ABSTRACT
Speech Communication is very important in the everyday life. In every-day speech communication situations, the transmission of speech can be affected by the environment noises. Speech recognition algorithms, though increasingly successful in a naturally quiet environment, fail in the presence of background sound which does not trouble a human listener. Recent years much effort has been directed to reducing this deficit in auto speech processing [1]. In order to further understand the nature of speech processing, it is critical to find experimental paradigms to build an effective computational method for evaluating speech intelligibility in the noisy environment.
