Title :
Generalized Optimal Multi-Microphone Speech Enhancement Using Sequential Minimum Variance Distortionless Response(MVDR) Beamforming and Postfiltering
Author :
Kim, Lae-Hoon ; Hasegawa-Johnson, Mark ; Sung, Koeng-Mo
Author_Institution :
Dept. of Electr. & Comput. Eng., Univ. of Illinois at Urbana-Champaign, Urbana, IL
Abstract :
A theoretical basis for optimal multichannel speech enhancements presented, sufficient, flexible to be used with any assumed statistical model and optimality criterion. Any Bayesian optimal one-channel estimator for speech enhancement can be generalized to the multichannel case as a sequentially constructed minimum variance distortionless response (MVDR) beamformer followed by an optimal one-channel postfilter. We present experimental results using the minimum mean-square error log-spectral amplitude (MMSE-logSA) optimality criterion, applied to a statistical model with simplified channel but realistic inter-microphone noise coherence. Word error rate in the audio-visual speech in a car (AVICAR) corpus (moving car, windows open) is reduced from 18% to 9%
Keywords :
Bayes methods; array signal processing; error statistics; filtering theory; least mean squares methods; speech enhancement; statistical analysis; AVICAR corpus; Bayesian optimal one-channel estimator; audio-visual speech; generalized optimal multi-microphone speech enhancement; log-spectral amplitude optimality criterion; minimum mean-square error optimality criterion; optimal multichannel speech enhancements; optimal one-channel postfilter; realistic inter-microphone noise coherence; sequential minimum variance distortionless response beamforming; statistical model; word error rate; Amplitude estimation; Array signal processing; Bayesian methods; Coherence; Error analysis; Microphones; Noise level; Noise robustness; Speech enhancement; Statistics;
Conference_Titel :
Acoustics, Speech and Signal Processing, 2006. ICASSP 2006 Proceedings. 2006 IEEE International Conference on
Conference_Location :
Toulouse
Print_ISBN :
1-4244-0469-X
DOI :
10.1109/ICASSP.2006.1660591
