Techniques for handling background noise and reverberation when dealing with audio signals captured in the real world have gained tremendous attention during the last four decades. Recently, new and promising attempts at the problem have been reported independently in different areas, including adaptive filtering, speech enhancement, microphone array signal processing, and automatic speech recognition (ASR). While new methodologies for adaptive filtering are appearing to deal with time-varying colored source signals, model-based approaches are being investigated in the context of speech enhancement that are robust to changes of the room transfer functions. A practical solution designed to compensate jointly for noise and reverberation has been presented based on microphone array systems. New techniques for ASR have also been presented to handle reverberant speech signals appropriately in real acoustical environments based on statistical signal processing techniques. These emerging technologies are now prepared to be explored for useful applications in many audio signal processing areas, including hands-free ASR, online/offline high quality sound capturing, and natural immersive speech communication. Proper dissemination of these new techniques is important not only to accelerate their development, but also to offer a refreshed and yet cross-fertilized perspective to researchers engaged in related speech signal processing areas.
During the International Conference on Acoustic Speech and Signal Processing (ICASSP) 2009 there will be a special session titled: "Handling Reverberant Speech: Methodologies and Applications", organized by Dr. Tomohiro Nakatani (NTT Communication Science Laboratories, Japan) and Prof. Dr.-Ing. Walter Kellermann (University of Erlangen-Nuremberg , Germany). The aim of this special session is to offer an opportunity to link these techniques in different areas and to find effective ways of solving this pressing problem.