Use the Mimic3 text-to-speech TTS engine for Reader mode and accessibility on Linux, instead of eSpeak
Categories
(Toolkit :: Reader Mode, enhancement)
Tracking
()
People
(Reporter: nekohayo, Unassigned)
Details
User Agent: Mozilla/5.0 (X11; Linux x86_64; rv:109.0) Gecko/20100101 Firefox/116.0
Steps to reproduce:
- Toggle "Reader" mode on an article with Ctrl+Shift+R
- Click the headphones button on the left to ask Firefox to read the text
Actual results:
Let your ears weep as the output is apparently spoken with the vintage eSpeak TTS engine, which sounds incredibly robotic and unpleasant to listen to.
Expected results:
Use this new AGPLv3-licensed TTS engine released by MyCroft at the end of 2022: https://github.com/MycroftAI/mimic3
Listen to this demonstration for example, where it seems all the voices are generated by Mimic3: https://www.youtube.com/watch?v=nZsZ34KpFjY
It is an order of magnitude better than eSpeak (or even Mimic 1 and Mimic 2), the quality is pretty much on par with the audible results that were traditionally only heard of in Big Tech's proprietary mobile OSes or cloud services. Very much less robotic, almost indiscernibly human.
Corresponding ticket for Orca in GNOME: https://gitlab.gnome.org/GNOME/orca/-/issues/378
|
||
Comment 1•2 years ago
|
||
The Bugbug bot thinks this bug should belong to the 'Toolkit::Reader Mode' component, and is moving the bug to that component. Please correct in case you think the bot is wrong.
|
||
Comment 2•2 years ago
•
|
||
FWIW: It looks like reader mode in linux is using speech-dispatcher. As a quick experiment, I modified my /etc/speech-dispatcher/speechd.conf setting the default synthesizer to Voxin (a proprietary synthesizer I have locally installed). The default was espeak-ng. I killed existing speech-dispatcher processes. Finally I enabled reader mode and clicked on the headphones. Firefox read the text using Voxin rather than espeak-ng. Note that you may have to also restart Firefox to see the change.
Description
•