Combining Resampled Importance and Projected Solid Angle Samplings for Many Area Light Rendering

Abstract

Direct lighting from many area light sources is challenging due to variance from both choosing an important light and then a point on it. Resampled Importance Sampling (RIS) achieves low variance in such situations, however it is limited to simple sampling strategies for it's candidates. Specifically for area lights, we can improve the convergence of RIS by incorporating a better sampling strategy: Projected Solid Angle Sampling (ProjLTC). Naively combining RIS and ProjLTC improves equal sample convergence, however achieves little to no gain in equal time. We identify the core issue for the high run-times and reformulate RIS for a better integration with ProjLTC. Our method achieves better convergence and results in both equal sample and equal time. We evaluate our method on challenging scenes with varying number of area light sources and compare to uniform sampling, RIS and ProjLTC. In all cases, our method seldom performs worse than RIS and often performs better.

Implementation

Our algorithm is very simple to implement and requires slight modifications to the original RIS+WRS [Bitterli, 2020] provided implementations of LTC [Heitz, 2016] and ProjLTC [Peters, 2021] are given. We provide the modifications required below. The complete source code can be found in the following repository.

Results

We use three scenes with varying number and sizes of area lights for the evaluation: (1) Zero-Day, 10K lights, (2) Bistro Interior, 2K lights and (3) Bistro Exterior, 30K lights. For each scene, we test the method on a diffuse and glossy version of the scene. We compare our results against the following four methods:

Uniform: Uniform light & uniform area sampling
ProjLTC: Uniform light sampling & Projecte Solid Angle Sampling [Peters, 2021]
RIS: RIS with uniform light & uniform area sampling [Tablot 2005, Bitterli 2020]
RIS-ProjLTC: Naïve combination of RIS with ProjLTC

Equal Time Comparsion

Equal Sample Comparsion

Acknowledgments

We thank the reviewers of our SIGGRAPH Asia 2023 submission for their valuable comments and suggestions. The renderer was built upon the amazing vulkan_renderer by Christoph Peters. The image viewer is based upon interactive-viewer.

References

Justin F. Talbot, David Cline, and Parris Egbert. 2005. Importance Resampling for Global Illumination. In EGSR.
Benedikt Bitterli, Chris Wyman, Matt Pharr, Peter Shirley, Aaron Lefohn, and Wojciech Jarosz. 2020. Spatiotemporal reservoir resampling for real-time ray tracing with dynamic direct lighting. ACM Transactions on Graphics 39, 4 (2020).
Eric Heitz, Jonathan Dupuy, Stephen Hill, and David Neubelt. 2016. Real-Time Polygonal-Light Shading with Linearly Transformed Cosines. ACM Trans. Graph. 35, 4, Article 41 (Jul 2016), 8 pages. https://doi.org/10.1145/2897824.2925895
Christoph Peters. 2021. BRDF Importance Sampling for Polygonal Lights. ACM Trans. Graph. 40, 4, Article 140 (jul 2021), 14 pages.

BibTeX

@inproceedings{10.1145/3610543.3626165,
	author = {Shah, Ishaan Nikhil and KT, Aakash and Narayanan, P. J.},
	title = {Combining Resampled Importance and Projected Solid Angle Samplings for Many Area Light Rendering},
	year = {2023},
	isbn = {9798400703140},
	publisher = {Association for Computing Machinery},
	address = {New York, NY, USA},
	url = {https://doi.org/10.1145/3610543.3626165},
	doi = {10.1145/3610543.3626165},
	booktitle = {SIGGRAPH Asia 2023 Technical Communications},
	articleno = {2},
	numpages = {4},
	location = {, Sydney, NSW, Australia, },
	series = {SA '23}
}