An Improvement of HSMM-Based Speech Synthesis by Duration-Dependent State Transition Probabilities.

Jing Tao, Wenju Liu

Browse the full ISNN paper archive.