About noise in general

[pengo]

Thanks for the reply but if u permit me i have to tell the opposite about the global subds multiplier effect onto irrad hsph subds, and is Flipside to confirm in in his classic tutorial where specifically wrote: "Global subds multiplier affects irrad map too" in steps 6 at this link:
http://www.spot3d.com/vray/help/150S...s_interior.htm
Here the confusion sometimes someone tells one thing and after someone tells the opposite.But don't misunderstood I appreciate the suggest u gave i tested a lot a new way to use vray even if I still not find it so fast in rendering maybe in setting yes.
Thanks again

[pixela]

No, by no means. I would say most people haven't actually noticed any negative changes since 2.0. It's just a handful of people. But then again, it doesn't necessarily mean there is no problem, but perhaps that there is something in common in the workflow I, Pixela and others are using that leads to these problems.

I agree with you.

It is hard for me to point the problem too. Because I cant replicate it with small simple scenes.
One of the things I can say is that it happens with bright scenes full of GI.
Another thing is that "universal settings for vray" method isn't working for me as it used to work in Vray 1.5
So strangely even I crank up the maximum parameter (with a low threshold that should bring good detail), I can't have enough detail and/or decent AA.
I am aware of the compromise between detail and noise, all the grains and how DMC works. But what I am trying to say is that the method which I was able to get good results with 1.5 is not working in my scenes anymore with 2.

As far as I understand this is similar with you Bertrand - no?

[pengo]

Hi Pixela, if is possible I'd like to establish if there is some common things in the scenes we get problems, because I think is a fact that some users use to avoid some realistic GI giving importance to rendertimes (trying reducing them as possible) some other point to realistics effects to learn as well how to treat them and to optimize having a lot of those problems. I try to explain as well what I mean (with my english s not so simple), for example some users avoid curtains shadows or refractions from them and some instead want they refract to have perfect behaviour from this element in scene, and so with glass at windows and so on. Who avoid real behaviour have faster renders and cleanest but exist cases where anyway artifacts are hard to solve without technical support. So I suppose a scene for example to point the artifacts we are trying to avoid would be needed. Because I know from my experience and i read similar doubts from others (fortunately) that when a problem is hard to solve in GI or on some materials the first idea is that a user error in usage or some plugin that could cause problem (like sometimes was even verifyed color correct in some cases) or modeling error or a complanar planes or solids or so on and on complex scenes is really really hard to go to each point to discover what's new we used in parameters to get artifacts we never got.
Hope to have been clear. Trying to avoid micro-noise in bright areas a scene I'm working onto gives strange results sometime (in one frame gone , the other re-appears or the dr bucket from other pc loading the same irradmap and lc is clean and the original pc made starts the DR render proces gives in the same area buket noised)

[joconnell]

A lot of people have bought bertrand's chicago loft aopartment scene - would that be a good test case? It's got all of the problem areas, highly glossy materials, high quality detailed GI, depth of field and a huge exposure range with extremely bright hdri values for the exterior light with quite dark areas in the scene too. Sounds like a good hell test and also he mentioned he wasn't happy with the render times he was getting on it?

[pengo]

Would be a good test scene but expensive to test in therms of rendertimes, I'd say maybe a bedroom or a room with only one windows to give light with a domelight as most of the user explained here with an exr in it like Guthrie's map, would point better the problems because less lighted (with less light sources or openings) ambient gives more problems to GI and maybe even more people could give proper support without need to have a renderfarm to do some tests. But this IMHO.

[joconnell]

Yep - if we setup a scene kind of like the old vray benchmark scene and then try a few different approaches to getting the quickest, cleanest, most detailed render possible in a reasonable time.

[greysheep5]

hi Bertrand,

what i meant is i use the same high subdivision all time, so nothing to tweak and very fine for premade libs, we use all presets too with that settings.
what is an alternative maybe is to use the global multiplier but, i havent tested on this in detail yet (from pure logic should be the same).

the only value i adapt is threshold in AA, sometimes to 0.003, (but not changing the dmc threshold, it is not the same as the AA threshold and i dont need it has low as AA thr. in 2.0 core, DMC thrs. i keep usually at 0.01 or lower not more than 0.005), in my works it renders faster noise free in that setup than with low subdivisions. it seem then to get easier noise free in shorter time.
as some already stated this system seems not to suit all work flows, for me it is definitely faster, for others not though, they seem to have better results keeping it at 8.
so only trying which is better for yourself might bring something.

cheers
stefan


of course it might be not for all type of work flows, so i can say only from my working.

cheers
stefan

hi stefan,

thank´s for clarifying that. your answered my question too with this reply. as we were/are not sure what exactly this clr trshold thing in the AA adaptive dmc settings is we were always keeping it linked with the "use DMC sampler thresh." checkbox to the general DMC sampler settings and only adjusting the image quality with this value (from 0.01 down to 0.005 just like brett)... so maybe it´s worth a try to unlink those settings, keep the DMC noise threshold at 0.01 and play with the clr threshold value to optimize quality vs speed.

just like you said, one can only find out trying for yourself, for our workflow it works better with keeping the subdivs of the lights and materials at 8, that´s fast for us and noise free enough... but of course, when you have prebuilt libs and objects you are using all the time with the subdiv settings at 128 or 256 it´s just the same as using the default value and also a very universal approach. i guess it really depends on the kind of work you have to do and on the light setup (for example if you use hdr´s and how you light interior spaces, i mean if you use fill lights or the purely realistic approach and so on...).

thanks for the insight from your perspective stefan! how is vrayforc4d going? we would have definitely tried it if we hadn´t bought our max and vray licences before it came up... sounds kinda cool to have the support in the same city and working in the same industry.


cheers,

christoph.

[PIXELBOX_SRO]

i actually think it has to be a bigger scene where light has to bounce a lot.....you wont get these troublesome situations in small interiors (at least i dont).

When it comes to using Bertrands scene for testing...i am up for it!!!

Few things i noticed in the setup that are sort of different:
* in LUT pref, BB is using 2.2 in the outputs checkbox instead of 1.0...in other words, the LUT setup is not the same as the standard one
* instead of using sun, BB is using standard direct light in this particular scene + an HDRI map for the environment skydome
* no area lights in the window openings

when it comes to my workflow...i find it generaly hard to get rid of the noise when i dont place the plane lights into the windows
sadly i did not have much time to dig in the scene deeply yet but those things just were standing out cos they differ quite a lot with my workflow

[joconnell]

Okay well perhaps if it's not too much to ask then would yourself, pixela, bertrand or Peter guthrie put together a simple scene with a typical glass, floor, ceiling and wall material. It could be filled out with geometric primitives and simple stairs objects to give the GI something to work around - the intention is not to have a scene that looks good as that'd take a long time. It's more a case of having all of the ingredients that'd typically appear in one of their renders and cause problems.

[pengo]

I agree with Joconnell, even if those artists have not all this free time to spend for tests I hope some of them will find it

[pengo]

PIXELBOX_SRO not all have that scene and me too
But anyway to follow your workflow or test as u'll do i'd like to understand as well:
-Does the hdr/exr map used in vray environment slot or in max environment or as BBB3viz uses often in vraydomelight?
-Does he changed the gamma value of the hdr/exr map used with gamma option in the vrayhdrimap gamma or he used that map into color correction map? (because some plugin like color correct gaves artifacts)?
-the vraylights subds are at wich value?
I mean if u are interested to share tests would be even preferable to indicate all parameters used
Thanks for the attention

Stefano

[PIXELBOX_SRO]

guys i am not the creator of the scene and it would be quite time comsuming probably to create scene that is problematic
as much as i would like to share some other thoughts about BBs scene i cant, it would not be fair as a lot of things in that particular scene are part of BBs knowhow and special workflows.
I only pointed out things that differ to typical scene i would create myself.

But yes, i agree it would be really great to have a scene we could all work with.....

It would be great, on the basis of the new function is vray to create a NEW PROBLEMATIC BENCHMARK SCENE ))

[lllab]

here a small test i made a test with Bertrands Chicago scene,
19min with bruteforce-LC using my method at 1600px, 1h 38h at 5000px wide.
of course it is not fully textured and has no furniture elements,
for the test i put a light gray glossy reflecting material (ior 1.6, intensity 0.5) on all objects, and hdri outside.

brute force, glossy reflection and one hdri dome light, all a subdiv of 256,
AA 1x100x 0.004, DMC all default(threshold 0.01)
colormapping: subpixelmapping on (on makes it dramatically faster for me even though not 100% correct in light then), reinhard 1/0.25/2.2

http://images-stefan.s3.amazonaws.co...BRLC_19min.jpg

i am not sure this says some for fully textured and crowded scenes, but it seems quite ok from noise to me,
AA thrh. at 0.003 would make it more clean of course, render slightly longer

Stefan

[pengo]

lllab I agree about subpixel mapping using it switched on is a lot faster and prevent that micro-noise in brightest area that was evidenciate as vray 2.00 came out, but the warning message in vray window during render always keep me far from use it.
About the vraymaphdri is used with its gamma lowered? Like 0,8 because the sun visible to came so sharp would need lower gamma value.
Did u used a blurred version for light in dome and the same map not blurred for reflections c4d environment slot?
The choice of a glossy material is perfect to give the right complexity the scene would have with all its proper materials applyed how many bounce u leaved? default 5 or lower like 2?
What pc did u used? cpu-ram-hdd/ssd?
seems enough for now
Thanks in advance and thanks for sharing knowledge

I forgot are present glass at windows? If yes are those casting shadows? Are those excluded by GI? Are those excluded from light casting shadows too?
It specifically gaves problem to BBB3viz as he wrotes in previous page darker render and more noisy GI.

[bardo]

lllab..it's a good result as noise level per rendertime for me.
Could you share also your hardware specs?? and if DR, how much pcs and their hardware specs?


Thanks in advance.