- User Since
- Jun 2 2009, 2:10 PM (428 w, 6 d)
Jun 30 2017
Ideally the adaptive sampling should take color management into account indeed, though in general it's unknown what kind of image transforms might be applied in compositing. A simple color_scene_linear_to_srgb() or sqrt() might already see a noticeable improvement.
Jun 29 2017
Frankly, all the tests look absolutely gorgeous, not just the headlights :)
What I miss is a comparison at same render time: plain vs adaptive.
The original render takes hours but deciding what is 100% quality sounds arbitrary here.
One thing I didn't understand: is the check done at pixel level or at the whole tile?
May 22 2017
So I understand that Norm here is a sort of threshold, right? Then, setting Norm, is like setting the "main level of smoothness" of the image.
Could this include a time-halt condition?
Apr 12 2017
I found a weird shading bug for wide surfaces. It behaves alike when using procedural texture for bump mapping or in displacement slot
I found a weird shading bug for wide surfaces. The black reflections come back in this file: http://pasteall.org/blend/index.php?id=46239
Mar 9 2017
Occasionally testing this branch (build 26e906d) it appears to have a major bug rendering SSS with Branched Path Tracing, as shown in pictures here: https://blenderartists.org/forum/showthread.php?403342-Cycles-Disney-Brdf&p=3162218&viewfull=1#post3162218. Basically branched takes care of base color only, as if it is ignoring the SSS color input.
Mar 2 2016
Sep 28 2015
Aug 29 2015
May 5 2015
I think that a feature like this would be better than making two or more renderlayers and compositing just because two render layers mean two different renders in a row, which means more time. I don't know how blender works under the hood, I imagined that with material override there's some "switch material" at the end for every object, so I thought that it would be easy to just mark some objects to be excluded from this switch.
Hi Lukas, great to see you giving some time to this, thank you a lot!
As posted on BA, there could be another option for this, that is having the override driven by a group instead. What do you devs think is more appropriate?
Apr 6 2015
Sorry if this can be seen as a feature request: currently the rendered preview redraws on *any* change happening in the node-tree even if the node added/removed/changed is not connected or evaluated (mix=0/1 conditions). This is sometimes annoying since long and clean previews get cleared (and slow the system down while re-rendering) while one is making some node operation that don't affect the final material appearance.
For the sake of this kind of optimizations, do you think that this can be addressed?
Mar 24 2015
Feb 27 2015
Hi, being a former Maxwell user I can say how much progressive refine can be useful, giving you the chance to stop the process anytime and have a full-frame render anyway. The plus in Maxwell is that for each step it actually writes a temp image, to keep the job safe in case of a crash or power failure.
Oct 17 2014
Hi Lukas, regarding the UI debate, what about adding just a slider named "AA samples multiplier"? It could come along with a text exposing the evaluated total, as well as how it happens in the bottom of the samples panel when you enable branched path tracing.
Also, I don't know if the word Confidence comes from the papers or from some JND/MonteCarlo/math-related glossary, but for us final users wouldn't be Convergence or Smoothness more obvious?
Oct 13 2014
Sorry, I forgot to mention, the behavior is both in GPU and CPU rendering
Oct 1 2014
Jul 8 2014
Still i can't see benefits in my daily usage ...probably beacause it's not clear to me what this patch is actually doing!
Jun 24 2014
Anyway. How can Blender users have a try at your patch?
Hi, I'm quite interested in these adaptive solutions.
How do you manage to avoid visible tile boundaries in a low sampled render?
Using (entire)tile stop condition easily lead to have adiacent tiles with equal threshold but visible difference in terms of noise, as I experienced with the other experimental method developed by likas.
Jun 10 2014
About me asking earlier if it was possible to have per-pixel stop-condition, I thought of a workflow like this:
May 8 2014
Yes indeed, my tiny test showed way lower speed gains. Indeed almost no speed gain for bit-noisy renders. But this is probably my fault.
Do I read well: now your CPU rendering is 4 times faster than GPU... instead of 8 times slower??
This means 32 times faster in CPU/CPU comparison!! Or in other words, 3200% speed boost, or also rendering in ~3% of the time!
Ok, let's say that I misunderstand something here...
May 5 2014
How hard - if ever possible - is it to implement per-pixel stop condition?
Apr 17 2014
as m9105826 says. test could be done on a more difficult scene as this: http://www.blenderartists.org/forum/showthread.php?331149-The-new-Cyles-GPU-2-70-Benchmark
Lukas, what is actually the difference between "adaptive stopping" and "adaptive stopping+sampling" ?
Apr 10 2014
IIRC LuxCore implementation uses max error inside each tile, and it sounds reasonable too. This way you always know how much the error might be.
Can't wait to test this out. Can you (ora anyone) please upload the build once it is ready?
Apr 5 2014
There is also one more opportunity: time halt condition. To make this possible the engine should know how long each tile takes to render, so a first pass (with a minimum sample number to be decided) should be done for all the tiles. This would give two benefits: a proper render-time estimation, and a first rough look at the whole render output.
Apr 4 2014
@Jason Clarke (jtheninja)
I actually don't know ;)
BTW, Yafaray has something similar since ages. Wondering what is the noise-aware implementation there...
And it's what I've been waiting for a looong time... :)
Mar 31 2014
Yep thumbnails are better than nothing, but often i feel the lack of a good (big) view of textures (Not to speak about the time needed to see all the thumbnails in a crowded folder.). So I have to reach the folder via system and search the images with an external app (a trivial viewer). Having a view that is fullsize and scrollable would be a gift. A thought this because Blender already has its own viewer
Mar 26 2014
Any chance to have a preview area in the file browser?
Feb 21 2014
@Lukas Stockner (lukasstockner97) no, unfortunately i have no clue about performance
@Jason Clarke (jtheninja) can you please make another BMW test with no adaptive, letting cycles run for 3:03 as well, so we can see how better the algorithm distributed the samples in the same amount of time?
Feb 20 2014
I actually did a search... Google this: "entropy based adaptive sampling paper"
Wow! Can't wait to test it (whenever a build apparso on graphicall). About more advanced difference predicatore, i remember i posted somewhere a link to a paper named something like "Entropy variance". That could be something worth to look at
Feb 16 2014
Regarding difficulty vs remaining noise, what about the idea shown here? (http://blenderartists.org/forum/showthread.php?236453-Measuring-Noise-in-Cycles-Renders&p=1985872&viewfull=1#post1985872)
Jan 30 2014
Jan 8 2014
Dec 19 2013
Hi, i'd like to resume here a mockup i made a while ago on BA UI enhancements thread.