I have been trying to vocoder a few lines of technobabble for Grace Lee Whitney in a Star Trek the Motion Picture fan edit using Deepfacelab.
I am not a coder so I struggled to train the voice from scratch due to error messages I just didn't understand. In the end I managed to do something passable using pre-trained models in the demo but the accent just isn't quite there.
I only have about 10 minutes of speaking from various shows and interviews. Has anyone successfully trained a voice from scratch and what were the results like.
A brief vocoded line for Shatner elsewhere in the edit doesn't sound too bad. I think Grace just had a very unique voice!
I am not a coder so I struggled to train the voice from scratch due to error messages I just didn't understand. In the end I managed to do something passable using pre-trained models in the demo but the accent just isn't quite there.
I only have about 10 minutes of speaking from various shows and interviews. Has anyone successfully trained a voice from scratch and what were the results like.
A brief vocoded line for Shatner elsewhere in the edit doesn't sound too bad. I think Grace just had a very unique voice!