I’ve installed Mimic3 by pretending it’s MaryTTS and using http://localhost:59125/process
to call it successfully. I mention that, because the API call on some doc seems different.
Anyway I’d now like to install the other voices from here: GitHub - MycroftAI/mimic3-voices: Voice models for Mimic 3 text to speech system but not sure where to put what maybe something in the repo readme would be useful too?
TIA Hugh
I think its all goes in the profile dir hugh from the docker run command.
You can prob work it out from where current is.
docker run -d -p 12101:12101 \
--name rhasspy \
--restart unless-stopped \
-v "$HOME/.config/rhasspy/profiles:/profiles" \
-v "/etc/localtime:/etc/localtime:ro" \
--device /dev/snd:/dev/snd \
rhasspy/rhasspy \
--user-profiles /profiles \
--profile en
"$HOME/.config/rhasspy/profiles
is bind mounted in the container so have a look there are prob will become clear as ends up as /profiles
in the container
Thanks, so do I, but good to have a bit of support. I’ll back up my profile stuff and try. I managed to stop the whole thing working with a profile mistake about 3 weeks ago, so I’ve become a little cautious!
Look in profile and there’s nothing non-rhasspy there.
I’ve downloaded the extra voices, run make
and put the voices directory into ~/.local/shared/mycroft/mimic3/
. Actually, mimic3 (as server, anyway) is in a separate container. Then restarted the mimic3 container and Rhasspy.
I now get all the voice options in the drop down but TtsException: file does not start with RIFF id
I’m still actually not quite sure how to prepare and where to put the extra voices?
To answer myself: Mimic 3 - Mycroft AI looks as though it has some help, at least for non-docker Mimic3.
Greetings to all
Is the a way to insert my own words and phonetics to mimic 3 dictionary file?
It’s really urgent, I would appreciate any answer ASAP
docker exec 5f832d0d23ab /home/mimic3/app/.venv/bin/mimic3-download en_US/vctk_low
INFO:mimic3_tts.download:Downloading en_US/vctk_low
config.json: 100%|██████████| 3.50k/3.50k [00:00<00:00, 4.99MB/s]
generator.onnx: 100%|██████████| 133/133 [00:00<00:00, 227kB/s]
Unfortunately this is no longer enough. For each voice that was installed via mimic3-download the file ‘generator.onnx’ needs to be replaced by the one, that can be found on ‘mukowaty/mimic3-voices at main’ e.g. for German with three voices the ‘generator.onnx’ from each of these voices must be downloaded. As far as I understood this is required because ‘mimic3-download’ downloads only links for these files.