Mistakes were made these past two weeks, but I nevertheless have valuable pointers for the next steps I plan to take.
To save myself the time and the trouble of sourcing or modeling and rigging a new arm, I dug through some files from an undergrad organic modeling class and found an old rigging file for a full humanoid. This model was provided for students by our instructor for a rigging project, and the arm has an IK rig. I figured that for this project, the quick and dirty “hacking off the rest of the body and rig” method for focusing on the arm would suffice—more on this later.
These were the controllers I left in and the mesh’s appearance in its default A pose:
I picked a few test poses in order to test the range of the deformations.
This specific pose has the arm raised above the shoulder, and shows both a small amount of clipping at the inner elbow (even with the relatively open elbow angle) and also a strange divot in the shoulder, right above where the deltoid meets the shoulder top on the model, where the deltoid doesn’t move fluidly with the upward motion of the arm.
In contrast, this pose brings the arm down and close to the body. We still see a little bit of strange deformation in the inner elbow, but at this angle we can also see some odd ballooning happening to the mesh next to the elbow joint, where it juts out instead of settling smoothly like flesh. Additionally, there’s a small amount of clipping at the armpit, where again, the skin next to the pectoral isn’t properly being displaced.
This pose only exists to illustrate the elbow situation in an exaggerated pose. The elbow seems to fold like a plastic straw that’s been chewed on a bit, rather than smoothly deforming. LBS and standard weight painting seem to really struggle with acute angles and flesh on flesh displacement—the elbow, the armpit in the earlier photos. I would assume it would struggle greatly to achieve a nice, natural displacement for something like the squash that a crouched leg gets when the calf and hamstring press together.
Unfortunately, I quickly found out that hacking and slashing at old fully rigged models was a rather stupid idea when I actually went and opened the weight painting tool and discovered that for whatever reason, I could not easily find nor edit each joint’s influence over the mesh, which put my endeavors to experiment with weight painting at a standstill. This honestly isn’t a big setback to me, since my original intent was to weight paint the arm from scratch anyway, and I already disliked the lack of control over the test poses the IK rig was giving me.
Now, my new plan is to quickly re-rig the arm with an FK rig, and work on weight painting with the problem areas I identified with my old rig in mind. I’m going to focus primarily on seeing how accurately I can get the elbow and armpit to squash, but I also specifically want to see if I can tweak the shoulder joint’s influence over the shoulder, clavicle, deltoid, and pectoral area to get it to behave more naturally by pulling upward properly when the upper arm is raised.
While I am behind schedule and intended to start working on blend shapes by now, I’m fairly confident I’ll be able to both rework the weight painting and supplement it with blend shapes relatively quickly, and should still have plenty of time for Maya Muscle later.