Dave Griffiths wrote:
On Tue, 2010-01-26 at 15:54 +0100, Rob Bothof wrote:
I've been reading on quaternions but this is still beyond my
understanding of 3d math.
I would imagine that the flipping you describe with maim is gimbal
lock[1] occurring. The proper way to avoid these problems is to use
quaternions, but there are often quick hacky ways to solve it.
Re: "how to get the local rotation of a primitive as a vector??" - the
problem is that a 3D vector as a direction doesn't include information
about any twists around that vector, so it's not enough to describe an
orientation. So sometimes you'll get extra twisting/flips that you
don't want, (a result of gimbal lock). Quaternions include the extra
twist information (as far as I understand it), and you can convert a
quaternion to a rotation matrix fairly easily (though using the
quaternion directly maybe more computationally efficient).
Using quaternions isn't too hard (though understanding them might be, I
get very confused). There is one little algorithm I found the other day
which might apply here, to calculate the quaternion that takes one 3D
direction onto another 3D direction along the shortest arc between them:
http://www.gamedev.net/community/forums/topic.asp?topic_id=429507&whichpage=1�
Interpolating between quaternions (aka slerp) is a bit tricky, but you
only have to write the code once (or once per language...). I have some
code I use for this (but not in Fluxus) which is essentially just an
implementation in C of the equations from this:
http://www.lce.hut.fi/~ssarkka/pub/quat.pdf
It would be good to have a library of quaternion functions and
conversions to/from various vector and matrix representations available
in Fluxus, if there isn't already one.
Claude
--
http://claudiusmaximus.goto10.org
