3 ms·
Any plans to offer speech to speech models which keep prosody, intonation, and timing intact? ElevenLabs is getting expensive for this.
by oidar 2y ago
Any plans to offer speech to speech models which keep prosody, intonation, and timing intact? ElevenLabs is getting expensive for this.
- jeffharris 2y agowe'll keep expanding these GPT-4o based models with more controls. Is the main feature missing we're missing custom voices?
- oidar 2y agoNo, not custom voices - but voices that can be influenced by a recording. As in, a male voice actor records a part, and the model transforms it to a female part - keeping all the prosody, intonation and timing in the original recording. This would allow one voice actor to do many roles.