Ray-Traced Shadow Denoising With Variance-Adaptive Temporal Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional ray tracing methods for rendering images result in noisy render data due to sparse sampling of shadow rays, leading to inefficient use of computing resources and increased rendering times, while existing denoising techniques require costly bandwidth for storing and processing hit distance data.
Innovation Solution
Adaptive spatiotemporal shadow denoising filters are employed, where filter values are defined based on variance in temporally accumulated ray-traced samples, allowing for the exclusion of irrelevant samples and reducing the need for storing hit distance data, thereby conserving computing resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If shadow rays are sparsely sampled to conserve computing resources, then rendering time is reduced, but noise in render data increases
Solution Approach 1:
The patent introduces an intermediary denoising filter that processes the noisy but computationally efficient sparsely-sampled shadow rays. This filter acts as a mediator between the sparse sampling approach and the final render output, removing noise artifacts while preserving the computational efficiency of sparse sampling.
Solution Approach 2:
The patent creates a copy of the shadow ray sampling process by generating multiple pseudo-random samples that statistically represent the lighting conditions. Instead of casting many actual shadow rays, the system uses copied sample data that mimics the statistical properties of fully-sampled shadows, achieving visual fidelity with reduced computation.
2Manufacturing precision
If conventional shadow denoising filters use ray-hit distance to guide filtering, then denoising accuracy improves, but bandwidth requirements and computational overhead increase
Solution Approach 1:
The patent extracts and removes the requirement for storing and processing hit distance data from the denoising pipeline. By eliminating this data dependency, the system achieves denoising without the associated bandwidth and computational overhead, while maintaining effectiveness through alternative filtering guidance mechanisms.
Solution Approach 2:
The denoising filter is designed to be self-sufficient by not requiring external hit distance data. The filter adapts to the local characteristics of the shadow data itself, using the available shadow ray samples to guide the denoising process without needing additional geometric information from the scene.
Data Source
AI summary
In examples, a filter used to denoise shadows for a pixel(s) may be adapted based at least on variance in temporally accumulated ray-traced samples. A range of filter values for a spatiotemporal filter may be defined based on the variance and used to exclude temporal ray-traced samples that are outside of the range. Data used to compute a first moment of a distribution used to compute variance may be used to compute a second moment of the distribution. For binary signals, such as visibility, the first moment (e.g., accumulated mean) may be equivalent to a second moment (e.g., the mean squared). In further respects, spatial filtering of a pixel(s) may be skipped based on comparing the mean of variance of the pixel(s) to one or more thresholds and based on the accumulated number of values for the pixel.