i''m constructing a view path editor which uses piecewise cubic bezier curves for the positional curve and the quaternion squad derivative scheme for the orientation curve. now, when choosing the control points of the bezier curve to ensure C1 continuity of the positional curve, you need to adjust the lengths of the control points if the keyframe time intervals aren''t uniform. this is discussed on page 491 of real time rendering, 2nd edition. my question is, do i need to do something similar for the orientation curve control points? when the t values between the orientation keyframes (which are just the curve arc lengths) are non-uniform, i''m noticing orientation speed discontinuities when traversing the view path. i have directional (G1) continuity, but not C1 continuity. this is the exact problem that arises with non-uniform time intervals in the positional curve, and is why you need to adjust the lengths of the control points. i think that i need to somehow alter the "length" of the orientation to take into account the non-uniform t intervals, but i don''t understand the equivalent concept for quaternions of vector length. unfortunately, none of my books mention this problem. surely someone has implemented a view pathing system and encountered this problem, or perhaps you can tell me why i don''t need to worry about adjusting the orientations at all. i also posted this over at flipcode, but the forums there seem pretty dead. any help is appreciated. thanks. -steve