In this tutorial, I go in depth into the process I use for breaking subsurface scattering into renderable layers for compositing. The reason for this has a lot to do with my general CG workflow. The fewer times I have to render my final output, the better; rendering takes time and time is money. So, if I can render my output once, with enough control to be able to fix whatever small tweaks need to be made in compositing, all the better. As you'll see in the example images I provide at the end of the tutorial, you have a great range of control on the final look of your output based on your rendered SSS layers with a few simple compositing tricks. But remember, SSS is based on your light positioning, so, if it's something as drastic as wanting your light source to change direction, that will have to be re-rendered. I'll go over some tricks for changing things like a light source direction in comp during a later tutorials on normals passes and how they're useful in compositing.

For now, the tutorial is available in PDF version and over at CGTalk here. Very shortly, I will write up an upated and full walkthrough here on my own site as well as update the PDF source too.