|
lw ik seem unusable? tips from someone who agrees with you....
1. 3 bone leg is unusable in a real setup. well, you can if you use 2 ik goals, one to control the leg and another one with a weaker strength to control the facing of the knee. but if you want to use the bank of the thigh bone, you're out of luck. as soon as you take bank off of ik control and fk it, any animation you've done goes to hell and poporama ensues.
2. 4 bones is the minimum for a leg or arm. the 4th previously unmentioned bone is a 'buffer bone' that allows you to set the thigh or upper arm to be controlled by ik on a SINGLE CHANNEL ONLY (and i would imagine this is the only way that lw ik will be usable in real world settings).
so knee is ikd on pitch, thigh is ikd on pitch, buffer bone is aimed down the z axis (either + or -) and it's BANK is controlled on ik to allow the leg to lift to the sides.
3. now, you can control the aim of the knee by rotating the heading of the buffer bone. and although the knee ends up not pointing where the buffer bone is, if you can just tune out the buffer bone, you'll be able to control the facing of the knee.
4. PARENTING THIGH TO THE BUFFER BONE:
make all your bones as skelegons but DO NOT parent the thigh to the buffer bone in modeler. the thigh should be placed at the ROOT of the buffer bone (fat side) so that when you rotate the buffer bone on heading, the thigh rotates from the buffer bones pivot point.
but if you parent in modeler, the leg chain will jump to the END of the buffer bone, not stay where you want it to, at the root.
so, parent in layout with PARENT IN PLACE option turned on. viola
5. DRAWING A BEND AT THE KNEE DOES NOT ACTUALLY DO ANYTHING AT ALL!!!
or this is what i used to think when i drew that bend and the chain would flip and invert the other way as if there was no recognition of a 'preferred angle'.
actually, during the 5.6 days, i complained about our ik and brought up how other apps had a preferred angle bend that can make ik solving much more reliable. and then, some time later, i heard people mention tutorials in lw that mentioned instituting such a bend... and i always thought that more than anything, it was a placebo effect!
happily, that is not the case anymore.
the bend does work...
BUT ONLY IF THE THIGH BONE AND THE SHIN BONE ARE ON IKing ON PITCH!!!
in the past, whenever i parented the thigh bone to a buffer bone with 'parent in place', the orientation of the thigh's pitch FLIPPED... but i didn't know about it.
this resulted in ik experiments that sometimes seemed to take the bend into account and other times not. now i know why.
to make sure that when you attach your thigh bone to the buffer, your thigh's pitch orientation remains intact, make sure that the buffer bone's pitch is oriented as you would want the thigh to.
(p.s. it might be that the bend would work on other channels too (heading and bank) but that both bones involved in the calculation must use the same one... haven't tested that theory but i'm pretty sure that both must match)
i wrote another thread earlier that made me come to the conclusion that lw's ik is essentially unusable. but i'm happy to see that i'm wrong about that.
it is still clunky and nonintuitive and primitive and makes you jump through more hoops than you should....
but i'm happy to report that it is functional!
thanks jeanphi and castius for pointing out the finer points.
jin
|
Linear Mode
