ICASSP
2011
IEEE
12 years 8 months ago
2011
IEEE
Our previous analysis of speaker-adaptive HMM-based speech synthesis methods suggested that there are two possible reasons why average voices can obtain higher subjective scores t...