The problem of enhancing speech degraded by uncorrelated additive noise, when only the noisy speech is available, has been widely studied in the past and it is still an active field of research. Wiener filter, which is the most fundamental approach, has been delineated in different forms and adopted in diversified applications. An improved wiener filtering algorithm is proposed in this study, which utilizes band-partitioning spectral entropy to achieve accurate and robust speech endpoint detection and a dynamic noise power spectrum is estimated for updating a priori SNR. Experimental results reveal that the proposed algorithm can extract the embedded speech segments from utterances containing a variety of background noise successfully.
展开▼