Strange buckets in VrayMatteShadow Pass

[caypiranha]

Hi Guys,

I am having here strange Buckets in this rendering. Max 2019 / Vray 3.7. Testscene, wich shows still the same Problems is attached.

I would be thankful, if you find the problem. If not you, who else? =)

The containing scene shows it more obvious.


best regards,
Kay

[hermit.crab]

I seem to be unable to reach the link, could you please re-upload the scene?

[caypiranha]

VrayMatteShadowPassProblem.rar

[^Lele^]

For the unevenness in noise level, ensure you raise the min AA subdivs a bit (to, say, 10).
Because those REs aren't taken into account for antialiasing, the minimum AA will ensure they stay noise free.
Further, feel free to reduce the Dirt map subdivs to 3, as 8 is unnecessary (in general, post V-Ray 3.3) to get clean results (you'll notice that RE *is* considered for AA).

The oddities in the shadow pass are still being investigated.

[caypiranha]

Thank you all for the reply. Well and I uploaded it once again and it should work now.
Any ideas for the Bucketproblem?

here I met today another one - totally different scene - same Prob

[^Lele^]

Ok, wait, i think i misunderstood what the problem is then (we did get your scene, ofc. it's what we looked at and worked on.).

In the first post, you show something with green in some buckets, i thought that that was the issue (which we couldn't replicate here.).

If it was not, but rather the problem was the uneven noise level, then the solution is what i posted above: raise min AA to something around 10 (more or less will depend on your target noise threshold), and that will sample all buckets in good enough a way so as to not leave visible differences.

Or upload another scene if the issue can't be cured as suggested.

[caypiranha]

=) .. nono - the Problem is the BUCKETS (see my last Post wich differs in luminaceinterpretation) - or more precise: these are "bucketgroups", since the renderbucket even was smaller. We reproduced this error here on 2 machines in our office and from a renderfarm I got the same results.
its all here not about noise or color (the first pictures with the cylinderlikething isn´t gammacorrected, thats why it looks so harsh) see the Pictures at1:1 size and look around: bucket almost noisefree are there and if you render my sphere in the given scene - it´s right obvioswhats going wrong.

[^Lele^]

Oh ok, so the way to fix it is raising min AA subdivs, for the reasons explained above.

[PIXELBOX_SRO]

i cant resist and have to stick my nose into the debate because my head cannon comprehend this new info, Lele. Woudl you perhaps mind to elaborate on the following please?

would not one expect that raising min AA values result in drastic impact on the render times?
also i woudl have thought that "a bit" would mean to 2 or 3 and not 10
is it not the same like using fixed at 10 back in the old days?
May i also ask why RE are not taken in account when doing AA?
Or did i miss something and you guys are speaking about somethign special???

Thanks a lot!
M

[^Lele^]

2 or 3 is fine for some cases: tiny, or moving geo may need 9 rays to be found inside a pixel.
10 subdivs will converge the shadow/matte shadow passes when there isn't enough info in the RGB to go on by (the scene above has non-gammaed pixel values of ~0.01f. An auto-exposure would boost that image exposure by at least a factor of ten, and then adaptive routines would work in the shadow pass too.).
REs which make sense (and are quick enough) to be taken into account for AA have their corresponding checkbox.
If V-Ray took every RE into account for AA the render times would explode: the behaviour is identical to competitors, by the way.
A min AA of 10 is indeed a fixed sampling "padding".
Which is how data *passes* have historically been rendered: fixed AA for render elements where contrast is not nearly enough to have an adaptive algorithm do its job.
Nothing has changed in this as far as i recall.
Alas, mattes and masks were *always* relegated to special setups, *never* calculated with the beauty, on the jobs i witnessed and directed technically.
Speed and iterability were always worth the extra prepping effort (which, by the way, could be fairly easily scripted.).

[^Lele^]

caypiranha i forgot to mention: you can (and really, really should!) set the subdivs multiplier for the plane (and all other geo) to 1, down from 12.
Or that render will never finish.
If you can't achieve the result with adaptive AA, look towards fixed sampling, and try your hardest to find a cure to the issues from within the global render rollout.
I say this as debugging otherwise becomes hell(tm.), for anyone but the scene author, and likely only if the scene was set up recently.

[caypiranha]

Hi Guys. It´s all about the the "Buckets" in the shadow pass - thats the point. I don´t mind here in this thread about AA in RE and Multiplier thing... sorry. So far rendeing is fast and fine. I also use Vray since 10 Years or so. (wich doesnt mean, that I am right with all I do!)

and HERE, It´s all about the VRAY MATTE SHADOW wich delivers here this strange luminacediffering "bucket" - thats the bug here! =)

Aleksandar directly tried to pick up the scene after publishing this threead.
Did you tried it again, any Ideas what up with this thing? Any Progress?

thanks in advance to all!

Kay

[^Lele^]

Hi Guys. It´s all about the the "Buckets" in the shadow pass - thats the point. I don´t mind here in this thread about AA in RE and Multiplier thing... sorry.

Fine then.

thats the bug here! =)

There is no bug.
But you do not care why it behaves like it does, so i cannot help.

Set min AA to 10 to fix your issue with the scene.

[hermit.crab]

What Lele said. We collaborated for the testing of the scene. Try the provided solution and report with feedback.

[caypiranha]

sorry Lele, no right time in the moment - no (needed) care...

Well, I tried it & indeed, it will get away. Strangely when I set the Multiplier @ Objectproperties to 1 again, leave everything else as it was, the Shadow now behaves very different.(*) If I set the minSubdiv in the bucket image sampler as you said to 10 - well then it is all fine. But rendertime goes all over the scene and now it takes a while. Problem so said solved.

now: if I open myuploaded Scene and and do nothing else than set the number of probialisticlights from 8 to 12 - it´s gone again.

If I use a polyplane instead of the vrayplane - the world is ways slower on one hand but on the other it looks in reference of the light´s direction way more correct but totally different. (and is rendering absolutlysmooth with AA to 1 , ObjectMultiplier also set to 1 and everything leave as it is.

In case of performance for visual motives like Product shots e.g. I find now a vrayplane lokal subdiv to - somewhat to ever what ends smoothlike and the adaptive lights to 12 for avoiding these Buckets (maybe this "bug" is a technical impossibility, that Chaosgroup might explain somehow?)

(*) it´s possible that I am having trouble while previewing my renderings in the FB when it comes to see a right gamma. Once rendered, stored and recalled, thy show now also a "sharper" shadow. So the light grey aura, the appeard - better: was shown during the rendering will be dimmed away... well then it´s not easy... =)

the first picture shows AA and objectsamplerset to 1 ( dont know, why these artifacts are now there apear)
the second Picture is the diffenence between VrayPlane and Polyplane