I've seen the "won't reflect displacement" error b4. Have you tried scaling the scene by 10 or 100? May help.

Interesting....Fresnel should be invisible to the eye (except on extreme close-ups). I suspect for the tutorial I used a high fresnel color value (RGB:114) - which is too much. For the revised RealSkinShader and HyperReal fresnel color defaults to around RBG:51 - so another approach to try.

Mmmm, I'd need to see a render of the "blotchyness" to comment. If the texturemap is "non-human", then I'd say a few tweaks would be needed to get good results. Both the spots and sss use that fact that human skin already has some redness in it to get the results. If you use the script (or nodes) on a "green shimy skinned monster from the deap" - then I wouldn't have thought the results would be quite right! As a startingpoint - turn the spots off - if the blotches are still there - then it is something else.

AntoniaTiger, regarding the spots....That was what I thought too. However when I made the spots larger for the head materials, it looked very wrong....so I changed them back to being the same as for the body. I don't really have an explaination for it....

This is unlikely to work, since python relies on correct indenting (which is lost in these forum threads).

The script that builds the node network actually renames the nodes so that the whole network is readable. Unfortunately you can't rename a node through the P5 material room.

Well the first real reference to Fresnel is in the "Fresnel Effect" section....which reads "Combined_Fresnel is a maths_functions node. The Fresnel_Spots node is a spots node"...so it tells you what nodes to use to get the fresnel effect. To get your Combined_Fresnel node, simply add a maths_functions nodes, and enter the parameters as per the screen shot. I've renamed the node name from "math_functions" to Combined_Fresnel in the screen shoot so you can tell what node are doing what.

MMmmmm, which "fresnel" are you refering to? Give me the context from the tut - and I'll clarify.

tag...

PBM...there was a slight detail due to the python script being oddly modified by certain email programs (some of whom were doing some testing of the update). Totaly bizzare. Anyway, changing a ".4" to a "0.4" in the script has got around that problem. There is a potential issue with the German version of Poser and it's probably not far off finished from there. I hope to get it into Renderosity testing late this week.

DivisionRAIN....quote from the tut...."Combined_Fresnel is a maths_functions node. The Fresnel_Spots node is a spots node"...this is in the Fresnel Effect section. Sorry if it's complicated. In the tut I tried to indicate what sort of nodes were used where - but may not have got them all.

Message edited on: 01/30/2005 16:22

Hi There....My 2 cents....I'm not sure that we are quite at the realism stage yet. There are too many factors to get right in order to fool the eye into believing an image is real. These factors include skin tone, eyes, teeth, shadows (ugh), background, clothing, pose, morph, joint deformation, hair - all of which we can get close - but putting it all together in one image is tough. Maybe the easiest would be a head only shot of a male, mouth closed, eyes shut, black background. There was an image a while back on cgtalk (of Tom Cruise?) that was pretty close.

Mmmm - cool idea. The answer is "maybe"...well "probably", but you would need to get your hands dirty. I am current doing some work on the special maps at the moment, and found they required a bit more work than I thought in order to get them right for a different character. You could probably get 75% of the way there with the UTC (easier to convert to another female character than a male one). Then there would be the issue of changing the script - so it knows which materials to plug where. If you were doing that for M3 it would be a code challenge to cope with this extra bits. For V2, probably pretty simple.

White spec could either be displacement mapping (turn it off a rerender to see if that fixes it). Set your min dispalcement bounds to be whatever your displacement amount is in the shader. Or it could be stray highlights - caused fom a backlight somewhere. Let us know how you go with this.

Eirian - great comparison (better than the one I recently put in my gallery). Firstly, you can see that specular highlight don't work nearly as well in RSS compared with HyperReal. Secondly - it's interesting that her teeth look different in each render (although I'm sure they are the same). Shows how teeth depend on the skin, and how hard (impossible?) it is to get them right with just a single mat pose. Also - on the noise bumpmaps option - since the HyperReal P4 mat poses apply a bumpmap, the RSS script will use these bumps instead of the noise node. As for texturemapping - I've found that it is best to avoid it at all costs - since it causes problems when skin stretches - particularly at the shoulders. So a far shot showing shoulder skin will be tough - it needs filtering for the hair, but that impacts shoulders.