Monika voice TTS test. Cutoffs are due to bad linesplits, probably fixable.
https://vocaroo.com/17TX4dKWWp7R
Resplitting/retiming lines if I implement this in the game is going to be an absolute pain, I don't think there's any models for STT that have text timing info, and there's certainly none that retime to an existing transcript, though I can certainly try and guess with an LLM and maybe something like https://cloud.google.com/speech-to-text/docs/async-time-offsets (30 seconds to process + TTS time...)