Loading…
3
Wasted $460 on an AI voice cloning tool for my podcast and it still sounds like a robot
So about 3 months back I got it in my head that I wanted to clone my own voice with AI and use it to dub my podcast episodes into Spanish and Portuguese. I did maybe 2 hours of googling, got hyped, and dropped $460 on a yearly plan for one of those voice clone platforms everyone was posting about on Twitter. You upload like 30 minutes of clean audio, it trains a model, and then you type text and it talks in your voice. Sounds sick in the demo videos. In real life? Every single episode, even after I re-recorded my training samples in a closet with a $200 mic, came out sounding like a GPS that just woke up from a nap. The pacing was all over the place, it kept putting weird pauses in the middle of words, and when I played it for my cohost in Fresno he laughed and asked if I was doing a bit. I tried tweaking the settings for like 3 weekends straight, and the best result still made me sound like I was reading a terms of service agreement underwater. Canceled the plan last month and ate the loss, and honestly the $460 stings way less than the 20 or so hours I'll never get back. If you're thinking about cloning your own voice for a show, please test the free tier with a full paragraph first, not just the sample they give you. Anyone here actually get a decent clone out of one of these, and if so what did your setup look like?
0 comments
Log in to join the discussion
Log In0 Comments
No comments yet
Be the first to share your thoughts on this discussion.