About SSS in blender:
I started without the use of nodes.
I had to use extreme parameters for this, else I could loose all details and be in front of a piece of wax. Even so, waxyness is still visible and I still need a much more translucent effect.
So I tried nodes method.
At first, I need for example two identical materials, the one with sss, the other without. I can combine then via mixer. But how can I do this? When I select sss on the first material, the second does the same automatically, these two are linked. I did it once but never again. In blender 2.5.3, please help. I really can’t understand how blender nodes work with materials. And no tutorial in blenderwiki. I can’t understand much there.
It’s really hard to help you,as I don’t understand really what you are saying, but,probably,your problem is that you don’t use 2 materials,but 2 instances of the same material…so:
Go to the node editor(for material),and check the use material node.
After create 2 new materials(in the node view) and look carefully how many instances are used for the materials(look at the name)
If every materials have a different name,you can have separate control for everything,if the materials have the same name you have more instances of the same material,and you have to check the +(near the name).
After this operation,you should have the name changed,and you can actually change parameters.
So you can combine more material for doing a decent skin shader.
If it’s not this problem please try to give more information,because I don’t understand.
OK thanks I already found it. It was hard for me to understand. Exactly as your advice.
So a new test here, using nodes, not exactly what I have in mind but no waxiness, I think.
Thanks Frank_robernson, this was really helpful.
But this wasn’t working in this scene, I had to use maps in all nodes, I didn’t use the epidermal node with sss. So three nodes like this. I still don’t like what I see. I think that GI is the solution but really a mess combined with SSS.
Thanks womball, I’ll do it, but its PP. I already tried this, lets see what a single pass can do first. All these look complicated in a way they make me think that something is wrong. LOL Maybe textures and lighting set up.
I tried sss for aged (dirty) marble. Shouldn’t be more. A three passes combination. -nodes-
I also tested shrinkwrap to project high tri mesh from voxels (3DC) to low retopo (6k) mesh. Much better than displacement_maps - see in hair areas and a multi-res model for further sculpting. Some wires too, I wish we could have an auto-retopo tool in blender some day.
That look pretty good. Try placing lamps further from mesh, to avoid burned areas (or change lamp falloff to linear).
About shrink wrap- you add it in multires, to render mesh with hires detail? Isn’t is slow? I mean you probably have like 2mil polys with multires enabled for render.
Cool sculpture.
Thanks JoseConseco,
Yes I burned it, I know but I liked it. Normally I hate burning. About multires, I didn’t go so high, ~ 400k, the same could happen using displacement am I right? Its rendered pretty fast 4 passes in less than 2 min. On a dual xeon 16 threads machine of course. The four times subdivision is the default of many renderers like mentalray. The trick is to bake normal maps from this point, if two subdivisions in mind then I have to bake normals for this level.
The faun is part of a bigger composition, cooking right now lol
I think he imported in blender the high res mesh. Then he created a low res mesh and wrapped this low mesh using shrink wrap on the high mesh. So this is his way of retopologizing very quickly that is the tool he lacks in blender.
@Bao2, exactly so, here I already constructed a low retopo mesh in 3dc too. But this is the point. Blender is ready 100% for excellent re-topo and we don’t have it. Some times I even constructed a simple box modeling and worked in most cases (not very complex though, it couldn’t work in this case)
But I did try this. A simple lattice like box modeling, subdivided at 10k and after shrink wrapping its almost here. I could spend some more minutes for better results. The snapping method of re-topo tool is completely different, its not working like this. I don’t like it anyway. We are so close, someone could build something decent here.
The same technic is very common among zbrush users. (cage or project all)
I tested the excellent zbrush and compared it with blender shrinkwrap. These tools work exactly the same. They even have this issue (you have to project one by one for every sub level) for better results.
But these results are much much better than displacements. And take exactly the same time to render.
I usually bake two normal maps. One for low def and another for 3-4th sub level. You may have it as a background object in the end.
I find rebuilding that skin shader a little tricky to get the same results.
My results are way to translucent - which I think can be an issue
with the scale.
Change units to Metric (scene panel)
Now select the head and in the material nodes select the Back Scatter material and look
the SSS values in material panel for RGB Radius: 2 meters, 1 meter and 50cm (that is the length the light penetrates in the mesh and radiates illuminating it).
Now look at the SSS Scale you are using: 0.500
Scale when you model in real world measures must be 0.001 and your mesh is in real world measures (about 25 cm height).
Try also to change the Error parameter in SSS from 0.05 to 0.5 and you will have a render near three times quicker if you change it in the three materials with SSS you have, it will lower the quality in some areas so it only is good for test renders not final image.
I think the SSS trick is almost there cekuhnen. But where is the texture? This makes the difference. And of course it depends on scaling and of course a pass without sss is needed. I use one as you can see in nodes setup. You may trim the whole effect from this node. And scaling of course as bao2 mentioned.
I think I must confess that the scale still confuses me.
for different layers it has a different values.
I gues sthe main problem is that I do not even know what value
for size is needed for which skin layer.
Epidermal 0.05 (causes also most of the wax like effect setting it to 0.005 lowers the translucent effect a lot)
Back Scatter 0.001
Dermal 0.018
michalis
in my file you will see that here is one unscattered pass as well!
