The second part here is a pretty simple technique and could probably be summed up as: using a normal map from max to find out where the mesh overlaps itself. If you know exactly what I'm talking about with that statement, you can probably skip this tutorial, but I thought I'd include it for completeness.
