XNormal: object nm to tangent issue
polycounter lvl 6
Hi,
got issues to convert an object space nm to tangent space nm with xnormal.
The converted map is calculated totally wrong.
I did tesselate the bake mesh to avoid skewing and want to re-bake the OS NM to my lowpoly to get a proper tangent space map.
I did export the lowpoly as sbm with default settings.
Do I have to uncheck normals, bi-normals and smoothing groups for object space normal maps?
thx
got issues to convert an object space nm to tangent space nm with xnormal.
The converted map is calculated totally wrong.
I did tesselate the bake mesh to avoid skewing and want to re-bake the OS NM to my lowpoly to get a proper tangent space map.
I did export the lowpoly as sbm with default settings.
Do I have to uncheck normals, bi-normals and smoothing groups for object space normal maps?
thx
Replies
I did explode the mesh on frame 2 and did export from frame two. Maybe this only works with the baker and not with the converting tool?
EDIT: If I export a tangent space nm from the tesselated mesh, the tangent space nm looks good.
I used handplane with obj and it works, but only with obj. If you use fbx with tangents+binormals. you'll get the same error like in Xnormal.
No, the final low poly was not rotated. Just did a re-import with fbx, obj and sbm files into max - no flip or rotation.
I always reset x form before export and it was the same mesh. I just deleted the tesselation modifier.
Maybe xnormal scales or flips the mesh on import within the tool?
You can try exporting the mesh after you've rotated 90 degrees on X, see if that helps?
Handplane might be working better because it's able to detect which way is meant to be up and account for any differences.
I baked the OS with -y ( for Max; Cryengine) in Xnormal.
I flipped it in PS to +y, and it works now. Mesh axises weren't the issue.
Didn't know, that a flipped channel with OS NM causes problems due to tangent nm conversion.
THX