12.235 pip

Piper TTS is an Open Source neural-net TTS engine. This module exposes Piper TTS to Emacs.

12.236 Piper TTS

  1. Sandbox for https://github.com/rhasspy/piper
  2. pipx install piper-tts.
  3. Set location of voice data in pip-data-dir. Default is ~/.local/share/voices-piper
  4. ;; Create the data directory above if needed.
  5. Run piper there using the –update-voices flag once for each desired voice to get the voice models.
  6. M-x pip-speak to speak.
  7. At present piper is not suitable for use as a primary TTS engine for Emacspeak, given its high latency.
  8. However it might be interesting to use it for lower-priority speech where quality trumps latency.

12.236.1 Pip Commands

12.236.1.1 pip-device-select

Command: pip-device-select (device)
Select default from available choices.
Restarts piper pipeline if already running.

(fn DEVICE)

12.236.1.2 pip-model-select

Command: pip-model-select (voice)
Select default from available choices.
Restarts piper pipeline if already running.

(fn VOICE)

12.236.1.3 pip-speak

Command: pip-speak (text)
Speak text

(fn TEXT)

12.236.1.4 pip-start

Command: pip-start
Start the Piper process

12.236.1.5 pip-stop

Command: pip-stop
Stop Piper TTS