https://github.com/candlewill/rawnet

RawNet: Fast End-to-End Neural Vocoder
https://github.com/candlewill/rawnet

Last synced: over 1 year ago
JSON representation

RawNet: Fast End-to-End Neural Vocoder

• Host: GitHub
• URL: https://github.com/candlewill/rawnet
• Owner: candlewill
• Created: 2019-04-04T13:18:25.000Z (about 7 years ago)
• Default Branch: master
• Last Pushed: 2019-05-29T11:50:06.000Z (about 7 years ago)
• Last Synced: 2025-01-18T01:25:32.701Z (over 1 year ago)
• Size: 3.68 MB
• Stars: 42
• Watchers: 28
• Forks: 1
• Open Issues: 2
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# RawNet: Fast End-to-End Neural Vocoder

Samples of Paper [RawNet: Fast End-to-End Neural Vocoder](few), which is submitted to Interspeech 2019. 

RawNet, is a truly end-to-end neural vocoder, which use a coder network to learn the higher representation of signal, and an autoregressive voder network to generate speech sample by sample. The coder and voder together act like an auto-encoder network, and could be jointly trained directly on raw waveform without any human-designed features.



More detail is explained in the paper.

Code will be released soon...

## Samples

In the `RawNet samples` directory, you can find the demo samples generated by RawNet.

## Author

Yunchao He

yunchaohe@gmail.com