Adaptive Pixel Sampling for Temporal Rendering Artifacts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current ray tracing techniques for computer-generated imagery face challenges in achieving high refresh rates and efficient rendering, particularly in real-time applications, due to limitations in spatial and temporal sampling rates, leading to artifacts like flickering and blurring.
Innovation Solution
The proposed solution involves a dynamic per-tile sampling order based on motion vectors, using techniques like hourglass and bowtie patterns to optimize pixel ordering, and combining subframes with prior frames to create full-resolution frames, reducing spatial and temporal aliasing effects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If temporal antialiasing uses regular subpixel offsets and accumulates samples over many frames, then visual quality is improved, but rendering speed decreases and computational complexity increases
Solution Approach 1:
The patent applies dynamics by making the sampling pattern adaptive rather than static. The system dynamically adjusts the sampling order based on motion vectors detected in the scene, switching between different sampling patterns (e.g., hourglass, bowtie) depending on the motion characteristics. This allows the system to optimize between visual quality and rendering speed in real-time based on actual scene content.
Solution Approach 2:
The patent changes the sampling pattern parameter based on motion vector magnitude. When motion vectors indicate significant movement, the system switches to sampling patterns that better handle motion artifacts. This parameter adaptation allows the system to maintain visual quality while reducing the number of frames needed for accumulation, thereby improving rendering speed.
2Measurement precision
If ray tracing performs continuous updates of world simulation for each sample, then temporal resolution is improved, but computational complexity and rendering time increase
Solution Approach 1:
The patent segments the pixel array into multiple tiles, each with its own sampling pattern. This segmentation allows different regions of the image to be processed independently with optimized sampling orders based on local motion characteristics. By dividing the large computational task into smaller tile-level operations, the system reduces overall computational complexity while maintaining high temporal resolution.
Solution Approach 2:
The system applies partial action by selectively applying different sampling densities to different regions. Rather than uniformly sampling all pixels at maximum resolution, the system adjusts sampling intensity based on motion vector magnitude in each region, applying higher sampling rates only where motion indicates it is necessary for quality maintenance.
3Ease of manufacture
If a fixed sampling pattern is used for all pixels, then implementation simplicity is maintained, but visual quality deteriorates due to aliasing and flickering
Solution Approach 1:
The patent implements local quality by assigning different sampling patterns to different pixel regions based on their motion characteristics. Each tile or region receives a sampling pattern optimized for its specific motion vector properties, rather than applying a uniform pattern across the entire image. This significantly improves visual quality by reducing aliasing and flickering in motion-prone areas while maintaining implementation simplicity through systematic pattern assignment.
Data Source
AI summary
A method dynamically selects one of a first sampling order and a second sampling order for a ray trace of pixels in a tile where the selection is based on a motion vector for the tile. The sampling order may be a bowtie pattern or an hourglass pattern.