search

mkiol / dsnote

Speech Note Linux app. Note taking, reading and translating with offline Speech to Text, Text to Speech and Machine translation.
Mozilla Public License 2.0
566 stars 20 forks source link

[TTS] Add Mimic 3 models #25

Closed JamesClarke7283 closed 1 year ago

JamesClarke7283 commented 1 year ago

I would like to see the Mimic 3 models in this app

A link to the GitHub is HERE.

It does a better job than Piper in my opinion and sounds more real.

P.S Awesome Project, keep up the good work.

mkiol commented 1 year ago

I will look in to it.

BTW, I always thought that Piper is a successor of Mimic3 and eventually will replace it. Maybe I'm wrong. Both engines seems to be developed by the same people 🤔

JamesClarke7283 commented 1 year ago

I will look in to it.

BTW, I always thought that Piper is a successor of Mimic3 and eventually will replace it. Maybe I'm wrong. Both engines seems to be developed by the same people 🤔

They are made by different people.

Piper is made by the rhasspy people: https://github.com/rhasspy/piper

Mimic is made by the Mycroft people: https://mycroft.ai/mimic-3/

In my opinion the Mimic 3 voices sound better than Piper. but that's just my opinion.

You can find a demo on my instance here: https://voice.james-clarke.ynh.fr/

mkiol commented 1 year ago

Implemented in 203d13ae02f6b3c06dd6cb7a91b125ad21d2d2c8.

This change is included in "beta" version. You can install and test "beta" version from flathub-beta channel.