GART: Games and Applications using RayTracing

Status
Not open for further replies.



pics look darker for some reason, but in game it looks fine. Its a nice job done on this. The game is really heavy at full render resolution which you can select in the options. Its a little bellow 60 at 1440p/3080. FSR bumbs it up to 90-100

This guy is already doing Half Life, he has a trailer his channel
 
pics look darker for some reason, but in game it looks fine. Its a nice job done on this. The game is really heavy at full render resolution which you can select in the options. Its a little bellow 60 at 1440p/3080. FSR bumbs it up to 90-100
Just tried it on my 2080Ti, it runs 4K60 using DLSS performance (effectively 1080p native). It looks awesome!
 
Anyone examined the presentation? How's he getting 30 bounces? That in itself sounds excessive for lighting and only warranted for numerous reflective and transparent surfaces.

Because of the fur

This is a screenshot from a live capture of our research path tracer.



The scene shows an animated tiger walking through a dense jungle. The jungle has 3 billion instanced triangles.

For the tiger we’ve modeled the fur with explicit geometry. Every single hair strand is a curve. This is tessellated on-the-fly to a total of 3.5 million hair segments.

The light transport is simulated with up to 30 bounces of path tracing and then denoised. This many bounces are required to get the correct shadowing and lighting deep within the fur.



Note the FPS counter in the top-left corner. This renders at slightly above 30 FPS.

The settings are 1920x1080 and rendering is done on a single NVIDIA RTX 3090 GPU.
 
It says “up to 30”. That’s probably a fixed maximum. No way the average number of bounces is close to that.

The bit about tessellating the fur on the fly is interesting. They couldn’t be doing that every frame right?
 
Nice technique to complement the slow software traverser. But the visual difference is quite noticeable in some scenes.

Seems clever. From their own provided tests, the difference is negligible. The simplified scene representations can look quite horrid when shown directly, but as I understand it, they are only used for very long rays from very diffuse reflections. Sharp mirror-like reflection rays and ones that traveled short distances will still retrieve data from the full resolution polygon representation of the scene, no?
 
Status
Not open for further replies.
Back
Top