been working on this for a while - regular joint chain (+y, aiming down the child n +Z up), with spline IK on it, cluster the curve, have them parented to Ctrl_Curves. i have attached a sample file(.ma file).. in which i have a COG curve, shoulder, mid_neck & head curves. i know you must be familiar with this method..
my problem is, when i try rotate my COG in y, in either direction, my neck bones kinda twist ??! ggrrrrr..........#@$@()*()*&%^%%^&% .. grrr.....
i have the local rotation axis visible.. you check them..in the .ma file... in rotateX, its stable.. can anyone tell me why this happens ?
now i tried these too : i know my bone chains r having +Y aiming towards the child...but, i tried doing the same setup on a bone-chain, with +X aiming down the child.. it happens in that too.. but in this second case, i select the splineIK handle, in the advanced twist attributes, when i put it to scene UP > +Y , the flip is controlled ...
i noticed that in my chain with +Y aiming down, when my COG is rotated -90, and on the splineIK, i make the roll to 45, the bone's twist is nullified.. at COG > RY = 90, roll shud be -45 ... but at COG.ry = (- or +) 180, roll has to be 180 , and at COG.ry = (- or +) 360, roll shud be 0. i cud'nt deduce any relation or ratio kinda pattern from these values..
is this a problem with the splineIK or with my method itself ? or is it that splineIK needs the bone chains to be having +X aiming down, only ? is this anyway connected to limitation of the adv,twist attributes, which work only on the bone chains with +X aiming down ?
but the same setup , i have on my spine area, which works fine.. but same setup on the neck.. is giving me the problem. how many times i tried to reproduce it, it always, twists.. can any one guide me what could be happening wrong ?