Enhancement of reverberant speech in noisy acoustical environments

Abstract

The propagated sound waves in an indoor environment hit the surfaces of solid objects and produce reverberant speech signals. Reverberated speech signals in noisy acoustical environments cause some problems such as reducing speech intelligibility, distinguishing speakers, locating source, quality for hands-free telephony, hearing aid, etc. Adaptive filters can be applied to suppress the interfering signals and reduce the reverberation effects or to dereverberate the received speech signals at microphone. In this paper we propose a two stage algorithm for enhancement of reverberant speech from one microphone recordings. In the first stage, Bayesian State-Space Kalman filter has been employed to reduce the effects of noise on received speech signal before applying it to dereverberation procedure. In the second stage, an inverse filter has been applied for reducing the reverberation's effect on received speech signal. In this paper different types of noises have been added to the reverberated speech signal in order to compare and analyze the results of noise reduction and dereverberation processes. In addition, different values of the Room Impulse Response are used in the dereverberation procedure and the results are compared. The results show a significant improvement in the quality of the received signal for the desired reverberant environment. Furthermore, the mean opinion score (MOS) of objective quality score from the perceptual evaluation of speech quality (PESQ) of two different values of RIR coefficient are compared.