Don't bother mentioning the Edge_Blend with Blinn - just mention Blinn. I saw face_off do that, but I never really understood why. I think it was a roundabout way of conserving energy, which we know is accomplished much better and more directly by subtracting the output of the specular node from the diffuse value. Although that still isn't correct, but its reasonable.

Oh and the Fresnel equation we were using is only for a microscopically smooth surface, such as water or glass. For micro bump, as found on skin and just about everything else, the amount of reflection does not follow that idealized curve. The Blinn algorithm is a direct attempt to capture the idea of the Fresnel effect for a micro-bumped surface. That is what the eccentricity parameter is for. The Specular node does not model this, and so it does not produce the nice "rim light" effect. It can't, even if you amplify it via the Fresnel equation. It just doesn't capture what happens at the edges of a micro-bumped surface.

So my first suggestion would be to dump the Fresnel nodes and just use a Blinn. Not only because its less nodes, but also its more accurate for the specular effect on human skin (and a lot of other things, too.)

---

Renderosity forum reply notifications are wonky. If I read a follow-up in a thread, but I don't myself reply, then notifications no longer happen AT ALL on that thread. So if I seem to be ignoring a question, that's why. (Updated September 23, 2019)