Neural BRDF Integration and Analysis in Pixar’s RenderMan
DOI:
https://doi.org/10.58445/bj6pp502Keywords:
rendering, Artificial Intelligence, animationAbstract
Capturing both the detail and accuracy of a complex scene while also ensuring an efficient and cost-effective rendering model has become an aspiring goal for the field of computer graphics. A bi-directional reflection distribution function is the ratio of reflected radiance in direction ωo to incident irradiance from direction ωi. (Nicodemus, F. E., Richmond et al., 1977, 3). An attempt at creating a BRDF with neural networks has been attempted in recent years, mainly with high-end hardware and high computational costs. We investigated the mathematics of physically-based shading and why evaluating those models during ray/path tracing is computationally expensive. We identified the specific mathematical operations in a multi-layer BRDF evaluation that create computational bottlenecks during path tracing and used RenderMan to see these lighting terms in action, and analyzed whether a neural network could more efficiently render in an offline path renderer’s architecture. We present an alternative application of a NBRDF incorporated into Pixar’s RenderMan trained on the MERL dataset for metallic materials. We evaluate the benefits and drawbacks of this neural network and also propose an extension of the work to include a greater all-in-one neural renderer that is usable by the average consumer.
References
AMD GPUOpen / Kazmierczyk, K., et al. (2024). Neural supersampling and denoising for real-time path tracing. AMD GPUOpen. https://gpuopen.com/learn/neural_supersampling_and_denoising_for_real-time_path_tracing/
Bauer, D., Wu, Q., Gadirov, H., & Ma, K.-L. (2025). GSCache: Real-time radiance caching for volume path tracing using 3D Gaussian splatting. arXiv:2507.19718.
Burley, B. (2012). Physically based shading at Disney. In ACM SIGGRAPH 2012 Course Notes.
Christensen, P. H., Fong, J., Shade, J., Wooten, W. L., Schubert, B., Kensler, A., Friedman, S., Kilpatrick, C., Ramshaw, C., Bannister, M., Rayner, B., Brouillat, J., & Liani, M. (2018). RenderMan: An advanced path-tracing architecture for movie rendering. ACM Transactions on Graphics, 37(3), Article 30. https://doi.org/10.1145/3182162
Immel, D. S., Cohen, M. F., & Greenberg, D. P. (1986). A radiosity method for non-diffuse environments. ACM SIGGRAPH Computer Graphics, 20(4), 133–142. https://doi.org/10.1145/15922.15901
Kajiya, J. T. (1986). The rendering equation. In Computer Graphics (Proceedings of SIGGRAPH ‘86), 20(4), 143–150. https://doi.org/10.1145/15886.15902
Matusik, W., Pfister, H., Brand, M., & McMillan, L. (2003). A data-driven reflectance model. ACM Transactions on Graphics, 22(3), 759–769. https://doi.org/10.1145/882262.882343
Nicodemus, F. E., Richmond, J. C., Hsia, J. J., Ginsberg, I. W., & Limperis, T. (1977). Geometrical considerations and nomenclature for reflectance (NBS Monograph 160). National Bureau of Standards.
Sztrajman, A., Rainer, G., Ritschel, T., & Weyrich, T. (2021). Neural BRDF representation and importance sampling. Computer Graphics Forum, 40(6), 332–346. https://doi.org/10.1111/cgf.14335
Veach, E. (1997). Robust Monte Carlo methods for light transport simulation [Doctoral dissertation, Stanford University]. http://graphics.stanford.edu/papers/veach_thesis/
Walter, B., Marschner, S. R., Li, H., & Torrance, K. E. (2007). Microfacet models for refraction through rough surfaces. In Proceedings of the 18th Eurographics Conference on Rendering Techniques (pp. 195–206).
Zeltner, T., Rousselle, F., Weidlich, A., Clarberg, P., Novák, J., Bitterli, B., Evans, A., Davidovič, T., Kallweit, S., & Lefohn, A. (2024). Real-time neural appearance models. ACM Transactions on Graphics, 43(5), Article 118. https://doi.org/10.1145/3659577
Downloads
Posted
Categories
License
Copyright (c) 2026 Research Archive of Rising Scholars

This work is licensed under a Creative Commons Attribution 4.0 International License.