Abstract
This paper proposed a new single-channel speech enhancement in variable noise-level environment. This new method possesses the advantages of the spectral entropy analysis and temporal spectrum energy variation information. Hence the background noise level estimated by the proposed method is more precise than that estimated by the pure spectrum energy method [4]. Experimental results show that the proposed speech enhancement with considering the MS-entropy produces the enhanced speech signals with higher SNR values than that without considering the MS-entropy.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have