So, I think I basically figure this out: Again, follow the same strategy outlined above, I created a simple edge detection node and use it to blend a few common values, like base color, normal, height, AO etc. The input map is simply a ID map baked in SP, but specially without dilation or diffusion, so basically a map of…