Block-Based TVAR Models for Single-Channel Blind Dereverberation of Speech from a Moving Speaker | IEEE Conference Publication | IEEE Xplore