Adaptive Beamforming Using Zelinski-TSNR Multichannel Postfilter for Speech Enhancement
S. Siva Priyanka, T. K. Satish Kumar · 2018
In this paper, a speech enhancement algorithm combining Generalized Sidelobe Canceller (GSC) beamformer with zelinski-Two Step Noise Reduction (TSNR) postfilter is proposed. The main goal of this method is to suppress the directional and diffuse noise. In generalized sidelobe canceller for noise cancellation, a subband feedback controller is used which is based on speech presence probability. GSC suffers from transient interferences and diffuse noise which is suppressed using GSC with zelinski - TSNR postfilter. In this paper desired signal is enhanced from directional noise using GSC with subband feedback controller and diffuse noise is suppressed using zelinski-TSNR Postfilter. The proposed algorithm show better results than existing methods in terms of PESQ, segmental SNR and LSD.