a cool experimental formant TTS based on old TTS papers
A huge Shoutout to Tamas Geczy for the help on getting high rate synthesis working
Code from TGSpeechBox was used as reference to tune these high rate improvements
i work on this in my free time, if you want fixes and feature requests, contributions are appriciated.
this repo is MIT, do whatever you want with the code, i don't care
you can read the Documentation Here
- English
- Japanese (hiragana only, experimental, looking for native speaker feedback)
PRs are welcome
if you find a pronunciation issue, you can report it either in the pronunciation tracker issue or in the discord
you can pick from one of the two links below:
discord
Pronunciation tracker
All code in this project is based on public Academic literature.
All tunings of this Synthesizer are by ear, by spectral analysis, or by formant logging debug words