May 2008

To continue my series of crappy, 1 or 2-hour practice blocks, here’s a guy being hit in the head with a ball:

I tried to do a anticipation-accent-final cycle here, though I’m not sure if I succeeded (it was kindof hard to do the accent, especially as this is not a cartoon rig).  Also, the timing doesn’t seem quite right to me.


As part of my ongoing effort to improve my very sucky animation skills, I made this one-hour block of an animation yesterday:

It basically followed some ideas I read in the Animator’s Survival Kit (great book btw).  It also pointed out a weakness in my rig, in that it was fairly difficult to invert the arch of back and have the hand stay in the same position, without using IK on the arms.   I also used a parent constraint on the box to move it, which caused problems of its own (I had to disable rotation inheritance to have it look even halfway decent because of the IK); I think it would’ve been better to hand-animate the box in this situation.

I’m thinking of doing several of these one or two-hour blocks in the future, since I don’t really have time to really produce polished pieces of animation, especially since I’m not even sure I can, without more practice.

Vector IK is a great technique for some situations where normal IK just doesn’t cut it.  The technique basically uses one control to rotate and “curl” an IK chain.  It’s also fairly simple to do.


My first real introduction to rigging was Javier Salson’a reverse-footlock tutorial.  It’s an interesting technique that results in a fairly easy-to-use foot rig.  What was significant about this tutorial is it introduced me to the idea of using an ik pole vec constraint to control where knees and elbows point.  With this tutorial, I produced my first actually usable rig (a simple carrot, that I animated my first walk cycle on).

I’m not sure if reverse footlock is the best foot rig or not.  Certainly I’ve learned a lot since first being reading this tutorial.  But this is still where I really started to understand rigging for the first time.

A while back someone wrote a siggraph paper on using dual quaternions for a more advanced skinning.  Unlike linear skinning, dual quarternions don’t suffer from the “shrinkage” problems and require far less corrective shape keys as a consequence.

I’ve used this in my own rig, and it seems to work great.  It’s implemented in Blender, and the site has links to plugins for Maya and XNA.

So, here’s my first post: Shoulder Deformation.

contrast between without and with good deformation

This is actually a topic I’ve spent a great deal of time on.  My theory for the main problem with shoulder deformation is twofold: a) people don’t realize you can’t lift your arm above your shoulder, going higher requires rotating your shoulder itself, and b) they don’t build support for twisting into their upper arm rig.

You can test both aspects of this theory yourself.  Try lifting your arm above your head without moving your shoulder.  And try twisting your upper arm and see what happens.  The upper arm twists, it doesn’t rotate in a block.  Trying to blend between the upper arm and the shoulder with weights won’t help either, a true volume-preserving twisting rig is needed.


Since I seem to have a lot of free time this summer, I’ve decided to start a rigging and animation blog.  This site will mostly feature various books and tools that sound cool, but will also document some of the knowledge I’ve accumulated in my past adventures learning the black art of character rigging.

Note that I’m in no way a professional or expert on animation or rigging (though I am fairly good for an amataur at rigging, I think).