Weight distribution on mirrored objects

Hi,

I am creating hands for a game, and stumbled across a strange thing I struggle to understand.

I created this simple model to isolate the problem.

1: First create a simple armature (two bones sticking out of a root bone)
2: Put a sphere each on the two bones sticking out. Then select them both + the armature and parent them using automated weights.
3: When selecting the first sphere+connected bone in weight paint, it is clearly strongly connected.
4: The same on the other side.



5: Now, ctrl+Z back until the spheres are not parented.
Choose one of the spheres in object mode, and then transform->mirror. (alternatively -1 on scale X)
Select them both again + the armature, and parent with automated weights as before.


6: When looking at the sphere that was mirrored before parenting, with weight paint+correct bone selected, there is something wrong with how the weights are made.

This is just a simplified example, when doing the same process on my arms and rig, the mirrored arm gets incorrect and weak-ish weight-distribution, while the original arm get a perfect distribution.

I looks like there is someting basic about mirrored objects in Blender I have not understood yet, having mostly used Maya.
Does anyone know why this is happening?

Ahh… never mind… You had to APPLY the transformation for it to weight it properly afterwards. So just using mirroring makes it still the old object, just with a mathematical invert in one axis. Auto-weighting doesn’t care about that, or doesn’t tell you. Trap for young player as they say. It is these fun days transitioning to a new program and tearing your hair as to why things doesn’t work as in your old program.