Abstract: One of the popular configurations for the statistical model-based multichannel speech enhancement (SE) is to apply a spatial filter such as the minimum-variance distortionless response ...