Normal map baking - problem
polycounter lvl 13
Hi.
I want to bake my normals for a weapon model.
this is highpoly model:
this is Lowpoly model:
and this one is baked model preview with xoliul shader: (baked using xNormal)
hard edge UVs are broken but as you can see I still have strange results.
I don't know why this happens but I guess strange borders near element collision places is because of cages. :poly141:?
do you think I have to bake elements separately?
what other thing should I do for hard edges (which are not really hard in Highpoly model)
this is my first try on hard surface baking, so any kind of tips are helpful..
I want to bake my normals for a weapon model.
this is highpoly model:
this is Lowpoly model:
and this one is baked model preview with xoliul shader: (baked using xNormal)
hard edge UVs are broken but as you can see I still have strange results.
I don't know why this happens but I guess strange borders near element collision places is because of cages. :poly141:?
do you think I have to bake elements separately?
what other thing should I do for hard edges (which are not really hard in Highpoly model)
this is my first try on hard surface baking, so any kind of tips are helpful..
Replies
(move individual pieces away from each other?)
now I'm trying to bake with scanline renderer and MR, results with scanline is much better than xNormal.. :poly142:
Make a keyframe, got to frame 100 (or whatever), move the opjects pices away and create another keyframe. Then bake the normals. When you need the model to be one pice aggain (for AO) just go back to your inital keyframe and bake there.
what he said, except it's better to just turn it off, as it's non destructive
It's just a common misconception that all meshes have to be in only one smoothing-group when you bake normals. Now, you could either use one smoothing group and throw in supporting edges if needed, or split it up into different smoothing-groups. Both can give good results, depends a bit on the mesh I think.
Not sure which one is the most efficient really, but guess that comes down to the mesh as well. In fact, other than that the one smoothing-group thing can capture corners better sometimes I have no idea why there even are two methods