info about Generation in DLSS5 #2
Replies: 2 comments
|
I think he mentioned in the description that Normal, Depth and Motion Vector (?) inputs are to be implemented on a later development stage. I tested the OFX today on final 3D renders and got some very interesting results. It definitely has a place in the toolset. Flicker can be significant but Resolve has some good deflicker tools that can somehow fix this. I hope that SAOG0721 will continue developing this! |
|
I’m getting some serious flickering and splotchiness in some 3D product beauty shots. I also tried the standalone version made by another user, “DLSS5 Visual Enhancer”, which doesn’t have this flickering issue, although the effect is less prominent: I’ll be eagerly looking forward to updates to this Fusion version! |
Uh oh!
There was an error while loading. Please reload this page.
Hi, thanks for this plugin, it looks really interesting!
I’m trying it out to improve some 3D renders, and I’m trying to understand a bit better how it works.
I imagine the DLSS 5 version for games takes advantage of various passes, like Diffuse, Normal, Depth, Roughness, etc., and then does its magic.
I was wondering if your version could allow us to provide these passes as inputs as well. That way, perhaps it could skip generating this information from the video itself and potentially achieve a more temporally stable result.
Is something like this possible, or is it something you’ve considered?
All reactions