While I waited for my computer to render out a test screen, I picked a random face shot off Google and did a little cut-and-warp to turn it into a face texture, which I then separated into the minimum detail frequencies needed to de-construct and re-construct the face. This meant a high-pass run to get the fine skin texture, a smart-blur run to get a painted-smooth hue and saturation layer, and black-and-white filter to get an optimal gray-scale layer. As I re-constructed the final texture, I added in a little noise under a hue and saturation adjustment layer, which I overlaid--together with the high-pass layer (I forgot to duplicate my gray-scale, which got eaten by the high-pass filter, but I did not need it in the end)--back onto the painted layer using a soft light blending mode. The final is a little sharper and brighter, thanks to the noise, and I did not mask out the eyebrows in this process, so they look painted, still; but on the whole, a good exercise.