Robust HMM phoneme modeling for different speaking styles.

Tatsuo Matsuoka, Kiyohiro Shikano

Browse the full ICASSP paper archive.