Closed aaronchantrill closed 1 month ago
There is a library called Dimits that makes using piper much easier. Just pip install dimits
then from dimits import Dimits
and dt = Dimits("en_US-amy-low")
and audio_file = td.text_2_audio_file(text, "speech_output", ".")
. The last command will generate an audio file and return the name that can just be passed to your audio playback method.
I have created a "Piper-TTS" plugin and added it to NPE.
Detailed Description
See https://github.com/rhasspy/piper
Context
It adds additional voices and supports a large number of languages (about 20).
Possible Implementation
Currently it has to be downloaded and compiled. Then a voice needs to be selected. It should be possible to select voices based on the language selected in Naomi, as the filename includes the locale code.