Raytrace SSS strange refractive behaviour?

[Vincit]

Hello again,

Raytrace mode seems to be behaving similar to the "raytraced refractive" in the prepass mode, when we want it to behave as a "solid", the image below ( a sphere with a cylinder inside) shows the solid render done in prepass mode which is how I want the raytrace to behave, the raytrace mode is refracting the cylinder inside the sphere.

This makes any kind of skin setup in raytrace mode very problematic, is this a common issue or am missing a step?

For now we will be sticking to prepass mode for our tests since raytrace seems to have other issues as well mentioned in previous threads.

[Svetlozar Draganov]

What option for Singe Scatter did you use when rendering with Prepass-based illumination map ?
I can only get the cylinder invisible if Singe Scatter is set to Simple/None in Prepass-based mode or None in Raytraced mode.

[Vincit]

What option for Singe Scatter did you use when rendering with Prepass-based illumination map ?
I can only get the cylinder invisible if Singe Scatter is set to Simple/None in Prepass-based mode or None in Raytraced mode.

I am using the skin shader - In the skin shader I have only one option to choose from, which is either raytraced or prepass (no options for the simple/none/etc.), prepass works as expected while raytrace shows the cylinder.

I am assuming you are referring to the fastsss2 shader which I am not using but if so I did test it and here are the results:

Fastsss2:

Prepass + none = no cylinder (correct)
Prepass + simple = no cylinder (correct)
Prepass + Raytraced(solid) = Cylinder visible (correct as per the description of the option in the manual)
Prepass + Raytraced(refract) = cylinder visible (correct as per the description of the option in the manual)

Raytrace + None = Cyliner Visible
Raytrace + simple = Cylinder visible
Raytrace + Raytraced Solid = Cylinder visible
Raytrace + Raytrace refractive = Cylinder Visible

Raytrace is still showing the cylinder in all modes?

Here is the test file link with fast sss2 applied if you wish to check, also try applying skinsss raytrace on the sphere to see what i mean: https://db.tt/IJ9BkPtD

Thanks,

[Svetlozar Draganov]

Thanks for the scene - I managed to reproduce it now.
We already have a similar issue into our bug-tracking system but since it is not absolutely the same I'll add this one too.

We will update you when we have news about it.

[Vincit]

Thank you for the heads up, for the moment we switched to fast sss2 shader with prepass mode, it is so far giving the best and most stable/predictable results.

After numerous tests on the new skin sss we thought some of the underlying layers it is using such as the "deep" scatter is not producing the predictable results expected, more tests would be required but there is something a little off with the over all setup of the shader we can't really put our finger on at the moment (having used arnold and mental ray shaders extensively in production both share the same philosophy of multilayered skin, While arnold also had minor issues with the deep scatter it was not this troublesome) also the fact that more control over the "phase function" is unlocked in fast sss2 while it is "hidden" in skinsss may not be helping perhaps in other cases, I hope we could pinpoint these areas better in the future and I will try to provide exact places which may still be causing trouble, for now the fast sss2 is producing best results of the two, however I managed to use the new skin sss reflection parameters which have seen much needed improvement overlaying it with a blend over the fastsss2 as a base,

Lastly I was forced to use "Shellac mode" in the blend material for reflection passes to visually "look" correct vs not having this option turned on, however turning this on Vray warns of breaking the energy preservation mode of the shader, while understandable that it is acting as additive, I am hoping there is a way to get the same result without having to force it to break the energy conservation, now in our case we are controlling it carefully through the parameters to preserve energy conservation in the shader but isn't the point of a coating blend shader to have it work correctly with all elements including reflections without having to resort to "cheating it"? or is there a step we are skipping in vray?

Thanks again and good luck.

[Svetlozar Draganov]

It would be a good idea to show us some real examples about what you are talking about in order to understand you more clearly.
For example what do you consider as predictable result and what did you get from Vray?

Every render engine uses different code so applying the same material setup between different engines will give of course different result. The result is usually unexpected because the users are used to working with another software. In any case a visual representation of the result you got with included scenes and textures would be very very helpful for us. Furthermore a description of expected behavior is also needed along with expected results and what you got from Vray.

Phase Function is indeed not included into Skin material although I am not sure if this is due to shader-simplification or some sort of limitation.

According to the Shellac Mode - this option is designed mainly because a lot of people are more interested in the final look of the shader rather than if it is physically accurate or not. Sometimes with additive mode we could get some very nice results without too much efforts and some effects are even impossible when the low of energy preservation is applied, so getting the same (or very close) result when shellac mode is disabled could be possible but not aways. Feel free to send us sample scene about this questions too.

[Vincit]

Yes Of course, samples would be a best way to explain issues, I certainly agree regarding your points.

Regarding additive mode it is certainly the final look we care about as you mentioned, once again i mentioned below i will do more tests and look into how the energy is being preserved, however when a flag comes up i do post it to get your point of view, one of them was related to this additive mode, our worries were questions such as what would happen if we have multiple layers of reflection additives in shellac mode each adding on top of the other (like how we used to do it traditionally) while it may work for one tweaked scene it may need to be tweaked again under slightly different light situations and as you can imagine on large scenes such things could become troublesome to re-adjust and tweak accordingly, in any case we will go through more tests and will try to keep you guys posted as well in the future.

Thanks for your input and feedback once again, have a good day.

[Svetlozar Draganov]

The issue with the visible inside geometry has been fixed in the latest 3.30.03 version.