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

TL;DR

8DNA introduces an 8D neural representation for detailed 3D asset rendering, capturing complex light effects under near-field illumination with efficient training and inference.

Contribution

It learns full 8D light transport using distribution learning, surpassing prior 6D methods in accuracy and efficiency for complex global illumination effects.

Findings

Rendering closely matches path-traced results across scenes.

Achieves lower training variance with fewer samples.

Provides faster inference on challenging assets.

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.