The Volume Fog sample shows the per-pixel density volumetric rendering

technique. The fog volume is modeled as a polygonal mesh, and the

density of the fog at every pixel is computed by subtracting the front

side of the fog volume from the back side. The fog is mixed with the

scene by accumulating an in/out test at every pixel -- that is, back-facing

fog polygons will add, while front-facing ones will subtract. If the value

is non zero, then the scene intersects the fog and the scene s depth value

is used. In order to get better results, this demo uses 12 bits of precision

by encoding high and low bits in different color channels.