Congrats on the launch! I just tried this and while the tuning examples actually sounded more like me (a bit robotic but not bad for 20 seconds), the actual generations sounded nothing like me. It was some kind of aussie accent - like you're just modifying some existing voice to get a little closer to mine. Tried it with 2 different versions of myself - reading from text vs conversational and both times got this weird aussie accent.
No comments yet.