SPEECH ENHANCEMENT USING MICROPHONE ARRAY WITH MULTISTAGE PROCESSING

A microphone array system with multi-stage processing for speech enhan cement is presented in this paper. Two beamformers with uniform direct ional patterns, one aimed at the target source and the other at the in terfering sources, convert the multi-channel inputs into two data sequ ences. A novel microphone array structure with a small aperture has be en designed to obtain the dual beamformers. The outputs of the two bea mformers are then presented to a post-processing stage to further impr ove the quality and intelligibility of the speech signal. The post-pro cessing stage can be selected from one of three different algorithms t hat are presented, which are suitable for different acoustic environme nts. Applications for such a system include hands-free telephony, tele conferencing and also special situations where speech signals must be picked up in an extremely noisy acoustic environment in which the micr ophones are hidden (e.g. in a forensic covert recording system).