ICE Rigging

[Pancho]

Just wanted to start rigging a simple character in ICE. Did I miss something? I can't find any ICE IK solvers! Aren't they part of ICE since 2011.5? So, everybody talks about ICE rigging. I'd like to do it! But how, if the ICE compounds aren't there? Can anybody enlighten me, please?

Cheers
Pancho

[Hirazi Blue]

I guess they are still in the "the ICE_Kinematics workgroup located in the Application/Workgroups folder of the Softimage installation location". ;)

[Pancho]

Thanks! But how do I make the compounds part of SI? In the version history I do see that there is a special folder in the ICE compound library. How do I add them like this and WHY aren't they part of SI already???

[Hirazi Blue]

You could try copying the compounds from the Workgroup to your factory compounds folder or the user compounds folder... As to the why...

[Pancho]

Thanks! Copying worked!

But why isn't it a native part of SI?

[Hirazi Blue]

I assume you know (but it doesn't hurt anyone repeating it, if you already do ),
the ICE Kinematics Workgroup Compounds are documented
on this page of the Softimage WIKI ;)

[origin]

afaik they're not supported (read: not tested enough) to be part of standard SI.

Thanks! Copying worked!

But why isn't it a native part of SI?

[Pancho]

I don't understand why Autodesk is offering ICE modeling, while ICE rigging is not an official part yet. Should be vice versa.

[origin]

I meant THESE compounds, not ICE rigging. You can built them on your own... If you expand them you will see it's only math

[Pancho]

I'm just a bit suprised that the solving is oriented on the x- and not the z-axis. Is this common in SI rigging or is it because the creator of the rabbit rig and the compounds comes from Max?

[dwigfor]

I was wondering about that as well (Calculated along X).. It was a pain changing my rig once I discovered that....

[Pancho]

I just corrected the 2 Bone IK compound so that it works with my regular rigging and coordinate system. I really don't get why this hasn't gone official yet. Change x- to z- and it works for the SI users who are used to rig this way.

What I don't get either is that there are almost no tutorials except the one from the rabbit rig creator. This is supposed to be one of SIs strength, but no one has any info about it. Reminds me of IK booster in LW.

[Mathaeus]

I'm just a bit suprised that the solving is oriented on the x- and not the z-axis. Is this common in SI rigging or is it because the creator of the rabbit rig and the compounds comes from Max?

If you mean 'axis that points to the next bone in chain' afaik that's the same in both , positive X. And Z for rotation of 2d chain. However, 'Direction to Rotation' ICE node has Y for 'point at' , X for up vector, there is no built-in switch, as it is with XSI direction constraint. You need to rotate vectors for 90 degree around some axis, then to apply mentioned node.

[Pancho]

O.k. I started the other day to try and rig a character with ICE. The rabbit rig is my only inspiration and guidance except from the rabbit rig videos. The lack of examples is frustrating.

First of all: The documentation on the Softimage Wiki Page is a slap in the face. It's the same frustrating experience as with Newtek and Lightwave. I don't need to see the node inputs and what kind of data input it allows. I want a decent explanation what each node does and how it can be used. This mininmalistic technical sheet is utterly useless.

Well, now for the ICE part. I like the way the leg, arm and toes get calculated by the same node with the help of arrays and there is no need for ice trees for each finger, arm, leg. Great!

The wired thing is that somehow the "bone/env" nulls jump around as soon as the rig gets activated. Usually bones just change the side. I don't know how to stop this or how to bring in all the nulls in a correct order, so that the left side gets calculated before the right side and everything stays where it should stay.

In the group folders the left bones are alway listed first, so somehow neither the order in these folders nor the alphabetical order does matter in terms which bone gets calculated first. There seems to be an "invisible" ID which determins the order.

Can anybody enlighten me?

Also, if I'd post the rig, would anybody like to participate in setting the whole thing up? Would be cool to generate a generic ICE righ which would be easily adaptable for everybody because of its simple nature and lack of expressions which would make any other rig almost impossibe for transformation.

Cheers
Pancho

[Pancho]

... and another question.

How do I combine the two IK chains for the leg and foot? As the ankle null serves both as the goal for the leg IK and the starting position for the foot IK chain, unfortunately the foot gets torn of the leg if the foot controller (to which the ankle null is parented) is further away from the leg than it's maximum radius. Should be properly very simple to fix this, if one does know.

With regular rigging the foot IK chain wouldn't be torn of the leg, as it's parented to the leg effector.

Cheers
Pancho

[Pancho]

I found a way while shopping in a grocery store.

Actually what I was looking for is the bone tip position (goal position) of my parent chain (leg). I connected this tip position to a tip position input in my compound which calculates the foot chain. As I just need the position and not the SRTs I just exchanged the position in the SRTs of my chain root for the foot and voila. It worked!

Any interest in working together on this?