The MVDR Beamformer for Speech Enhancement

Abstract

The minimum variance distortionless response (MVDR) beamformer is widely studied in the area of speech enhancement and can be used for both speech dereverberation and noise reduction. This chapter summarizes some new insights into the MVDR beamformer. Specifically, the local and global behaviors of the MVDR beamformer are analyzed, different forms of the MVDR beamformer and relations between the MVDR and other optimal beamformers are discussed. In addition, the tradeoff between dereverberation and noise reduction is analyzed. This analysis is done for a mixture of coherent and non-coherent noise fields and entirely non-coherent noise fields. It is shown that maximum noise reduction is achieved when the MVDR beamformer is used for noise reduction only. The amount of noise reduction that is sacrificed when complete dereverberation is required depends on the directto- reverberation ratio of the acoustic impulse response between the source and the reference microphone. The performance evaluation demonstrates the tradeoff between dereverberation and noise reduction.