When I was using the Jacobian-based FullBodyIK in UE4, I found that the rotation of the effectors seem to have little effect when I do not enable Force Effector Rotation Target. Only when I set Max Iterations to a very large number can I achieve the desired effector rotation.
By observing the source code, I noticed that during the calculation process, the scale of position values (0 to several hundred) seems much larger than that of rotation values (0 to 2π), causing a mismatch where the weight for achieving the position target needs to be much greater than that for the rotation target. I’m not sure if my understanding is correct.
In UE5, when using PBIK, it seems that the effector’s rotation is directly applied and does not significantly affect other bones. Does this mean that the three dimensions of rotation are irrelevant for solving FullBodyIK? Because I only need to calculate the position and then directly apply the target rotation to the bound bone, I can then set preferences for the intermediate bones through constraints/bone settings. Is that correct?