I’m relatively new to rigging & skinning, but am far enough along to be somewhat clever… still, far from skilled. I have studied ManCandy but don’t fully understand him (yet).
My current WIP here: http://blenderartists.org/forum/showthread.php?p=1554184
I’m currently working on the rig, more specifically for the wing. The FK chain is shown in the attached photo. There is an IK chain and the FK chain is set to copy and of course a IK FK switch with IPO driver. The other two constrained bones are to attempt to hold volume in the bicept and forearm.
Anyways, since the arm is so “flat” and “circular” (like a frisbee), when the forearm bends into the bicept the creased deformation is… well could be better.
Finally, my question. Would harmonic coordinates work better here (mesh deform modifier)? or do I stick with the straight armature? Also, what considerations are there when deciding from rig to rig? (ie. why use one over the other?)
I did do some searching here (in the forums) but didn’t quite find what I was looking for. My best guess at this point is when you want the “detail” in the creases you use the harmonics & when close enough is alright, you just use the armature.
I have only used the meshdeformer a little bit, but I will take a crack at this. The best advice I can offer for knowing when the armature should be in control of the mesh would be to look at the distance of the bones from the surface. The farther the bones are from the surface of the model, the more useful the meshdeformer will be. The closer, the more redundant it becomes. In your image I would suggest using the meshdeformer for the wing, and then switching to the armature just before the wrist.
Just in case you don’t know, the influence between the two can be blended with a weight group (usually called ‘NoMDef’). The weight group is entered into both the armature and meshdeform modifier ‘vgroup:’ fields, and one of them is set to inverse (depending on how you’ve painted the values). This will prevent overlapping influences leading to double transforms.
Thank you very much, that ought get me on the right track. I did just (quickly) try a cage on the whole thing (for kicks) for educational purposes. There are some minor issues with the base mesh that cause the bind to be confused, but I have to say it was pretty cool (as I’ve never seen a deform go before). I’ll try the blend (I didn’t know that), it does seem that would work well. Again, thank you very much for your time.
It looks like you’ve created the meshdeform cage only around the area to be deformed by the meshdeformer. First, it is a good idea to make sure the meshdeformer does not penetrate the deforming mesh anywhere - surrounds it completely. Second, it becomes harder to blend the influence if the meshdeformer abruptly ends at the blend point…
In this article which first introduced the meshdeformer there is a chinchilla example meant to demonstrate clearly how the meshdeformer works in conjunction with the armature. Notice that the deform cage surrounds the character entirely. If you go into weight paint mode and activate the ‘NoMDef’ group you can then see how the influences are being blended. The ‘NoMDef’ group is the influence vgroup on the armature, and the inverse vgroup on the meshdeformer in the modifier stack. This means the armature controls the red areas and the meshdeformer controls the blue. The influences can also be 50-50 this way where needed
