TY  - BOOK
AU  - Benesty,Jacob
AU  - Huang,Yiteng
TI  - A perspective on single-channel frequency-domain speech enhancement
T2  - Synthesis lectures on speech and audio processing,
SN  - 9781608456994 (electronic bk.)
AV  - TK7882.S65 B456 2011
U1  - 006.454 22
PY  - 2011///
CY  - San Rafael, Calif. (1537 Fourth Street, San Rafael, CA  94901 USA)
PB  - Morgan & Claypool
KW  - Speech processing systems
KW  - Noise control
KW  - Digital filters (Mathematics)
KW  - Single-channel noise reduction
KW  - Speech enhancement
KW  - Frequency domain
KW  - Linear and widely linear models
KW  - Interframe correlation
KW  - Wiener filter
KW  - Maximum signal-to-noise ratio (SNR) filter
KW  - Minimum variance distortionless response (MVDR) filter
KW  - Tradeoff filter
KW  - Linearly constrained minimum variance (LCMV) filter
N1  - Part of: Synthesis digital library of engineering and computer science; Series from website; Includes bibliographical references (p. 91-95) and index; 1. Introduction -- A brief history of single-channel noise reduction (SCNR ) -- Rudimentary problem of SCNR and approaches of this book -- Organization of the book --; 2. Problem formulation -- Signal model -- General framework for speech enhancement --; 3. Performance measures -- Noise reduction -- Speech distortion -- Mean-square error (MSE ) criterion --; 4. Linear and widely linear models -- Model 1: Conventional -- Model 2: Widely linear -- Model 3: Interframe correlation -- Model 4: Widely linear and interframe correlation --; 5. Optimal filters with model 1 -- Performance measures -- Wiener gain -- Tradeoff gain -- Maximum signal-to-noise ratio (SNR ) filter --; 6. Optimal filters with model 2 -- Performance measures -- Maximum SNR filter -- Wiener filter -- Minimum variance distortionless response (MVDR ) filter -- Tradeoff filter --; 7. Optimal filters with model 3 -- Performance measures -- Wiener filter -- MVDR filter -- Tradeoff filter -- Linearly constrained minimum variance (LCMV) filter --; 8. Optimal filters with model 4 -- Performance measures -- Wiener filter -- MVDR filter -- Tradeoff filter -- LCMV filter --; 9. Experimental study -- Setup and metrics -- Algorithm implementation -- Comparison of the Wiener filters with model 1 and model 3 -- MVDR filter with model 3 -- Perceptual quality --; Bibliography -- Authors' biographies -- Index; Abstract freely available; full-text restricted to subscribers or individual document purchasers; Compendex; INSPEC; Google scholar; Google book search; Also available in print
N2  - This book focuses on a class of single-channel noise reduction methods that are performed in the frequency domain via the short-time Fourier transform (STFT). The simplicity and relative effectiveness of this class of approaches make them the dominant choice in practical systems. Even though many popular algorithms have been proposed through more than four decades of continuous research, there are a number of critical areas where our understanding and capabilities still remain quite rudimentary, especially with respect to the relationship between noise reduction and speech distortion. All existing frequency-domain algorithms, no matter how they are developed, have one feature in common: the solution is eventually expressed as a gain function applied to the STFT of the noisy signal only in the current frame. As a result, the narrowband signal-to-noise ratio (SNR) cannot be improved, and any gains achieved in noise reduction on the fullband basis come with a price to pay, which is speech distortion. In this book, we present a new perspective on the problem by exploiting the difference between speech and typical noise in circularity and interframe self-correlation, which were ignored in the past. By gathering the STFT of the microphone signal of the current frame, its complex conjugate, and the STFTs in the previous frames, we construct several new, multiple-observation signal models similar to a microphone array system: there are multiple noisy speech observations, and their speech components are correlated but not completely coherent while their noise components are presumably uncorrelated. Therefore, the multichannel Wiener filter and the minimum variance distortionless response (MVDR) filter that were usually associated with microphone arrays will be developed for single-channel noise reduction in this book. This might instigate a paradigm shift geared toward speech distortionless noise reduction techniques
UR  - http://ieeexplore.ieee.org/servlet/opac?bknumber=6812652
ER  -