Motion 5 Test Project: Inverse Kinematics (Character Animation Rig)

Inspired by Mark Spencer's thoroughly ingenious MacBreak Studio episodes exploring advanced linking behaviors recently, I thought it would be fun to try and build a robotic arm with genuine inverse kinematics.

So here it is (the first version as simple blocks so you can see the translation and rotation more clearly, the second dressed up as a character arm):




In this case the rotation and translation of the upper and lower arm are being driven solely by the movement of the hand - it might seem simple but it's quite a complex issue:

en.wikipedia.org/wiki/Inverse_kinematics

Suffice it to say that with a lot of hidden nulls (actually I've unhidden them to show you some of their workings) and some "computational" link tricks that I hadn't tried before, it pretty much works. Not perfect, but good enough as "proof of concept".

I don't think that I'll be trying to build a full 2D character just yet - could be very time-consuming.

Still it's good to see that even pretty complex tasks like this are achievable in Motion - you'd think you'd have to have expressions to be able to do it but it just takes a bit of thought and planning and you can often get there.

Simon Ubsdell
www.tokyo-uk.com
Visit my YouTube Channel

Simon,

Quite impressive. Thank you for this.

- Ronny

Thanks, Ronny.

It took an awful lot of head scratching to get there and it probably doesn't have any useful application but it's an interesting test of just how much you can do in Motion.

Who knows, maybe I will try and build that complete character rig after all

Simon Ubsdell
www.tokyo-uk.com
Visit my YouTube Channel

Simon,

Do not underestimate the meaning of this. Just seeing this kind of motion can be created in Motion is a revelation to me, and I have been working with Motion since v.1. I"m really impressed. Yes, you definitely should build that character rig

- Ronny

Also, Grouping and sub-grouping layers, in conjunction with setting anchor points will create joints that hold together, too. I've done quite a bit of this before. Had a project about a year ago where we had to do stick figures, and it worked really well.

BenB wrote: Also, Grouping and sub-grouping layers, in conjunction with setting anchor points will create joints that hold together, too. I've done quite a bit of this before. Had a project about a year ago where we had to do stick figures, and it worked really well.

That's what I've done with this character rig here:



But that is a very different order of difficulty to the inverse kinematics route where the rotation and translation of the joints going up the chain is influenced by the end-effector, in this case the hand.

Consider that as the hand moves up or down, the lower arm needs to rotate to follow it but it also needs to move slightly as well - and neither of these are 1:1 correlations by any means.

Moving up or down and across makes that even more complicated.

Then add in that the upper arm needs to follow the lower arm but again in nothing like a 1:1 correlation and it gets very complicated indeed.

You can't use grouping to achieve this effect at all - you have to use a complex system of links, including some that are applied to nulls that are there simply to perform maths calculations.

Just to clarify, in the test animation the only thing that is being animated is the hand - the lower arm and upper arm are following along due to the inverse kinematic linkages.

Simon Ubsdell
www.tokyo-uk.com
Visit my YouTube Channel