Im looking into areas that can essentially be shared  – modular continuity between arms/legs etc on biped and quadrapedal rigs (even bird rigs). Its opened up some interesting ideas namely to first break the lower arm in half and some oddities in the foot.

Breaking the lower arm in half basically allows for a front arm to act like a front leg of a quadrapedal – why is this important well for one thing it allows a quadraped to act like a biped and vice versa. It doesnt mean you would nessesarily use it in a biped rig, but its a simple additional that allows for the control.

For the foot rig – ive seemed to find a stumbling block namely where the control of the foot goes? In human locomotion the pivot exists at the heel, but when its on the ground its at the ball. The problem is even more compounded in that we both swivel and hinge about the heel – hinging is a simple heel<>ball setup, but swivel is more of a problem. Do we change the transform space of the main control if its at the ball or add an additional layered control?

The problem is exactly the same and virtually opposite if the main control is the ankle, you get heel and swivel control but ball of foot swivel is lost.  Its not truely lost, you could drive its z rotation of the roll controls z rotation but, as where dealing with ik system our trigonometry plane space would freak out and the foot as a whole would twist – the old gimbal problem looming as two rotations overlap each other.

1. August 1, 2007

About the breaking the forearm you mentioned.. how will you handle the “locking” when it`s in biped “mode” I mean that the second joint must stays unanimatable then…

2. Charles
August 1, 2007

Basically its a very simple system – when in biped you just dont animate the lower lower (broken) limb. It relys on a floating ik chain about the fk hierachy of the broken limb, the limb rides along with it but allows for fk animation. So you have the standard conrol when its on the ground but in the passing position you animate the fk broken limb – the only issue is drift, but its not too much of a problem.

3. August 14, 2007

I was looking at my implementation of reverse footlock, based on:

onhttp://student.vfs.com/~m07goosh/Tutorials/reversefootlock/reversefootlock.htm)

. . .and I noticed it actually can handle both heel and
ball swivel (this rig just never ceases to surprise me).

Basically the rig leaves you with four controls, one
main foot control (that does some automatic
rolling of the foot) and three IK targets to tweak
things, one at the heel, one at the ball of the
foot, and one at the toe.

I’ve found if I rotate the three ik targets the
foot swivels around the ball of the foot. I’m
trying to tie them into a single bone for
swiveling, but maintaining independent
movement; I’m not having much
luck though as the necassary features to do that
in Blender are still new and rather buggy. In
fact they won’t even be in the next release,
I’m using a latest development build

Honestly I had no idea that was possible until