Mitsuba 2 is one of the few of research-oriented rendering system that supports GPU. In the past creating rendered data for deep learning was the bottelneck since path-tracing with ten-thousands of samples consumes a lot of time on CPU. Mitsuba 2 allows to render scenes using CUDA and Optix which reduces the rendering time.
To render scenes with various camara settings and materials, I managed to add and fix some parts of mitsuba2. Modifications and additions might be based on my onwn implementations, or other people's pull requests. I will mention the original implementations. Please see the issues and pull requests for more informations.
- Sunsky
- Coating
- Hair
- Cube
- Envmap
- SRGBReflectance
Please see the documentation for details on how to compile, use, and extend Mitsuba 2.
Mitsuba2 was created by Wenzel Jakob. Significant features and/or improvements to the code were contributed by Merlin Nimier-David, Guillaume Loubet, Benoît Ruiz, Sébastien Speierer, Delio Vicini, and Tizian Zeltner.