PSA: A guy on a podcast said their AI was trained on 4 hours of his own voice notes, and I tried it with mine
I clipped together about 90 minutes of me yapping into my phone on job sites and ran it through a free tool to make a voice model. It nailed how I say certain words but turned every sentence into the same flat rhythm. Curious if anyone has gotten better results by feeding it a script instead of random audio.