Histogram-based subband powerwarping and spectral averaging for robust speech recognition under matched and multistyle training.

Mark Harvilla, Richard M. Stern

Browse the full ICASSP paper archive.