Text to Speech Settings
These mockups are the result of an investigation of how something like spiel (a speech synthesis client library) could help towards giving users the ability to choose a better or preferred voice for the screen reader right from Settings.
Spoken Content Settings
One approach is to treat text-to-speech as a general use feature beyond the screen reader for reading content on the screen, and having a configuration per-language. This is not unlike a feature on Android[2] for general TTS output.
Accessibility Settings
Another is to focus solely on the screen reader as an accessibility feature and integrate with the Seeing subpanel which would open a dialog that allows users to pick a speech provider from a list of those installed, as well as tune the speed, pitch and volume of that voice.
Provider Installation
While I had the consideration of doing the installation of new voices right here in Settings (much like Microsoft [1]), I think the heavy lifting of installing new voices would be better taken care of by GNOME Software.
Examples
[1] Microsoft voice installation
[2] Android Speech output