For me, the current IK/FK implementation in XSI has reached a dead end, but that's probably just me. To most animators, looking at the current set of IK chain options is a daunting experience. Rather than adding more controls, I'd prefer a rethink of IK that simplifies the interface and expands our abilities.
As an XSI rigger with a reputation for experimentation, invention, and high end production experience, here's where I am with respect to IK in XSI:
-I only bones when it is necessary to solve for IK. They can become processor expensive in large numbers.
-I rarely use 3D chain behavior.
-I rarely use IK on more than 2 joints. I may use multi-bone chains, but they are almost always constrained to curves or manipulated by expressions.
-I always provide an outlet for blending from IK into FK or other control setups
-I am frustrated by the fact that IK can only be associated with special scene objects... bones.
What I'm hoping for Softimage to do, and I've put in a formal request, is to have simple, 2-joint IK behavior abstracted out into a constraint operator. Applying and removing IK should be the same workflow as applying or removing constraints. And as constraints, the IK constraint should behave the same as all other constraints in XSI where you can apply as many as you want to the kinematic stack, and blend them any way you need. You should be able to store them into Action sources and control them from the mixer.
However, as with any functionality, it has to be more than one person making the request. As long as it's just me, the idea of the "IK constraint" is going to be low priority in the huge pile of feature rec's. :-)
-Brad
> -------Original Message-------
> I was putting together a hybrid IK/FK reverse foot rig on friday when I
> noticed some odd design choices with the integrated IK/FK blending in XSI.
> It brought on some questions and ideas.
>
>
> previously, I had assumed a chain has three states of manipulation
> behavior... unkeyed hybrid ik/fk... keyed ik, and keyed fk. I had been
> going by the assumption that the way to get between these modes was by
> keying the IK/FK blend, to IK, FK, or not at all. However, it turns out
> there is a fourth state... when you key the ik/fk blend to fk, but have
> not keyed the rotations of the bones yet. In this mode, the bones can be
> rotated. however, if you pull their root, they snap to their end effector
> as if they were in IK mode ( if the effector is constrained).
>
>
> now, these crazy cool but sometimes annoying hybrid manipulation modes
> seem to have prompted the inclusion of the "Forc! e IK" switch.
> However, there is no "Force IK/FK" switch which eliminates all the other
> weird semi modes and leaves the bones in either pure IK or pure FK based
> on the blend, without keys.
>
>
> My frustration is that I can't create a rig that uses the internal XSI
> IK/FK switching that acts correctly when it comes in without keys. It
> seems silly to me to have to have fcurves on my legs in order for them to
> act correctly.
>
>
> So, anyone have any thoughts? Unless I'm missing something (entirely
> possible) it seems to me there need to be a "force IK/FK" switch on there
> so I can constrain the IK leg but blend it to the FK leg which is totally
> free, due tp ignoring the constrained end effectors completely, without
> the need for fcurves to be on the rig already.
>
>
> btw, please don't tell me "just make a compound leg." I know that
> already. The point is to make one that u! ses pure internal XSI IK/FK
> blending.
>
>
> Thanks.
>
>
> -brad
---
Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in body:
unsubscribe xsi
