Nightly TTS voice files switched from espeak-ng to piper, feedback needed

[speachy]

This has been in the works for some time now; I finally made the switch (from espeak-ng+mbrola) globally last night.

For most languages Piper provides multiple voice options.  I tried to pick the highest "quality" one on offer, but for anything other than English it was largely just a random choice since I personally can't evaluate how good/correct a language sounds when I don't speak it.

In other words, Non-English voice files need proper testing and feedback. Italian and Chinese are the only others I've received feedback on, and that helped settle on the choice of TTS engine+voice. 

So, please try the 20240719 (or later) nightly with your language of choice, and see how that sounds.   You can compare it to the 20240718 (or older) builds in comparison.  We can choose the engine (and voice) on a per-translation basis.

Here are the voice options that Piper currently provides: 

  https://github.com/rhasspy/piper/blob/master/VOICES.md

And here are the TTS options/voices we're using on a per-voice basis:

   https://git.rockbox.org/cgit/rockbox.git/tree/tools/builds.pm#n575

If there are things not translated or poorly translated, please pop over to https://translate.rockbox.org and make corrections!