Suggestions for Text-to-Speech Engines needed

As you may be aware LineageOS 16 and above does not include PicoTTS.
This mean there is no Text-to-Speech integrated support.
This is now showing up on the /e/ Pie builds we are testing for upgrading our devices.

There is a feature request raised for this on Gitlab

Since this would be a good feature to have in the OS for users with special needs besides those of us who regularly use this feature we are looking at suggestions on best options available that provide good voice quality and has a decent language choice.

Do you have any recommendations for good Text-to-Speech options which can be added to the /e/OS ?

1 Like

Unfortunately, I’ve not yet found an open source text-to-speech engine that is equally suitable for multiple languages.

Current exchange of experience here …

I only found F-Lite TTS and it works for me, cause i only use Osmand voice prompts
But it’s not great, don’t have many language TTSs, and is not mantained and probably won’t be. We needed a good alternative, but until now, there’s none, i guess.

There’s:

  • MARY, a Text-to-Speech System (MaryTTS) developed by a German public research centre (Open Source): http://mary.dfki.de/index.html
    It currently supports 9 languages (more are in preparation).
  • ESPEAK, Open Source as well. It supports some 20+ languages, but sound quality is considered to be quite robotic: http://espeak.sourceforge.net/
3 Likes

Actually, flite is a fork from Espeak, so they are quite same in many ways…
I think the only way to deal with this is the born of a new FLOSS TTS, and Mozilla and Mycroft are already working on this(Common Voice&voice.mycroft).
You can always contribute to them! No programming skills needed!