Ugh, the NotebookLM “hosts” sound so cheesy (also very recognizable). I understand that it takes time to record voice-overs, but maybe using a model trained on your voice would give it a more personal touch without taking up so much time? Keep at it, though! Love to see so much progress!
Yeah, they have the "temperature" a bit too high too as it glitches out occasionally. I've seen clone tools already, but none of the voices are great-great. I think OpenAI will likely have something in their API offerings soon with their advanced voice ... someone could make something even better I bet... But yeah, just tinkering... Thanks for the comment! An AI voice trained on my voice would save my voice too! haha
Ugh, the NotebookLM “hosts” sound so cheesy (also very recognizable). I understand that it takes time to record voice-overs, but maybe using a model trained on your voice would give it a more personal touch without taking up so much time? Keep at it, though! Love to see so much progress!
Yeah, they have the "temperature" a bit too high too as it glitches out occasionally. I've seen clone tools already, but none of the voices are great-great. I think OpenAI will likely have something in their API offerings soon with their advanced voice ... someone could make something even better I bet...
But yeah, just tinkering... Thanks for the comment!
An AI voice trained on my voice would save my voice too! haha