Massively Parallel Rendering of Complex Closed-Form Implicit Surfaces (2020)
1–10 of 19 posts
Re: Massively Parallel Rendering of Complex Closed-Form Implicit Surfaces (2020)
#2Really cool!
Need neural radiance fields. Then add super resolution, then add motion prediction and you are on your way to a synthetic visual cortex.
In the future GPUs will be spec'd by how far in the future they can predict given a power and thermal envelope.
Re: Massively Parallel Rendering of Complex Closed-Form Implicit Surfaces (2020)
#3Super cool to see Matthew's library as well: https://libfive.com/studio/
Re: Massively Parallel Rendering of Complex Closed-Form Implicit Surfaces (2020)
#4This is important for vision ML because sufficiently advanced simulations may comprise the real world as a special case: https://arxiv.org/abs/1703.06907
For example, gaming pixel maps have been used for semantic simulation, and have been rendered scenes: https://arxiv.org/abs/1608.02192 https://www.cv-foundation.org/openaccess/content_cvpr_2016/p...
This concept has not (yet) been applied in audio ML. We have a paper in submission---will be on ArXiv soon---where we share a GPU-enabled modular synthesizer that is 16000x faster than realtime, concurrently released with a 1-billion audio sample corpus that is 100x larger than any audio dataset in the literature. Here's the code: https://github.com/torchsynth/torchsynth
Re: Massively Parallel Rendering of Complex Closed-Form Implicit Surfaces (2020)
#5Very very impressive work. I wonder if it would be feasible for vendors to implement the interpreter in hardware in future.
Re: Massively Parallel Rendering of Complex Closed-Form Implicit Surfaces (2020)
#6Looking forward to watching this when I have some down time... But, whew, I did a double-take when I saw "VR/ML workstation"
Re: Massively Parallel Rendering of Complex Closed-Form Implicit Surfaces (2020)
#7Fabulous work. The video presentation is only 18 minutes long, well organized, and very accessible -- the author does a great job of explaining how and why the rendering works so efficiently using a simple example in 2D:
https://www.youtube.com/watch?v=_6CnaugAcCc
Highly recommended.
Re: Massively Parallel Rendering of Complex Closed-Form Implicit Surfaces (2020)
#8This is important for vision ML because sufficiently advanced simulations may comprise the real world as a special case: https://arxiv.org/abs/1703.06907 For example, gaming pixel maps have been used for semantic simulation, and have been rendered scenes: https://arxiv.org/abs/1608.02192 https://www.cv-foundation.org/openaccess/content_cvpr_2016/p... This concept has not (yet) been applied in audio ML. We have a pape…
How is it related to ML?
Closed form implicit surfaces are actually one of the most difficult way to model the real world. They are neat because they are very compact and the creation process is close to modeling with (mathematical) clay. But they are hard to use if you want to model the real world with all its complexity, resulting of a variety of chemical and physical processes happening over time. There is a reason why they are so popular in the demoscene, for which technical achievement and art is more important than realism, and not much elsewhere.
The paper is about making rendering of these primitives more efficient, which may prove to be a great addition to an artist toolbox, and maybe for scientific imagery. However, I don't really see applications for ML anytime soon.
Re: Massively Parallel Rendering of Complex Closed-Form Implicit Surfaces (2020)
#9Fabulous work. The video presentation is only 18 minutes long, well organized, and very accessible -- the author does a great job of explaining how and why the rendering works so efficiently using a simple example in 2D: https://www.youtube.com/watch?v=_6CnaugAcCc Highly recommended.
The 2D part is well explained, however, I am interested in how they made it 3D. They have voxels, but the actual rendering is a bit unclear.
I guess they simply didn't do much work on that part, using some bruteforce-ish raymarching technique, which their fast evaluation and nicely bound objects allows. They mention further work though, like sparse voxel octrees, improving culling, etc... So I guess that will be for a "future episode".
Re: Massively Parallel Rendering of Complex Closed-Form Implicit Surfaces (2020)
#10Fabulous work. The video presentation is only 18 minutes long, well organized, and very accessible -- the author does a great job of explaining how and why the rendering works so efficiently using a simple example in 2D: https://www.youtube.com/watch?v=_6CnaugAcCc Highly recommended.
The 2D part is well explained, however, I am interested in how they made it 3D. They have voxels, but the actual rendering is a bit unclear. I guess they simply didn't do much work on that part, using some bruteforce-ish raymarching technique, which their fast evaluation and nicely bound objects allows. They mention further work though, like sparse voxel octrees, improving culling, etc... So I guess that will be for…
Check out section 4.2 of the paper: 3D is very similar to 2D, using the same strategy of big regions -> small regions -> voxels.
(There's also a bit of extra logic to skip regions which are occluded in Z, plus a final pass to render normals using automatic differentiation)