FUBAR Rig and I can't find the reason?


So I made a basic dinosaur model and put a rig to it but it appears to be influencing random parts of the mesh (along with the parts it’s supposed to influence).

-My first thought was bad weights but everything seems to be ok (?) and I did my best to paint all parts to zero influence (shouldn’t this actually mean that it’s impossible to move any part?)

-The model is admittedly kind of crappy as it was more an excercise in rig learning than modellng. But bad topology cause the issue? But I hardly think it is and it’s not actually terrible.

-No double vertices, apparently.

-Am I making a bone hierarchy mistake?

-Could it be that it’s just the one mesh that’s randomly broken?

-I tried a different thing in the second attachment where I have fewer connected bones – same problem.

Known issues (although I wouldn’t imagine they’d cause rigging problems):

-I created a weird thing down the middle of the mesh with the mirror modifier.

-The legs don’t have inner thighs.

v 2.78
Mac OSX 10.6.8

Marco :smiley:

dino2.blend (1.32 MB)

dino5.blend (2.82 MB)

Sorry – v. 2.79

dino5 doesn’t have a rig. dino2 isn’t parented to its rig, doesn’t even have any weights, and the rig itself moves fine. What you should do in the future is give a copy of the exact file that’s giving you problems. Otherwise people have to guess what your problem is.

I’m looking at dino2. And what’s my guess? My guess is that you used automatic weights to parent the dino to the rig, and automatic weights didn’t work very well because your mesh has non-manifold vertices. Which, yup, is probably because of that weird thing you created with the mirror modifier.

Fix the non-manifold vertices and try again. In edit mode, vertex select, ctrl-alt-shift-m to display all non-manifold vertices. Fix what you see. Which will involve some combination of deleting them and replacing some of them and removing doubles.

Thanks! :slight_smile:

What the hell. I must have deleted the rig accidentally before the very final save then (dino5) and I’m positive I parented it on dino2.

Sorry for the confusion.

Yeah – automatic weights.

In the meantime made another one even more quickly (attachment) and it works much nicer despite me having gone through the same motions but I didn’t use a subsurf on the mesh and it’s in version 2.78.

It also doesn’t appear to have the non-manifold vertices.

dino8.blend (533 KB)

  • The first .blend doesn’t have a rig and the weights are bad
  • Second .blend has a rig but no weights
  • Bones at the back aren’t inside the mesh and ones at the front are on the side
  • Mesh type object has unapplied object scale, normals are inside out
  • The mesh is non-manifold, the interior faces probably come from incorrect mirroring

Armature modifier matches vertex weights and bones by name. The vertex groups have the same name as the bones. When vertices are weighted to one bone with higher than 0 weight, that bone will move those vertices. When vertices are weighted to two bones, that is enough to cause deformation that is controlled by those two bones.

The body vertices have many groups and multiple non-zero weights

Selected vertex is weighted to 5 bones.

It has 70

That’s a whole bunch of stuff I didn’t look into. Thanks for the heads up!

I didn’t know you could get that granular on analyzing weights.

No I meant the new one I made just now (dino8) that works well.