Minimum Generation Error Training for HMM-Based Speech Synthesis.

Yi-Jian Wu, Ren-Hua Wang

Browse the full ICASSP paper archive.