I’ve been trying to find a way to make rhasspy sound like Jarvis from the iron man movies, but i haven’t had any luck. I’ve got a somewhat powerful server to run rhasspy on (an actual cpu, much more powerful than a pi). If anyone is aware of a model that sounds like what i’m looking for, or a way to tune a voice myself, please let me know.
I have not tried it myself yet, but efficientspeech (GitHub - roatienza/efficientspeech: PyTorch code implementation of EfficientSpeech - to be presented at ICASSP2023.) appears to offer a way to generate custom voices. From what I’ve read, you can generate a voice from a few labeled samples on a desktop, then deploy the model to an edge machine. I don’t know that Rhasspy has a plugin for efficientspeech, though. I’ve long wanted to generate a few voices for my assistant, so I may try playing with this the next time I have a term break. If you have time to play with it and feel like sharing, I’d love to hear what you found out.