Hi,
I just read this tweet:
https://twitter.com/max_tarpini/status/ ... 1328095232
Could this be interesting to speed up (caustics) rendering in LuxCoreRender?
Langevin MonteCarlo rendering
- MetinSeven
- Posts: 137
- Joined: Sun Aug 18, 2019 10:19 am
- Location: Netherlands
- Contact:
Langevin MonteCarlo rendering
visualizer • illustrator • animator • 3D designer — metinseven.nl
Re: Langevin MonteCarlo rendering
https://research.cs.cornell.edu/langevin-mcmc/
If it is an online learning method will be hard as Dade said before.
It is a trend in the industry to go through the road of online learning method the problem is that it is almost CPU only.
Weta Manuka render and disney hyperion have it ( corona render also). It is very efficient and flicker free ( Animation paradise
). But un friendly to GPU (Coding hell
).
SIGGRAPH 2019 Course : https://tom94.net/data/courses/vorba19g ... uiding.pdf
It look simillar to the other path guiding algo seen before from my POV but only a dev can tell if it is compatible with Luxcore GPU arch. And also if it is easy enought to be implemented in lux code.We introduce a suite of Langevin Monte Carlo algorithms for efficient photorealistic rendering of scenes with complex light transport effects, such as caustics, interreflections, and occlusions. Our algorithms operate in primary sample space, and use the Metropolis-adjusted Langevin algorithm (MALA) to generate new samples. Drawing inspiration from state-of-the-art stochastic gradient descent procedures, we combine MALA with adaptive preconditioning and momentum schemes that re-use previously-computed first-order gradients, either in an online or in a cache-driven fashion. This combination allows MALA to adapt to the local geometry of the primary sample space, without the computational overhead associated with previous Hessian-based adaptation algorithms. We use the theory of controlled Markov chain Monte Carlo to ensure that these combinations remain ergodic, and are therefore suitable for unbiased Monte Carlo rendering. Through extensive experiments, we show that our algorithms, MALA with online and cache-driven adaptation, can successfully handle complex light transport in a large variety of scenes, leading to improved performance (on average more than 3× variance reduction at equal time, and 7× for motion blur) compared to state-of-the-art Markov chain Monte Carlo rendering algorithms.
If it is an online learning method will be hard as Dade said before.
It is a trend in the industry to go through the road of online learning method the problem is that it is almost CPU only.
Weta Manuka render and disney hyperion have it ( corona render also). It is very efficient and flicker free ( Animation paradise
![Very Happy :D](./images/smilies/icon_e_biggrin.gif)
![Twisted Evil :twisted:](./images/smilies/icon_twisted.gif)
SIGGRAPH 2019 Course : https://tom94.net/data/courses/vorba19g ... uiding.pdf
- MetinSeven
- Posts: 137
- Joined: Sun Aug 18, 2019 10:19 am
- Location: Netherlands
- Contact:
Re: Langevin MonteCarlo rendering
Hmm, it would be a pity if this is GPU-unfriendly. ![Sad :(](./images/smilies/icon_e_sad.gif)
![Sad :(](./images/smilies/icon_e_sad.gif)
visualizer • illustrator • animator • 3D designer — metinseven.nl