Playing around with xVAsynth 2.0 a while back I wanted to test the voice synthesizer with a longer piece to see if if it could create a “polished voice” compared to the short and rather robotic examples the app creator offered.

There’s unreasonable amounts of work required to make it sound as natural and dynamic as this, every single breath behind every single syllable have to be “animated” so it’s not sustainable to use it “at that level” for larger projects in my opinion. I’m pretty happy with the results though, it sounds around 80-90% “right” - And that’s pretty good.