Hey guys. I've been trying to figure out an issue I am having for a while and haven't come to a solution yet, so I'm turning to the lovely Polycount minds for some insight.
I'll just keep the example simple for convenience sake, but essentially I have a shader that is vertex blending with an edge mask to give a nice procedural like edge. I've done this several different ways before, but I've never actually tried to create what is basically an 'offset' or fake normal for the information that is being created by the blend. I'm not even completely sure if it is possible.
I'm wondering if there is some kind of tech voodoo I can possibly use in conjunction with the expression that is being used to define the alpha of the LERP to essentially create an offset/normal edge distinction between the 2 textures. For example in the attached image, being able to create the illusion in the normal information that the 'brownish' area is sunken into the lighter 'tan' area.
An example of use would be having a chunk of drywall with destruction blended away, and making it appear as though the core behind the wall was recessed.
I can fake an AO shadow around the blended area just fine, I'm just trying to create some kind of lighting illusion to make it appear to not be completely flat.
Thanks!

Replies
Here's a package with the material in it. I've just tied it up to a sine from time so you can see it sweeping over the edge map. If you'd like harder edges use ifs instead of blending the normal map.
I'll upload some images in a moment, Imgur are the latest casualty of the "lol let's break our perfectly good website" revolution
Here's how it looks with harder normals: (two different types)
I'll play around with this later after work, but James that is pretty badass. Exactly what I was talking about judging from the screenshots