I know there are multiple threads out there outlining some of the SSS issues, I wanted to compile my tests and put them all under one thread, I hope this helps.
It has been quite some testing time and I found myself not being able to choose one single method from the current SSS settings in Vray 3.0 which would get me to the desired final results since every option gives pros and cons in such a way that it has been a pain to choose one and get it to look right all the way within the context of the scene (which includes animated characters faces etc.), so far from what i gather the only half baked option left to choose from which somewhat works is the pre-pass mode with scatter gi off but this as mentioned below has some major “look” issues on complex objects lit with GI, making darker shadow areas as well as subtle displacement areas seem incorrectly lit as compared to itself as well as nearby correctly lit non SSS objects, which forces us to use unpleasant workarounds.
It would be great to know if progress is being made on fixing these issues and if so how soon do you guys think we could see a release with at least some of these major issues fixed? a few months-weeks?, any info regarding time would be greatly helpful if possible, reason I ask is because we will be migrating from another render engine to Vray very soon for our work, however such current issues with SSS are kinda scary in production for us :).
Thank you for your constant support and effort!
All the following tests where done using Skin SSS and latest Vray 3 Demo version running on 3ds max 2014:
The first and second problem (darkening on non-attached objects and brightening on attached ones) are actually easily fixable and will be done for the first major service pack.
The third issue with the cutoff seems like a bug (unless it’s reflection cutoff). Can you post the scene here, even though it is simple?
On the whole, my goal is to fix the raytraced mode; the other modes are mostly for compatibility reasons.
The third issue is not related to cutoff parameter - it is still there even if all cutoffs in the scene are set to 0.
The case has already been reported under ID: 0010955
It’s because there is another object inside the SSS material that blocks the SSS rays. Technically, in a physically accurate world, this would indeed affect the final object look. But if that’s not desirable, it’s not too difficult to ignore objects inside the SSS. Then again, there have been a number of requests for the opposite… maybe we need an option for that.
I wondered the same thing but unfortunately from my tests the problem with intersecting object and sss with raytraced sampling is still there in build 3.00.08.
In various work this has been a problem for us and we had to go back to the prepass illumination map one, i hope it’ll get fixed soon.
I just contacted our developers to get more information about the state of this issue.
But it seems that the fix is not added to the latest builds, so most probably they are still working on it.
We will inform you as soon as any update is available.
From what I understand from your reply that this issue is still pending and may not yet be addressed in the coming sp1? I have had the opportunity to check out the latest nightly stable build while there are good enhancements over all, unfortunately the issues of SSS are still present, if this statement is correct, is it possible to give us a quick workaround “fix” for this by the time you guys get the time to fully address it? the workaround if possible could be the following:
An exclude option in the SSS material itself to completely ignore any other object that comes into contact with, this way for instance lets say you have water droplets scattered over an SSS object, if this is rendered as is now it would produce dark edges which would get refracted and produce a muddy droplet feel instead of water, and because they are so small they would render almost entirely black, so we have to resort to many compositing tricks to render in passes etc.. so by having an exclude option we could simply tell the sss material to ignore the droplets therefore avoiding any dark edges per contact.
Is this possible as a quick update to help us work around this issue by the time the official fix is released?
I will be able to work on this after the service pack is released - it’s too late now for the official build. I will let you know as soon as it is in the nightlies.
here is an issue, latest demo in 3dsmax 2014, SSS2, basic setup, one light, prepass mode (raytraced mode useless for character creation, it produces even more errors already described on this forum).
I have one material applied on two objects, looks fine except for the strong bluish/green tint which is typical for SSS2 but lacking on the real skin of course:
Here I have the material copied in next slot of the material editor and applied to sphere, hence there are
now two identical materials applied on two objects, you can notice total error where the cylinder penetrates sphere. Where is a shadow should be SSS, and where is strong SSS should be a shadow:
Here again one material applied on both objects, this time cylinder does not penetrate sphere, is only very close:
Here again two identical materials applied on two objects, again the error.
So the main issue (except for the blueish/green tint) is that is impossible to get two SSS materials close together. That´s quite bad for character creation, eyes are either glowing or covered in shadows and eyelids too. Using of HDR as source of light better covers this issue as use of “flashlight” front lightning, but is still there. The only workaround is to use one single material for all objects - very problematic solution. The issue is always there - raytraced mode, VraySkin, multi/sub-object material, IDs, …)
You can control this SSS2 behavior by the “Prepass ID” option.
By default “Prepass ID” is 0 and the material will compute its own local illumination map, if you set both SSS2 materials with “1” then you will get the requested render result.
Please let me know if this helps in your scene.