PDA

View Full Version : LW9.6 Breaks IK of this 7.x Scene



Chrusion
02-26-2009, 07:54 AM
Load up the attached LW 7.x scene and notice how all the IK-driven hydraulic cylinders are screwed up... that is, all the IK params are the same as when in 7.x, but nothing is going to their goals all the way.

What setting(s) was added that I must be overlooking to fix this? I've tried several permutations of goal targeting. For instance, if I change a piston to Point at Goal, it pops right to it's goal, but then it's partner cylinder is thrown like 90 deg. off base. When both are set to Point at Goal, the whole IK system is "disengaged."

Notice, too, that when freshly loaded in, each IK pair appears to be doing a 50/50 blend with it's partner, so I tried upping goal strength. Doing so on just one of the cylinder/piston pairs makes the one part get closer to its goal and sends the other farther away. When both are set to high values, like 1000 or more, they behave the same as at 1... "50/50 blended."

Any ideas?

jay3d
02-26-2009, 09:11 AM
I think there's something wrong in this scene, using bank instead of pitch seems not natural for this kind of IK, I think those cylinder arms should be aligned again to use pitch rotation, and a simple point at target should work for cylinder motions

Chrusion
02-26-2009, 01:25 PM
If using IK on pitch wasn't wrong in LW7 (and probably 8), then how did it become wrong in 9.6?

What you wrote says to me that LW cannot now properly perform IK rotations on pitch and/or heading... that you must change/revised/remodel existing assets in some way (that most likely will break past uses of such models) to only use bank rotations for IK. How can this be right and normal?

Confused and feel the answer was a bit evasive.
.

jay3d
02-26-2009, 01:33 PM
Sorry for the confusion, what I meant is that the current scene is using IK on the bank controller, and should be using it on pitch instead,

I've done some experiment on the scene to correct the broken IK, and i'll post it shortly

Chrusion
02-26-2009, 04:58 PM
That still doesn't make sense. All the IK objects are supposed to rotate on bank, not pitch.

Thanks for helping to figure this out.
.

Chrusion
02-26-2009, 06:49 PM
HA! Found the culprit. Unaffected by IK of Descendants was Off for the Parents of the cylinders and pistons. Soon as those were enabled... BLAM! Everything aligned (save for the tiny one whose pivot point didn't translate in and needed repositioning on Z).

I have no idea if Unaffect by IK was on for those objects in LW 7 or 8, but at any rate, let's call this one Solved. :D
.

Castius
02-26-2009, 07:12 PM
Unaffected is much more important in the new IK solver. This is about the 5th time I've seen this. Old rigs need to to use this to help focus the IK solver.

I think at this point i want to call this a bug/feature request. I think the IK solver should see that it has no more channel above it that need IK information. IK should not be breaking this bad because this setting is not set.

jay3d
02-27-2009, 12:13 AM
HA! Found the culprit. Unaffected by IK of Descendants was Off for the Parents of the cylinders and pistons. Soon as those were enabled... BLAM! Everything aligned (save for the tiny one whose pivot point didn't translate in and needed repositioning on Z).

I have no idea if Unaffect by IK was on for those objects in LW 7 or 8, but at any rate, let's call this one Solved. :D
.

oh, thanks for the discovery! :thumbsup:

this one did not come to my mind, sorry