This is an exciting piece of work. During testing I noticed that the character’s mouth corners gradually pull outward while speaking, which degrades lip‑sync quality. I also came across the recent MoDA paper (https://github.com/lixinyyang/MoDA) and Ditto (https://github.com/antgroup/ditto-talkinghead), which reference your work, but where the lip‑sync accuracy of echomimic appears suboptimal。 Do you recommend any additional settings, preprocessing steps, or hyperparameter adjustments to improve lip‑sync fidelity? Any practical tips you can share would be very helpful.
This is an exciting piece of work. During testing I noticed that the character’s mouth corners gradually pull outward while speaking, which degrades lip‑sync quality. I also came across the recent MoDA paper (https://github.com/lixinyyang/MoDA) and Ditto (https://github.com/antgroup/ditto-talkinghead), which reference your work, but where the lip‑sync accuracy of echomimic appears suboptimal。 Do you recommend any additional settings, preprocessing steps, or hyperparameter adjustments to improve lip‑sync fidelity? Any practical tips you can share would be very helpful.