Wow this is impressive, any ideas why the first demo sounds a bit "hollow" like there is an echo? Thanks for sharing

Thanks! The model learns a lot from unsupervised (as well as supervised) audio, so technically low-quality and high-quality audio are both just as likely to the model as music, background sounds or really anything else including echos or bad microphones :) will be interesting to learn how to control for these things, either through prompting or other switches during training/inference

This is awesome! So excited to play around with it!

Thanks :) Let us know if you come across interesting findings!