Real-time Global Light Transport via Pre-computed Transfer Functions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Rendering global light transport in real-time is computationally intensive due to the need to simulate multiple light bounces and handle dynamic scenes with local light sources, which exceeds the resources available on most computing devices.
Innovation Solution
Pre-computation of machine learning models that learn point regression functions to determine indirect shading values, allowing for real-time rendering of diffuse and specular components using modest computational resources, enabling efficient evaluation on devices like smartphones and gaming consoles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional techniques are used to render indirect illumination with multiple light bounces, then rendering quality is improved, but computational resources and time required increase prohibitively
Solution Approach 1:
The patent pre-computes transfer functions that model indirect illumination behavior for various lighting conditions and scene configurations during an offline preparation phase. These pre-computed functions are then applied in real-time rendering, allowing high-quality indirect illumination effects to be achieved without performing computationally intensive multiple bounce simulations during the actual rendering process, thus resolving the contradiction between rendering quality and rendering speed
Solution Approach 2:
The patent creates simplified mathematical models (transfer functions) that copy and approximate the complex behavior of multiple light bounce simulations. Instead of performing actual physical light transport simulations during rendering, the system uses these pre-computed functional copies to reproduce the visual effects of indirect illumination, achieving both quality and real-time performance
2Adaptability or versatility
If pre-computed light transport data is stored for each scene configuration to handle dynamic scenes, then adaptability is improved, but storage requirements increase to several hundred megabytes
Solution Approach 1:
The patent changes the representation parameters from storing complete pre-computed light transport data (requiring hundreds of megabytes) to storing compact transfer functions with a limited number of coefficients. This parameter transformation allows the system to maintain adaptability for dynamic scenes while reducing storage requirements to a manageable size, as the functions can be evaluated for any scene configuration without storing exhaustive pre-computation results
3Manufacturing precision
If multiple light bounces are simulated to achieve photorealistic rendering, then visual richness is improved, but computational complexity increases beyond available resources
Solution Approach 1:
The patent extracts the computationally intensive multiple bounce simulation process from the real-time rendering pipeline and separates it into an offline pre-computation phase. The complex physics simulations are performed once beforehand to build transfer functions, which are then used during real-time rendering without requiring the full computational resources, thus resolving the contradiction between visual richness and device complexity
Data Source
AI summary
Some implementations disclosed herein provide techniques and arrangements to render global light transport in real-time or near real-time. For example, in a pre-computation stage, a first computing device may render points of surfaces (e.g., using multiple light bounces and the like). Attributes for each of the points may be determined. A plurality of machine learning algorithms may be trained using particular attributes from the attributes. For example, a first machine learning algorithm may be trained using a first portion of the attributes and a second machine learning algorithm may be trained using a second portion of the attributes. The trained machine learning algorithms may be used by a second computing device to render components (e.g., diffuse and specular components) of indirect shading in real-time.