Hi Polycount,
Using XNormal to bake a Normal map, but I'm getting some horrifying smoothing errors and I don't know what's causing them.
- My cage on the low poly mesh looks fine. Everything should be triangulated correctly.
- UVs look fine to me.
- Normals look fine on the High and Low poly.
- Test bakes in 3DS Max look OK to me. Never great bakes but they're showing no errors
High Poly and Low Poly:
What XNormal has spat out for me:

What 3DS Max produced (just a 1k test render to error check):

I've uploaded the FBX objects to
Sendspace
NOTE: For some reason they export with the "OpenSubDiv" modifier. You can just delete that modifier from the stack when you import and apply a Turbosmooth
The topology for the high poly is a bit all over the place but it smooths correctly :P
Any advice/help is much appreciated as always.

Thanks!
Replies
I have the low but not the high - good shout. I'll give that a go now
Never look at a normal map in 2D to diagnose problems, apply it to your mesh and see what it actually looks like.
I've got padding set to 8 at the moment. I've fixed most of the stuff. The remaining issues are where the UVs have just skewed for some odd reason - presumably because there were normal errors.
I've got it open in Marmoset too (awesome work on that by the way!)
Also this thread may help with skewing issues: http://www.polycount.com/forum/showthread.php?t=147227
Just a couple of little issues left:
The octagon shape with the circlular extrusion might need a few more sides to help smooth it nicely.
Couple of smoothing errors here - probably smoothing groups or too little geometry?