Page 9 of 25
Re: CGI tech news box
Posted: Tue Jan 29, 2019 8:49 pm
by B.Y.O.B.
I have split the posts about OpenImageDenoise into a new topic:
viewtopic.php?f=10&t=872
Re: CGI tech news box
Posted: Mon Feb 11, 2019 10:53 am
by lacilaci
Re: CGI tech news box
Posted: Sat Feb 23, 2019 2:51 pm
by kintuX
The Unreasonable Effectiveness of Quasirandom Sequences
"I present a new low discrepancy quasirandom sequence that offers substantial improvement over current state-of-the art sequences eg Sobol, Niederreiter,…"
http://extremelearning.com.au/unreasonable-effectiveness-of-quasirandom-sequences/
Re: CGI tech news box
Posted: Sat Feb 23, 2019 3:31 pm
by Dade
Uh, I looked and tried the code, it is incredible simple. It may be worth trying even if people where complaining of regular Sobol patterns ... I don't want even start to think what they will say about this.
Re: CGI tech news box
Posted: Sun Feb 24, 2019 7:40 am
by lacilaci
This looks nice, I don't mind pattern but don't you think it could affect denoiser?
And if it would work with oidn, then maybe R-dithered could too? They claim there that it is comparable with blue noise dithered masks in terms of performance...
Re: CGI tech news box
Posted: Sun Feb 24, 2019 11:21 am
by kintuX
Well, i find R pattern to behave more natural (samples sorted / space filled alike with grains in photographic film coating) and believe it would work just fine with OIDN. I also assume less memory consumption.
On the last note, maybe this could also help?
"Mesh - A memory allocator that automatically reduces the memory footprint of C/C++ applications."
https://github.com/plasma-umass/Mesh
Oh, almost forgot about
Progressive Transient Photon Beams
https://cs.dartmouth.edu/~wjarosz/publi ... ssive.html
Abstract In this work, we introduce a novel algorithm for transient rendering in participating media. Our method is consistent, robust and is able to generate animations of time-resolved light transport featuring complex caustic light paths in media. We base our method on the observation that the spatial continuity provides an increased coverage of the temporal domain, and generalize photon beams to transient-state. We extend stead-state photon beam radiance estimates to include the temporal domain. Then, we develop a progressive variant of our approach which provably converges to the correct solution using finite memory by averaging independent realizations of the estimates with progressively reduced kernel bandwidths. We derive the optimal convergence rates accounting for space and time kernels, and demonstrate our method against previous consistent transient rendering methods for participating media.
Re: CGI tech news box
Posted: Sun Feb 24, 2019 7:59 pm
by goofoo
Evenly Distributing Points in a Triangle.
"Most two dimensional quasirandom methods focus on sampling over a unit square. However, sampling evenly over the triangle is also very important in computer graphics. Therefore, I describe a simple and direct construction method for a point sequence to evenly cover an arbitrary shaped triangle."
http://extremelearning.com.au/evenly-di ... -triangle/
Re: CGI tech news box
Posted: Sun Feb 24, 2019 8:15 pm
by B.Y.O.B.
kintuX wrote: ↑Sun Feb 24, 2019 11:21 am
Oh, almost forgot about Progressive Transient Photon Beams
Their video is pretty awesome!
Re: CGI tech news box
Posted: Sun Mar 03, 2019 9:40 pm
by kintuX
B.Y.O.B. wrote: ↑Sun Feb 24, 2019 8:15 pm
Their video is pretty awesome!
Was blown away myself... amazing to see how photons travel, often i try to imagine how it would look expanded (visible+IR+UV) or full EM spectrum
then i remember it is what we "see" or rather experience... light and excess Energy materialized, just so damn fast only rare can imagine interactivity of it all. Ever flowing.
Ah well, back to science, this time test results from different samplers for "proper" comparative studies:
SampleZoo - Simple implementations of sampling patterns, with objective tests to compare them
... and
Practical Path Guiding for Efficient Light-Transport Simulation
Video &
Online test suite with an interactive viewer
Re: CGI tech news box
Posted: Mon Mar 04, 2019 7:30 am
by epilectrolytics
Thanks for that link!
Apparently this is accelerating unbiased path tracing including SDS paths.
If it is feasible to implement this I'd suggest to ditch the biased caustic branch of PGI in favour of this.