Cage edit and mesh normals

Rhino Rhino for Windows
1

It looks like _CageEdit doesn’t do the right thing with the mesh normals:

Pasted image617×691 420 KB

_RebuildMesh heals it, I know.
But it should be right after _CageEdit, especially as one can’t think of anything went wrong.

Thanks,
-C-H-A-R-L-E-S-

CP_MeshCage.3dm (67.8 KB)

2

Hi Charles- thanks, I see that. For now, RebuildMeshNormals may work as well without throwing away all the other info on a mesh that might be there- it’s worth a try RebuildMesh is ‘a bigger hammer’

-Pascal

3

Thanks for the hint!

If it is expected behavior of CageEdit, then this should be mentioned in the help.
Or is it a bug?

-C-H-A-R-L-E-S-

4

Hi Charles - it seems like a bug to me, off hand, but I’ll test some more.

-Pascal

5

Hi Pascal,
while at it I remember that Flow (alongsrf) expresses this behavior as well. (Not rebuilding normals of deformed meshes).

-Willem

6

Hi Willem- it looks to me like FlowAlongSrf gets it right- do you have an example where the normals do not get flowed properly?

thanks,

-Pascal

7

Hi Pascal,

if it is a bug, how about recording it in youtrack?
The chances to get fixed are higher than without… :smile:

-C-H-A-R-L-E-S-

8

Hi Pascal,

See attached example.FlowMesh.3dm (410.4 KB)

Pasted image718×427 75.6 KB

Thanks
-Willem

9

Hi Willem- thanks- I see that it does not shade right and indeed it is not 100% correct, so there’s a problem there but I would say, at the moment, that this is not the same as the CageEdit problem where the vertex normal vectors are apparently exactly the same after editing as before.

Hmm- I just tested again and CageEdit did about the right thing… still digging.

http://mcneel.myjetbrains.com/youtrack/issue/RH-30815
thanks @Charles

1/4/2016:
@Charles, @Willem - see Mikko’s comment on the above bug track item…

-Pascal

10

Hi Pascal,
Thanks for the update.

@Mikko:
What is your argument against ‘rebuilding’ the mesh vertex normals? Or better stated:
What is the purpose of leaving vertex normals unchanged?
The vertex normals are obviously wrong from a user perspective and I dare say also from a geometric perspective.

Pasted image1135×255 23.1 KB

If I were to deform a NurbsSurface there is also no issue with distorted vertices on the render mesh.
Deformed meshes with skewed vertex normals will not render as expected, misrepresenting the mesh shape. As an example also see my post above with the flowed mesh

My arguments to let Rhino handle this in the background, are:

  • The command _RebuildMeshNormals can be considered for advanced users and should not be needed to make a deformed mesh render properly
  • To fix the ‘bad’ normals one needs to run _RebuildMeshNormals yet that will release the mesh from the cage or flow history

Thanks
-Willem

11

Indeed interesting that this obvious bug is now considered as not existing.
State = Closed ?

@Mikko
Must be a misunderstanding.
Or?

-C-H-A-R-L-E-S-

12

@Willem and @Charles, I hear you. Based on your reaction I changed the behavior. It should work better in a near future build:
https://mcneel.myjetbrains.com/youtrack/issue/RH-30815

1 Like