Setup for the comparison of different path guiding optimizations. The implementation is based on OpenPGL.NET and SeeSharp by Pascal Grittmann.
Precompiled binaries for openpgl and its dependencies can be downloaded by running the make.ps1 (all platforms) or make.sh (Linux and OSX) scripts. If you want to supply your own binaries, check these scripts for details.
Currently, this project focusses on several methods of optimizing the guiding probability. Current implementations are based on:
- Jiří Vorba, Johannes Hanika, Sebastian Herholz, Thomas Müller, Jaroslav Křivánek, and Alexander Keller. 2019. Path guiding in production. In ACM SIGGRAPH 2019 Courses (SIGGRAPH '19). Article 18, 1–77. [https://doi.org/10.1145/3305366.3328091]
- Pascal Grittmann, Ömercan Yazici, Iliyan Georgiev, and Philipp Slusallek. 2022. Efficiency-aware multiple importance sampling for bidirectional rendering algorithms. ACM Trans. Graph. 41, 4, Article 80, 12 pages. [https://doi.org/10.1145/3528223.3530126]
- Mateu Sbert, Vlastimil Havran and László Szirmay-Kalos. 2019. Optimal Deterministic Mixture Sampling. Eurographics 2019 - Short Papers. [https://diglib.eg.org/handle/10.2312/egs20191018]