So, I found out that to see G3F face bones, you have to turn invisible nodes on in the Scene tab params (I had no idea there could be invis nodes!). I then set to try out the whole expression thing, using Mjc’s cute Anna custom character morph with obligatory elven ears from the Starter Essentials set (dialed to maybe half their max because gods, that morph is WoW-based or something…).
Y’know, I was hoping the expression bones would be anatomy based (like: muscles), but they aren’t =( In other words, not much more useful for me than just dials. I think in muscles when doing expressions.
This got me thinking whether an anatomy expression rig could be approximated with weightmapped deformers. In theory, yes; but in practice… that’s an idea I will be toying with, but I’m historically slow, so don’t hold your breath.
Then, I was again trying to get realistic pale skin. I took the lightest values from this great article with a notoriously difficult-to-use representation of results, and tweaked them even further because, well, they don’t have real-world ‘type I’ skin subjects in the research. Isn’t that discriminating?! how dare they! LOOOOOOOOOOOL
Close, but not exactly “me”. Ah well. I’ll keep on trying.
The skin is default G3F skin, I just divide it by its base tone in my shader, so that most of it gets replaced by the SSS result. There’s only 10% diffuse overlay – it helps reduce possible RT SSS noises somewhat, and is barely noticeable.
As you can see – it’s a truism but well worth reiterating because I had doubts myself – highlights are indeed hard to notice on skin which is pale enough. Because it’s already rather close to white.
I even had glossy reflection enabled on her skin, which looked so striking in isolated previews (against black), but the final result is very, very subtle.
My shader does some basic energy conservation. If I turned it off, maybe I would get more highlights, but then her skin would noticeably emit energy. Not what I really want.
The metals are excluded from GI (they do not bounce light) because that lead to scary fireflies. I’m thinking of trying this with photonmapped caustics… could be a bit more interesting. Reflective caustics seem to add less time than refractive. But I haven’t done _that_ many caustics to truly tell.
The bracelet I modeled in my Silo+Hexagon combo following Luxxeon’s awesome Blender tutorial, and the specular is too high on it, I think.
The image is rendered to EXR in 3Delight 12, but what I posted here is a simple sRGB JPEG export from i-display. I can’t seem to get the correct glow in Krita’s GMIC version on my 32bit EXR; I can’t figure out how to highpass the luminance level used for the glow. I guess I should check out what Natron has to offer.
I rendered at 12 pixel samples – with all the glossy reflections, even though the skin is limited to 1 bounce and the metal might be as well, it took 36 mins on my laptop (I wonder how many hours that would be in Iray??). As you can see, it’s still a difficult situation to resolve cleanly with DOF – the metal floor is raytraced displacement, to boot.
The dress is Mec4D’s celebration freebie one, and I guess I should try making an optimised UV for small breast size.