Uploaded November 2019 | Updated September 2026, 2 weeks ago
With the recent introduction of DirectX ray tracing and NVIDIA RTX, interoperability between the rasterization pipeline and the ray tracing pipeline can be applied in real-time applications.
The new technology does a lot of heavy lifting, but developers can’t just throw rays everywhere. They still have a limited ray budget to work with. They need to know how to place the rays to get meaningful, additional contributions from ray traced samples.
In Chapter 22 of Ray Tracing Gems, NVIDIA engineers Josef Spjut and Adam Marrs describe how to take advantage of the benefits that temporal anti-aliasing gives us, while addressing its failures. They also explain how to work within the constraints of a real game engine.
At SIGGRAPH 2019, Spjut presented a talk sharing his team’s research on improving temporal AA with adaptive ray tracing. This is a five minute excerpt, which introduces Spjut’s approach.
The full video can be viewed here:
developer.nvidia.com/siggraph/2019/video/sig937-vid
With the recent introduction of DirectX ray tracing and NVIDIA RTX, interoperability between the rasterization pipeline and the ray tracing pipeline can be applied in real-time applications.
The new technology does a lot of heavy lifting, but developers can’t just throw rays everywhere. They still have a limited ray budget to work with. They need to know how to place the rays to get meaningful, additional contributions from ray traced samples.
In Chapter 22 of Ray Tracing Gems, NVIDIA engineers Josef Spjut and Adam Marrs describe how to take advantage of the benefits that temporal anti-aliasing gives us, while addressing its failures. They also explain how to work within the constraints of a real game engine.
At SIGGRAPH 2019, Spjut presented a talk sharing his team’s research on improving temporal AA with adaptive ray tracing. This is a five minute excerpt, which introduces Spjut’s approach.
The full video can be viewed here:
developer.nvidia.com/siggraph/2019/video/sig937-vid
