bruce Considering the AI passthrough is an early iteration, i'm amazed at how good it is if the user is willing to accept a few issues, the scenes that you did for me are flippin incredible and i can't wait to see how this tech develops, and i'm still on Quest 2.
As the technology / algorithm improves will the scenes that have already been processed need to be processed again to take advantage of any updates to the algorithm ? or as the process is done live take advantage of it straight away ?
A few improvements / suggestions i'd like to see:-
1) For some reason there can be a a lot of black moving blotches when performers get close to camera which need improving.
2) If possible when performers have for example their hands on hips you can see the background of the original scene through the gap between body and arms.
3) A user option to feather / soften edges of the mask would be the best update if that is possible.
All these are just suggestions and i understand the tech is in early development and in no way a complaint as i appreciate the work you and the team have done with this.
