Re:Neural Actor: Neural Free-View Synthesis of Human Actors ...
by knicholes on Friday June 04, 2021 @11:23AM
(#27394499)
My mind went towards stuntmen. Now, studios can record natural actions for a large database of actions, and pick and choose. Instead of the Wilhelm Scream, you could see "Flip #564" and "Crash #827". If it doesn't even matter what shape the stuntman is, it seems to me like the industry might be looking to fill that occupation less and less.
[ Parent
| on HN ]
Re:Neural Actor: Neural Free-View Synthesis of Human Actors ...
by CyberDildonics on Friday June 04, 2021 @04:22PM
(#27398664)
There have already been CG stunts for multiple decades. You still have to match the lighting of the photography to integrate the CG even if you have something like this captured with flat texture maps.
The fidelity is also a big issue, flat textures and rough normals are not going to hold up to much scrutiny.
Re:Neural Actor: Neural Free-View Synthesis of Human Actors ...
by mrandish on Saturday June 05, 2021 @01:51AM
(#27402573)
> You still have to match the lighting of the photography to integrate the CG
True but dynamic relighting is rapidly improving, though it's not quite there yet for full frame hero shots.
Re:Neural Actor: Neural Free-View Synthesis of Human Actors ...
by CyberDildonics on Saturday June 05, 2021 @07:54AM
(#27404157)
What do you mean by "dynamic relighting"? I'm saying that even a flat texture captured mesh+video texture like this has to be lit to each shot it is placed in. That means it is being 're-lit' and any process that does that would be usable for regular CG.
Also again, the BRDF/normals of a capture like this are going to be very course. Cloth, skin, hair, eyes, metallic bits etc aren't going to look right if they have too much screen space over time. Quick motion blurred action might be applicable and I could see similar techniques being useful for crowds.