ICE Rigging

[dwigfor]

I've been trying to create upvector for 2 bone ik chain. I have it working with roll parameter; dunno how to use null control object.

[Hirazi Blue]

As a "learning aid" you might want to look at this post from the Mailing List, in which Ahmidou Lyazidi presents an "IK compound with upvector support". ;)

(BTW - with this post I do NOT claim to understand it in any way! )

[curly]

Some links I have found on the web


http://www.vimeo.com/21151418

http://www.xsibase.com/forum/index.php? ... ;start=5=3

Moderator edit: please use the URL tags - HB

[Pancho]

In the rigging section here a guy shares his rigs. Two or three of them are ICE based, so I need to disect them and check how he did things. Would be so cool to have an arm or leg compound and you just need to place a few nulls and it works.

Cheers!

[Lord Futzi Voldemort]

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???

I t might be a good idea for the common softimage user, to make himself familiar with the concept of workgroups in XSI.
It is not only a flexible way to manage multiple sets of Plugins, Shaders and Addons across a network but also a very convenient way to keep personal stuff, like for example compounds aout of harms way when installing or reinstalling new or old software versions, or manage the accessibility of tools to specific users.
The latter may be the reason, why the kinematics compounds are stored in a separate workgroup, since they are meant to be mainly used by td´s, who build compounds or riggs, which then are accessible through the workgroup for a given, well, workgroup.
Storing things in a separate workgroup doesn´t mean they aren´t part of the software at all. It just means you can switch it on and of very easily.

Achim

[dwigfor]

Woohoo!!

I think I finally got the 2 Bone IK working with an upvector!!

Thanks for the mailing list post. I haven't tried out Ahmidou Lyazidi's compound, but I did give his reference (http://www.google.com/url?sa=D&q=http:/ ... sept98.pdf) a read and tried to comprehend it... Failed to try to implement it, since it's only 2D.. I needed to figure out how to rotate around 3d..... I thought I could do everything with prebuilt compounds rather than calculate cos and stuff. After figuring out how to do it on my own, I should compare it with Ahmidou's.. His is probably more optimized and better....

I'll try to post my compound and hopefully get to make a video about it. The secret is, you need to calculate a "parent null" aligned first to Eff along x, and then aligned to the cross product along z. Cross product is calculated by 1: Eff Position Subtract by Root Position; 2: UV Position Subtract by Root Position. Not sure if it matters, but I normalized it. You then have to multiply bone1 and bone2's rotation by the "parent nulls" rotation. (Bone1 rotation * Parent null rotation). Order matters when multiplying rotations.

[Mathaeus]

btw, for IK chain setups, I'd believe it's most important to checkout how specific setup behaves with extreme low-high angles. When bones are parallel... For this one, I've finally used setup from Helge's 'ICE tutorials part 1' (I think it's still downloadable at this place). On the other side, 'cosine law' compound (as a 'hearth' of 2 bone IK setup) from first PhilPack release, had a problems with these angles.... Divide by scalar, together with square root, all on very small values - causes particles to disappear. At least, all that with point clouds, maybe ICE kine has a better precision....