My ideal would be an open source "deepfake toolkit" that allows me to provide pre-recorded samples of her speech and then TTS in her voice. Unfortunately most articles and tools I'm finding are anti-deepfake. Any recommendations?
Fallback would be recording her speaking "phonetic pangrams" and then using her pre-recorded phonemes to recreate speech that sounds like her. I feel like the deepfake toolkit is the way to go. Appreciate any recommendations... There must be open source tools for this??