8DNA: 8D Neural Asset Light Transport by Distribution Learning

2026-04-27

AI Summary

This work addresses the high computational cost of simulating global illumination effects—such as subsurface scattering, specular interreflections, and fibrous scattering—in high-fidelity 3D assets, which involve long scattering paths. The authors propose, for the first time, a complete neural representation of 8D light transport, overcoming the limitation of prior methods that support only 6D far-field illumination. By pre-baking light transport effects into a neural radiance field through distribution learning and training exclusively on forward path-traced samples, the method enables high-quality rendering under near-field lighting. It achieves close visual agreement with reference path-traced results across diverse complex scenes while significantly reducing variance and accelerating inference.

Abstract

High-fidelity 3D assets exhibit intriguing global illumination effects like subsurface scattering, glossy interreflections, and fine-scale fiber scatterings, which often involve long scattering paths that are expensive to simulate. We introduce 8D neural assets (8DNA) to pre-bake these light transport effects into neural representations. Unlike prior methods that assume far-field lighting and precompute light transport into 6D functions, 8DNA learns the full 8D light transport, enabling accurate rendering under near-field illumination. Our training leverages a distribution-learning formulation that learns light transport from forward path-traced samples, which produces less optimization variance with lower training budget than the prior regression-based approaches. Experiments show our 8DNA rendering closely matches path-traced results under various scene configurations, yet it achieves improved variance reduction and fast inference speeds on challenging assets.

Problem

• light transport
• global illumination
• neural representation
• near-field illumination
• high-fidelity rendering

Innovation

• 8D light transport
• neural assets
• distribution learning
• near-field illumination
• global illumination