French Text to Speech Voice (siwis)

Voice and vocoder models for larynx based on the SIWIS.

Used in Rhasspy in the rhasspy-tts-larynx-hermes service.

Samples

Usage

$ larynx \
    --model /path/to/tts-checkpoint.pth.tar \
    --vocoder-model /path/to/vocoder-checkpoint.pth.tar \
    --output-file /path/to/output.wav \
    'Merci beaucoup!'

Docker

Run a web server at http://localhost:5002

$ docker run -it -p 5002:5002 \
    --device /dev/snd:/dev/snd \
    rhasspy/larynx:fr-siwis-1

Endpoints:

• /api/tts - returns WAV audio for text
  • GET with ?text=...
  • POST with text body
• /api/phonemize - returns phonemes for text
  • GET with ?text=...
  • POST with text body
• /process - compatibility endpoint to emulate MaryTTS
  • GET with ?INPUT_TEXT=...

Model Details

• Type: Glow-TTS
• Sample rate: 22050 Hz
• Frequency range: 0-8000 Hz

See configuration for details.

Vocoder Details

• Type: Multi-band MelGAN
• Sample rate: 22050 Hz
• Frequency range: 0-8000 Hz

See configuration for details.

Files

Some files are split into multiple parts so that they can be uploaded to GitHub. This is done with the split command:

split -d -b 25M FILE FILE.part-

They can be recombined simply with:

cat FILE.part-* > FILE