Description
Feature Use Case
Use case is to give Leon a personality that is not just a voice, but an actual visible form that can be interacted with like a virtual person or anime character, a fantasy based completely imaginary character...etc
Feature Proposal
Using an animated gif image, incorporate the features of Wav-2-Lip below into the main Leon App:
https://github.com/anothermartz/Easy-Wav2Lip/releases/tag/v8.3_release
Text that is generated by Leon can be either spoken in audio only, or recorded as a wav file and then animated live via the "simulated" talking-face from the GIF image that is automatically animated by the wav-2-lip model.
Additionally, RAG can be used to give the personality specific lines to re-use as part of their core personality.
If real time TTS is needed for this to work, please consider Piper as it also supports multiple languages and is super easy to train on own machine, can run on super slow and weak PC's on CPU and still does near-real-time generation.
Thanks for the awesome work!