Can't solve baking issue

Nomadvisas

Hi,

I've recently done some tests baking a simple cube and made sure I fully understood the baking process to achieve a good clean bake. However, when I moved on to a test model of my own, no matter what I do, I cannot get rid of the edge artifacts on this bake.



I've UV mapped the mesh to have splits on any hard edges, exported using the recommended settings (excluding the triagnulate as that gives the smoothing groups issue). Tried baking in xNormal with a cage and substance using the Distance sliders.

Can someone please lend an additonal eye and some thoughts. I can't seem to see where I'm going wrong as the previous tests I did using the same method came out fine.

Files attached

https://1drv.ms/u/s!AotVsV5EHQaM3lVlHNtL18OR3p4n

Thanks

Zablorg

I took a brief look at your objects. If I understand correctly, I think the hard edges are your problem- it means the ray traces are shooting out perpendicular to the surface at each face, and thus they miss the high-poly corners entirely.

This thread should give you a good sense of the relevant concepts involved.

Nomadvisas

Zablorg said:

I took a brief look at your objects. If I understand correctly, I think the hard edges are your problem- it means the ray traces are shooting out perpendicular to the surface at each face, and thus they miss the high-poly corners entirely.

This thread should give you a good sense of the relevant concepts involved.

Hi, thanks for the reply.

I don't believe that should be the issue as I have followed the practice of making UV splits on any hard edges and giving them sufficient padding. My test bakes using simple cubes worked fine and this mesh is following the same setup. I've read through the wiki and can't seem to find a solution to the problem either

Zablorg

Nomadvisas said:

Zablorg said:

I took a brief look at your objects. If I understand correctly, I think the hard edges are your problem- it means the ray traces are shooting out perpendicular to the surface at each face, and thus they miss the high-poly corners entirely.

This thread should give you a good sense of the relevant concepts involved.

Hi, thanks for the reply.

I don't believe that should be the issue as I have followed the practice of making UV splits on any hard edges and giving them sufficient padding. My test bakes using simple cubes worked fine and this mesh is following the same setup. I've read through the wiki and can't seem to find a solution to the problem either

From your mention of UV splits and padding, I believe you think I'm referring to a problem with how to the texture is actually baked into the UV space, after capturing what are ultimately correct normals. Rather, I'm referring to the issue how how the rays are cast to determine what is to be baked.

Because it seems you have hard edges (correct me if I'm wrong, here), I suspect the bake is acting like the example on the left. Note in this example that because the normal direction is discontinuous at each vertex (i.e. a hard shading), entire sections of the high-res mesh are being missed by the rendering zones. I think a very similar thing is happening with your right angles- they're missing the corners of the high-poly, and that's what's causing the discontinuity in your bake.

If I'm incorrect in this, can you please describe how you've addressed this possibility? Did your cube test also have hard edges? And, what was the high-poly version you were baking from? Can you show us how that bake ultimately looked?