I’ve set up a local MaryTTS server and installed the UK voices in it, but the home assistant configuration rejects them as invalid:
2017-09-30 09:28:47 ERROR (MainThread) [homeassistant.config] Invalid config for [tts.marytts]: value is not allowed for dictionary value @ data['voice']. Got 'dfki-prudence'. (See ?, line ?). Please check the docs at https://home-assistant.io/components/tts.marytts/
Looking at the code I can see this:
SUPPORT_VOICES = [
'cmu-slt-hsmm'
]
which presumably means no other voices are actually valid. I have copied the marytts.py file to custom_components/tts and added dfki-prudence and the config loads correctly.
I assume no-one is really using MaryTTS as they’re not as paranoid as me but it would be handy to have a note on the documentation that cmu-slt-hsmm is actually the only voice available by default.
Ive just started looking at MaryTTS and preffer using noncloud services… I was thinking of creating a hassio bundled addon. But cant find any voices. Do you know which once come preinstalled with MaryTTS?
I stopped with it as I started looking at snips but I might come back to it again. To install other voices you need to use the component installer GUI, which is easy enough if you have a desktop on the machine you are installing, it’s in /path/to/marytts/target/marytts-5.2/bin/marytts-component-installer
If not, you have to export the GUI to a machine with an X server running (I use Linux on my laptop so that was easy enough) with something like DISPLAY=ip.ad.dre.ss:0; export DISPLAY where ip.ad.dre.ss is the address of your desktop.