Home | deutsch  | Legals | Data Protection | Sitemap | KIT

Gradient Estimation for Real-Time Adaptive Temporal Filtering

Christoph Schied, Christoph Peters, Carsten Dachsbacher

Proceedings of the ACM on Computer Graphics and Interactive Techniques


Results of our novel spatio-temporal reconstruction filter (A-SVGF) for path tracing at one sample per pixel (cyan inset in frame 404) with a resolution of 1280x720. The animation includes a moving camera and a flickering, blue area light. Previous work (SVGF) [Schied et al. 2017] introduces temporal blur such that lighting is still present when the light source is off and glossy highlights leave a trail (magenta box in frame 412). Our temporal filter estimates and reconstructs sparse temporal gradients and uses them to adapt the temporal accumulation factor alpha per pixel. For example, the regions lit by the flickering blue light have a large alpha in frames 406 and 412 where the light has been turned on or off. Glossy highlights also receive a large alpha due to the camera movement. Overall, stale history information is rejected reliably.


With the push towards physically based rendering, stochastic sampling of shading, e.g. using path tracing, is becoming increasingly important in real-time rendering. To achieve high performance, only low sample counts are viable, which necessitates the use of sophisticated reconstruction filters. Recent research on such filters has shown dramatic improvements in both quality and performance. They exploit the coherence of consecutive frames by reusing temporal information to achieve stable, denoised results. However, existing temporal filters often create objectionable artifacts such as ghosting and lag. We propose a novel temporal filter which analyzes the signal over time to derive adaptive temporal accumulation factors per pixel. It repurposes a subset of the shading budget to sparsely sample and reconstruct the temporal gradient. This allows us to reliably detect sudden changes of the sampled signal and to drop stale history information. We create gradient samples through forward-projection of surface samples from the previous frame into the current frame and by reevaluating the shading samples using the same random sequence. We apply our filter to improve real-time path tracers. Compared to previous work, we show a significant reduction of lag and ghosting as well as improved temporal stability. Our temporal filter runs in 2 ms at 1080p on modern graphics hardware and can be integrated into deferred renderers.


Supplemental video

Additional video material