[SOLVED] Relative orientations. Why this quaternion don't work?

Hi, I'm trying to mimic rotation from one object to another but with maintaining rotation difference between them.

On Start I calculate rotation offset

Code (CSharp):

1. _rotationOffset = Cylinder.rotation * Quaternion.Inverse(Box.rotation);

And on Update I apply parents rotation

Code (CSharp):

1. Cylinder.rotation = Box.rotation * _rotationOffset;

Problem is in case when both rotations aren't compatible (not the same and both are not equal 0,0,0 in euler angles). For example Box has (0,15,0) and Cylinder has (15,0,0). Rotation works just fine but in first frame

Code (CSharp):

1. Cylinder.rotation * Quaternion.Inverse(Box.rotation) != Box.rotation * _rotationOffset;

So Cylinder jumps. I attached very simple example with 5 lines of code. Any knows why it is not working?

 

It's working very well with the test scene. When i move or rotate the cube, the cylinder follows on relative distance and angle all the time.

 

Really you don't have this effect?

 

I do get that effect it seems... i have no explanation.

 

Its because Matrix Mulitpilcation is non communative
Rotation A * Rotation B != Rotation B * Rotation A

This should be flipped

Code (CSharp):

1. _rotationOffset = Cylinder.rotation * Quaternion.Inverse(Box.rotation);

Code (CSharp):

1. _rotationOffset = Quaternion.Inverse(Box.rotation) * Cylinder.rotation  ;

Or in the code you I downloaded from the first zip file change the start function to this:

Code (CSharp):

1.  private void Start()
2.     {
3.         _positionOffset = To.InverseTransformDirection(To.position - transform.position);
4.         _rotationOffset =  Quaternion.Inverse(To.rotation)* transform.rotation ;
5.     }

Your taking the Cube's current rotaion and rotating by your original rotation, then reversing Cube's original.. This is totally different than Take the Cube's rotation, reverse his original then rotate by your original. It seems like they are the same but it does end up in two different spots. The reason your getting that jitter is that its resetting the cylinder's original rotation so your first equation would work.

 

You'r absolutely right! Don't know how I could have missed that. Thumb up for you