Hardware-Accelerated Motion Blur Ray Tracing With Fewer Intersection Tests
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Combining ray tracing and motion blur in graphics processing is resource-intensive and power-consuming due to the need for extensive processor tests for ray intersections with moving primitives during motion blur intervals.
Innovation Solution
Implementing spacetime interpolation circuitry to generate interpolated bounding boxes for ray tracing, using both spatial and temporal nodes in the acceleration data structure, which reduces the number of intersection tests and improves performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If ray tracing is combined with motion blur using traditional bounding volume hierarchies, then rendering quality is improved, but processor resources and power consumption increase significantly
Solution Approach 1:
The patent extends the traditional spatial bounding volume hierarchy into a spacetime hierarchy by adding a temporal dimension. Each node stores bounding volumes at multiple discrete time points within the motion blur interval, allowing rays to be tested against primitive positions at specific moments rather than requiring complex temporal integration, thus reducing computational load while maintaining rendering quality
Solution Approach 2:
The system pre-computes and stores bounding volumes at multiple discrete time points during the motion blur interval in advance. This preliminary action allows the ray tracer to simply query pre-computed spatial bounds at specific times rather than performing real-time complex calculations, significantly reducing processor resources and power consumption during actual rendering
2Reliability
If bounding volumes cover all possible positions during motion blur interval, then completeness is improved, but the number of intersection tests increases
Solution Approach 1:
The patent segments the motion blur interval into multiple discrete time points, storing separate bounding volumes for each time point. This segmentation allows the system to test rays against specific primitive positions at specific times rather than using a single oversized bounding volume that encompasses all possible positions, reducing the number of false positive intersection tests while maintaining completeness
Solution Approach 2:
The system dynamically selects which bounding volumes to test based on the ray's time of intersection with the scene. Rather than testing all bounding volumes uniformly, the system queries only the relevant spatial bounds at the specific time(s) when the ray intersects the scene, optimizing the number of intersection tests while ensuring all possible primitive positions are covered
Data Source
AI summary
Techniques are disclosed relating to ray tracing and motion blur in graphics processors. In some embodiments, ray intersection accelerator circuitry is configured to perform traversal operations for an acceleration data structure (ADS) that includes hierarchical bounding volumes, where the ADS includes a first node that specifies: first coordinates of a bounding volume at a first motion blur time and second coordinates of the bounding volume at a second motion blur time. The Ray accelerator may determine a ray time interval representation indicating upper and lower bounds that represent error bounds for a quantized ray time value of a ray. Interpolation circuitry may operate on the first coordinates, the second coordinates, and the ray time interval representation to generate interpolated coordinates for the bounding volume. Box test circuitry may determine whether the ray intersected the bounding volume based on the interpolated coordinates and coordinates of the ray.