Temporal BVH Structures for Motion-Blur Ray Tracing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Ray tracing in graphics processing is time-consuming, especially when dealing with dynamic content, and rendering motion blur is challenging due to the complexity of finding object intersections at varying times during the camera's shutter time, particularly for moving objects like rotor blades.

Innovation Solution

The use of spatial and temporal data structures in a ray tracing architecture to efficiently handle dynamic content and motion blur by integrating over the camera's shutter time, improving the ray tracing kernel's ability to find proper intersections.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If Monte Carlo path tracing is used to handle motion blur by integrating over camera shutter time, then rendering quality is improved, but computational complexity and processing time increase significantly

Engineering Contradiction:
Improverendering qualityVSAvoidcomputational complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the temporal domain by dividing the camera shutter time into discrete time samples. Instead of continuously integrating over shutter time, the system samples at specific time points and combines results, transforming a complex continuous integration problem into manageable discrete segments that reduce computational complexity while maintaining rendering quality

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent pre-computes and stores object positions at multiple time samples before ray tracing. By preparing temporal data structures in advance that contain object positions at different time points, the system avoids real-time computation of moving object positions during ray tracing, significantly reducing computational complexity during the actual rendering process

Inventive Principle:
Principle #10Preliminary action

2Productivity

If traditional spatial data structures are used for static scenes, then intersection finding is efficient, but handling dynamic content with moving objects becomes challenging

Engineering Contradiction:
Improveintersection finding efficiencyVSAvoidhandling dynamic content
Core Design Contradiction:
ProductivityVSAdaptability or versatility

Solution Approach 1:

The patent extends traditional spatial data structures by adding a temporal dimension. The hierarchical spatial structure is augmented with time samples, creating a four-dimensional data structure that organizes objects not only in space but also across time. This allows efficient querying of object positions at any time point while maintaining the spatial hierarchy's intersection finding efficiency

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Solution Approach 2:

The patent transforms static spatial data structures into dynamic structures that can adapt to moving objects. By incorporating temporal sampling and allowing the data structure to be rebuilt or updated at different time points, the system maintains efficiency for both static and dynamic content, with the structure adapting to the specific temporal requirements of each rendering pass

Inventive Principle:
Principle #15Dynamics

3Measurement precision

If more time samples are used for motion blur integration, then rendering accuracy is improved, but processing time increases

Engineering Contradiction:
Improvemotion blur accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent pre-computes object positions at multiple time samples and stores them in temporal data structures before ray tracing begins. By preparing this temporal information in advance, the system avoids repeated computation of object positions for each ray and time sample combination, significantly reducing processing time while maintaining the ability to use multiple time samples for accurate motion blur rendering

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent implements adaptive temporal sampling where the number of time samples used can be adjusted based on scene requirements. For scenes with minimal motion, fewer samples suffice, reducing processing time. For scenes with significant motion or when higher quality is needed, more samples can be used, providing flexibility to balance accuracy and processing time based on specific rendering needs

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentUS20250299417A1Temporal data structures in a ray tracing architecture
Publication Date: 2025.09.25 INTEL CORP
  • US20250299417A1 patent drawing
  • US20250299417A1 patent drawing
  • US20250299417A1 patent drawing

AI summary

A graphics processing apparatus comprising bounding volume hierarchy (BVH) construction circuitry to perform a spatial analysis and temporal analysis related to a plurality of input primitives and responsively generate a BVH comprising spatial, temporal, and spatial-temporal components that are hierarchically arranged, wherein the spatial components include a plurality of spatial nodes with children, the spatial nodes bounding the children using spatial bounds, and the temporal components comprise temporal nodes with children, the temporal nodes bounding their children using temporal bounds and the spatial-temporal components comprise spatial-temporal nodes with children, the spatial-temporal nodes bounding their children using spatial and temporal bounds; and ray traversal/intersection circuitry to traverse a ray or a set of rays through the BVH in accordance with the spatial and temporal components.