AMD Neural Supersampling and Denoising

[Trunks0]

Neural Supersampling and Denoising for Real-time Path Tracing

Read our research for a neural denoising and supersampling technique, with the aim of achieving real-time path tracing.

gpuopen.com

Neural Supersampling and Denoising for Real-time Path Tracing​

• Originally posted October 28, 2024

Creating realistic images has been a persistently challenging problem in computer graphics, especially when it comes to rendering scenes with complex lighting. Path tracing achieves photorealistic quality rendering by simulating the way light rays bounce around the scene and interact with different materials, but it also requires significant computation to generate clean images. This is where neural supersampling and denoising come into play. In this blog post, we describe how our neural supersampling and denoising work together to push the boundary for real-time path tracing.

Expected to work on RDNA3 and newer according to Mike Burrows posting on Twitter/X
(source:here)

 

[metroidfox]

Well, they kind of have to do stuff like this. Their current ray tracing hardware does not work for the current scheme of path tracing effects, and performs worse than the 4 rays per unit that the 4xxx Geforce series can do.

 

[Trunks0]

It also doesn't help that game developers them selves are mostly not developing particularly great denoisers for their engines. Their are exceptions of course.... but with neither Khronos Group or Microsoft setting standards for how this is handled in Vulkan or DirectX API's... we get the current state of affairs.