5
$\begingroup$

I'm trying to find a way to apply skin texture detail to my model. I'd like to use normal mapping to represent the surface details, and thought I might use a ready made normal map of a skin surface like

enter image description here

However, in order to get good coverage for my model I think I need seams. But, if I were to simply tile on my skin map I'm pretty sure the seams would become really obvious.

enter image description here

So I'm left with a chicken and egg problem - in order to match up the normal map to the seams I'd need to bake one from a seam-aligned map. But if I had that, then I wouldn't need to bake one!

So

  • If I should abandon the seams, what should I do instead?
  • Or is there a way to make it work this way?
$\endgroup$
2
  • $\begingroup$ You can create another UV map where the seams are hidden and set the normal texture using this UV map. $\endgroup$
    – Denis
    Aug 20, 2015 at 2:22
  • $\begingroup$ But if I could hide the seams, wouldn't that solve the problem anyway? Or maybe do two bakes with the seams in different positions and then blend them together? $\endgroup$
    – jzx
    Aug 20, 2015 at 2:37

1 Answer 1

1
$\begingroup$

I was recently dealing with a similar problem. I found that I didn't really have to worry about it-- these kind of detail, high-res normal maps can stand a bit of noise. Specular rarely aligns closely enough with a seam for it to be visible.

If you try that and find you're having visible problems, you may want to consider masking your normal map with another (hand-made) texture. I'm not sure of your application, but with what I was doing, scaling the R and G components down smoothly eliminates the normal map (leaving only the meaningless depth component of the vector). You may want to consider an unorthodox seaming strategy for this, seaming along the smoothest parts of your model rather than natural lines, since these seam lines are exactly where you'll be masking the normal map. That may mean using an extra set of UV coordinates.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .