Phonon 0.7.1
Changes since 0.7.0.
- OpenAI server:
--gain-db(PTTS_GAIN_DB) makes the speech louder or quieter by a fixed number of dB. When it raises the level, a limiter keeps peaks below -1 dBFS, so nothing clips. It applies to the HTTP API and to Wyoming, so Home Assistant announcements can match the volume of other engines. The default is 0 (#342). - Voices: a checkpoint whose baked-in voices are only a prefix, without conditions of their own, now accepts other voices and voice cloning too (#341).
phonon-31m-en-coreml.zip and phonon-31m-en-qnn.zip are the ones attached to 0.7.0: the default model has not changed.