[Bf-funboard] Eular rotation vs Quarternion rotation

[Janek Kozicki]

Martin Poirier said:     (by the date of Mon, 21 Jan 2008 16:39:16 -0800 (PST))

> What I assume will be a problem is converting the
> interpolation of those values in a way that they
> remain editable euler angles without loosing the
> properties of quaternion interpolation.

hmm, I don't think it can get any better than:

If rotation calculated through quaternion interpolation
gets edited it becomes a new key on the path. However:

1. if it gets edited but not rotated at all - so it becomes a key
but has exactly the same rotation - whole rotation path must look
exactly the same

2. It must be possible to "un-key" this rotation: forget the
user-edited rotation so it once again becomes just a value
interpolated from its neighbours.


And by this way the user is doing rotations on Euler angle, while
internally the same rotation is represented as quaternion for the
interpolation purposes.


> > You know - internally blender still uses quaternions
> > (it's really not
> > possible to work with anything else internally).
> 
> Sorry to disappoint but quaternions are only used for
> bones, Objects' rotations are done in euler angles as
> this is "usually" more intuitive to edit and for
> simple rotations yield acceptable results.

heh, I wouldn't guess that. But I see the point :)

-- 
Janek Kozicki                                                         |