Voice Conversion by Cascading Automatic Speech Recognition and Text-to-Speech Synthesis with Prosody Transfer

Jing-Xuan Zhang, Li-Juan Liu, Yan-Nian Chen, Ya-Jun Hu, Yuan Jiang, Zhen-Hua Ling, Li-Rong Dai


With the development of automatic speech recognition (ASR) and text-to-speech synthesis (TTS) techniques, it’s intuitive to construct a voice conversion system by cascading an ASR and TTS system. In this paper, we present an ASR-TTS method for voice conversion, which uses iFLYTEK ASR engine to transcribe the source speech into text and a Transformer TTS model with WaveNet vocoder to synthesize the converted speech from the decoded text. For the TTS model, a prosody code is used to describe the prosody information other than text and speaker information contained in speech. A prosody encoder is adopted to extract the prosody code. During conversion, the source prosody is transferred to converted speech by conditioning the Transformer TTS model with its code. Experiments were conducted to demonstrate the effectiveness of our proposed method. Our system also obtained the best naturalness and similarity in the mono-lingual task of Voice Conversion Challenge 2020.


 DOI: 10.21437/VCC_BC.2020-16

Cite as: Zhang, J., Liu, L., Chen, Y., Hu, Y., Jiang, Y., Ling, Z., Dai, L. (2020) Voice Conversion by Cascading Automatic Speech Recognition and Text-to-Speech Synthesis with Prosody Transfer. Proc. Joint Workshop for the Blizzard Challenge and Voice Conversion Challenge 2020, 121-125, DOI: 10.21437/VCC_BC.2020-16.


@inproceedings{Zhang2020,
  author={Jing-Xuan Zhang and Li-Juan Liu and Yan-Nian Chen and Ya-Jun Hu and Yuan Jiang and Zhen-Hua Ling and Li-Rong Dai},
  title={{Voice Conversion by Cascading Automatic Speech Recognition and Text-to-Speech Synthesis with Prosody Transfer}},
  year=2020,
  booktitle={Proc. Joint Workshop for the Blizzard Challenge and Voice Conversion Challenge 2020},
  pages={121--125},
  doi={10.21437/VCC_BC.2020-16},
  url={http://dx.doi.org/10.21437/VCC_BC.2020-16}
}