This is the project page of our paper "MacST: Multi-Accent Speech Synthesis via Text Transliteration for Accent Conversion".

What's available?

• Paper: link
• Speech Demo: link
• The transliterated texts: link
• All datasets from MacST used in training Accent Conversion models: link

Todo

• Implementation (we update this upon acceptance of the paper)
  • Speech sample generation using MacST
  • A sequence-to-sequence VC Training

Acknowledgement

• Fantastic Demo Pages: zjlww/diffvoice-web and zjlww/ardit-web