Skip to main content

Table 1 Objective evaluation results

From: Reducing over-smoothness in HMM-based speech synthesis using exemplar-based voice conversion

  Average standard deviation of spectral magnitude (dB) Mel-cepstral distortion (MCD) (dB)
HMMSS 4.26 6.38
HMMSS + GV 7.43 4.76
HMMSS + VC 7.96 4.31
HMMSS + GV + VC 7.55 4.62