In my previous posts, I detailed how lighting can impact the artistic direction of a character. For that setup I used a greyscale model to focus on lighting, today I'm going to show one of, if not the most difficult aspect of computer rendering - skin.
While computer animation has improved drastically over the last 25 years, human skin is a demanding challenge that has yet to be mastered. We are exceptionally good at noticing minute problems with faces. We see faces everyday, and our brains are excellent at facial recognition. Here is our previous greyscale render.
Now lets add our texture. This is our colour information, and was captured along with the model scan. If all we do is add the colour to our model it looks a lot better than greyscale, but looks like bad VFX. The skin looks chalky and flat.
The way we fix this is by using Sub Surface Scattering. SSS is the principle of light penetrating a surface and scattering around below the surface before returning to the surface. Human skin looks the way it does because of the complex internal structure. Skin has three main components, the top layer or Epidermis, the middle layer or Dermis and finally the deep layer or Hypodermics. These layers along with blood vessels, fat, muscles and tendons all produce the final look of skin.
Larger studios will model the skin as closely as possible to get the closest result to human skin, including the creation of veins, tendons and muscles. However that level of complexity is a lot for us to go over now, so we'll stick to the basics and approximate each layer.
We digitally create textures for each of those layers, and combine them together into a SubSurface shader. Combining these results, and adjusting parameters we can produce this result. It is much softer and smoother, and feels a lot more like human skin.