View Full Version : IKB Bone Dynamics: Anyone use it?

03-18-2009, 04:32 AM
Are there any examples of bone dynamics out there I can look at? I looked at Will V's intro video. I am trying to build a bicycle that acts as if it's being picked up by it's front tire and being moved about. I'm trying different stuff, but I'm just getting variations of embarrasingly bad.

Is IKBooster's dynamics only good for tentacles? I guess what I am trying is akin to ragdoll physics. Are there any examples I can look at so I can say, "Ah HA! It IS possible!" Right now I'm souring on this prospect. :/

Edit: If it matters, I am using v9.6

03-18-2009, 05:59 AM
I wouldnt use IKBs dynamics for something like a rag doll simulation. I think that a 2point polygon skeleton run by cloth works better. The problem is that IKB doesnt do any dynamics on the root of the ragdoll. Now if you were shaking a doll in a dogs mouth etc I might use it.
What I have used quite a bit of is IKBs Dynamic Effect. Essentially its softbodies for bones or other items in a heirarchy. I used it on a couple of shots where a bullrider gets his hand caught in the rope wrapped around a bull.
I animated the body sort of following the bull with some delay. Then I went on each arm and leg used Dynamic effect to create the bounce on those based on what they body was doing. It worked out much better then expected.
The nice thing about doing this sort of thing with IKB is that you also have the tools to easily edit the results of the dynamics afterwards. I found that the dynamics also seem to respect any limits and such I put on the limbs or which channels were locked. This is useful to help keep things from going out of control.
You can also apply dynamics to any part of your rig so its not an all or nothing situation. This would allow you to use the dynamics a little bit at a time on your rig, build it up and tweak as you go.

03-18-2009, 06:05 AM
Here. You take a look at this and see if what I am attempting is realistically possible or not. Right now when I click calc the bike just disappears after frame 0.

I couldn't use a 2-point poly with ClothFX because I need to restrict the rotations of the hinges of the bike setup.. The front tire can rotate any which way, but the fork that connects to the tire can only rotate on the pitch, and the body meeting the steering column can only rotate on heading. If I grab an element and drag it, it works just as IKB should. It's just the dynamics are hosed.

Edit: Shaking a doll in a dog's mouth? That's what I want! I want to be able to pick up the bike by it's tire and carry it around.

03-18-2009, 01:52 PM
Ill take a look later today. :)

03-18-2009, 04:04 PM
Hope you don't mind. Had a peak myself.

It seems that it is the inclusion of the objects in the hierarchy that are causing the main issues. Likely their zero length is confusing the dynamics solver.

So I'd redo the scene with the bike in one layer and 100% weights assignments. See attachment. It seems to work better. Why the fork does an odd pitch rotation I haven't got a clue (I removed the shorter bone, but that wasn't it).

Cool setup/test btw.

03-18-2009, 08:03 PM
Hi Dan,

I looked at your version.. Much closer to what I am after.. Just tweaking the dynamics settings so it doesn't fly around quite so much.

Zero length? I don't understand. What element or part of this scene had zero length?

03-18-2009, 08:10 PM
You're right though.. something in my model is making it freak out. I loaded my version, cleared all of the geometry and it calculates with no problems! Very strange. Zero length.. Hmmm.

Edit: Via process of elimination, it's just the front tire object that is causing the problem. Even though the rear tire is a clone of the first, only the front tire is causing the problem. So strange.

Edit Edit: I loaded my scene, zeroed out the X and Y position values of the front tire and no more problems. Tres bizzare, no?

03-18-2009, 08:59 PM
One thing I've noticed.. Well, other than this dynamic solver being occasionally completely bizarre.. Rotation limits are not treated as collisions. If a hinge hits a rotation limit, it will be stuck there at the limit until the intended rotation swings back into the allowed limit range. It would be more natural if the rotation "bounced" off of any placed rotation limits.

03-18-2009, 09:04 PM
Cool. With zero length I was just thinking the solver needs something which defines a line. Bones do that naturally whereas nulls and items don't, since they only define a position. But it seems IKB just calculates the distance from parent pivot to child pivot, so it works nomatter what type of item. Sorry to be the cause of unneeded confusion.

I hope we'll get a chance to see the end result.

03-18-2009, 10:18 PM
Carm, select each of the bike parts individually, set the keyframe mode to current and go to dynamics edit. Turn off dynamics for each part of the bike but leave it one for each bone. Once I do that the bike stays together.

03-19-2009, 12:05 AM
Sorry to be the cause of unneeded confusion.

On the contrary, you put me on the path to solving the issue. Thank you!

I hope we'll get a chance to see the end result.

Unfortunately, this is an NDA project. :( No harm in showing you a bicycle.. But not the rest of it.