You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In the current interface, the expressive avatar has no transition between words. As a result, it tends to teleport around word-to-word, making it significantly harder to understand. We can probably add a few keyframes between each word where the avatar moves its hands and body from one location to another.
I'm not sure how this would visually work. Maybe calculating the starting point and endpoint between each frame and simply inferencing frames in between could work? I'm not sure if it would lead to inhuman
The text was updated successfully, but these errors were encountered:
Description
In the current interface, the expressive avatar has no transition between words. As a result, it tends to teleport around word-to-word, making it significantly harder to understand. We can probably add a few keyframes between each word where the avatar moves its hands and body from one location to another.
I'm not sure how this would visually work. Maybe calculating the starting point and endpoint between each frame and simply inferencing frames in between could work? I'm not sure if it would lead to inhuman
The text was updated successfully, but these errors were encountered: