We confirm that subband BSS achieves better performance than frequency-domain BSS. Moreover, subband BSS allows us to select a separation method suited to each subband. Using this advantage, we propose efficient separation procedures that consider the frequency characteristics of room reverberation and speech signals (3) by using longer unmixing filters in low frequency bands and (4) by adopting an overlap-blockshift in BSS's batch adaptation in low frequency bands. Consequently, frequency-dependent subband processing is successfully realized with the proposed subband BSS.