Juraj Borza

Maybe using something like this? https://github.com/lawrencenull/mobi I tried several local mobi files and it seemed to work. Of course it would require some work, as it's not as nice as...

@jeremiahsb did you manage to get Supertonic integrated? If yes, that's very cool

Speaking of other voice models, I'd like to try Supertonic. I'm not completely convinced adding another TTS engine than Kokoro wouldn't be painful, but I like the ebook processing, queuing...

@denizsafak and how do you see the PyQt version vs the web version?

@OpaTsupa000 used instructions here - https://huggingface.co/spaces/Supertone/supertonic - but it's really just the basics, so they have a file where you enter the text to synthesize and it outputs a .wav...

@denizsafak so I tried it, locally. I'm not super sure about Supertonic license - https://huggingface.co/Supertone/supertonic/blob/main/LICENSE - it's Open RAIL-M License. With regards to synthesizing, there are 10 voices, we can...

Yes @OpaTsupa000 , I see it at https://github.com/SUP3RMASS1VE/SuperTonic-TTS-Andriod/blob/main/app/src/main/java/com/sup3rmass1ve/supertonic/tts/VoiceGenerator.kt#L49 interesting, so it basically goes the same way as Kokoro, and we could mix voices from a formula like "0.654 * F1...

@denizsafak I'm not so sure - I suspect the language part over in the post ( https://huggingface.co/blog/YatharthS/llm-tts-models ) means the model sees audio the same way it sees text: as...

@jeremiahsb that seems useful. it would be good to still keep the code size manageable by a single developer. So something like TTSInterface with list of voices, available option, synthesize(text,...

@denizsafak that means that if I wanted to start using AWS Polly it would mean 100-150 lines saying how to start the synthesis and as a user to say where...