ISCA Archive Interspeech 2015
ISCA Archive Interspeech 2015

Towards minimum perceptual error training for DNN-based speech synthesis

Cassia Valentini-Botinhao, Zhizheng Wu, Simon King

We propose to use a perceptually-oriented domain to improve the quality of text-to-speech generated by deep neural networks (DNNs). We train a DNN that predicts the parameters required for speech reconstruction but whose cost function is calculated in another domain. In this paper, to represent this perceptual domain we extract an approximated version of the Spectro-Temporal Excitation Pattern that was originally proposed as part of a model of hearing speech in noise. We train DNNs that predict band aperiodicity, fundamental frequency and Mel cepstral coefficients and compare generated speech when the spectral cost function is defined in the Mel cepstral, warped log spectrum or perceptual domains. Objective results indicate that the perceptual domain system achieves the highest quality.

doi: 10.21437/Interspeech.2015-268

Cite as: Valentini-Botinhao, C., Wu, Z., King, S. (2015) Towards minimum perceptual error training for DNN-based speech synthesis. Proc. Interspeech 2015, 869-873, doi: 10.21437/Interspeech.2015-268

  author={Cassia Valentini-Botinhao and Zhizheng Wu and Simon King},
  title={{Towards minimum perceptual error training for DNN-based speech synthesis}},
  booktitle={Proc. Interspeech 2015},