Schobben D.W.E. Real-time Adaptive Concepts in Acoustics. Blind Signal Separation and Multichannel Echo Cancellation

Springer, 2001. — 160 p.This book deals with two topics; blind signal separation and multichannel acoustic echo cancellation. Both have received a lot of interest in literature lately. For multichannel echo cancellation this interest is mainly technology driven by the mass introduction of multichannel audio communication systems and speech recognition systems. Pioneering work on blind signal separation audio applications has been published less than a decade ago. The topic been booming ever since for the obvious reasons explained below.
The problem that is solved by blind signal separation is a magical one. This is especially true for acoustical applications where microphones for example pickup a mixture of various speech signals which are contaminated by room acoustics. The goal is to retrieve the independent speech signals without having knowledge of either (properties ot) the speech signals or the acoustical transfer functions. The numerous applications that can be based on this technology include voice control and teleconferencing. In 1995 the paper by Bell and Sejnowski, 1995a was published which gave both a theoretical background and reported promising results when working real-world audio signals. This is one of the early publications that motivated many researchers to join the quest for the holy grail that is called blind signal separation.
Echo cancellation is a well known technique that was developed about three decades ago. Its application can be found in any product that has both a and a loudspeaker, the purpose of the echo canceler being to cancel the contribution of the loudspeaker signal to the microphone signal. It has been widely used in many consumer products over the last decade now that this technology has become economic to implement. Just when most aspects of echo cancellation have been studied intensively a new problem arises with the introduction of multichannel sound systems; that of multichannel echo cancellation.Part I Background and IntroductionArray Processing Techniques
Efficient Filtering Using FFTs
Part II Acoustic Echo Cancellation
An Efficient Adaptive Filter Implementation
Efficient Multichannel RLS
Part III Blind Signal Separation
Blind Signal Separation, an Overview
A Blind Signal Separation Algorithm
Joint Blind Signal Separation and Echo Cancellation
Blind Signal Separation Algorithm Evaluation
ConclusionsA: Efficient Beamforming
B: Indeterminacies of Convolutive BSS based on Decorrelation
C: Computational Complexity of CoBliSS
D: Efficient computation of windowed FFTs
E: Contents of the Compact Disc
F: Glossary