"flipping" rotation problem

[MrGAG]

Hello, I am having a problem with some flipping rotations when I copy geometry on some deforming points.

No matter where the normal vector is pointing, it seems I can't avoid the problem.

I attached my hip file. hoping someone can point me in the right direction

lsystem_test_01.hipnc

[sanostol]

does this work for you?

lsystem_test_02.hipnc

Edited June 11, 2013 by sanostol

[MrGAG]

Hi, thank you for answering.

it seems to be a little better, but can still see some strange rotations on severals circles. And I must admit I don't fully understand what you did with these vops. I never played with quaternions. They scare me a little !

But I really would like to find a solution to this recuring problem I have. and quaternions might be the right answer, as scary as they may be ^^

thank again

[sanostol]

if you want non flipping rotations quaternions are a good way, they might be complex in theory, but in cg they are most of the times quite easy to use

i just get the orient quaternion attribute from the wiresim, ( you even get the twist) and transfer it to the leaves with a id, so every leave knows where it came from. now You convert You original orientation represented by up and N to a quaternion and multiply them

quaternion multiplication means adding rotations and write it back. the copy picks up the orient attribute

EDIT: try the new file, the normal got updated by the deform and made the rotation going wrong, by storing the orientation as quaternion in the beginning it works better now

lsystem_test_03.hipnc

Edited June 11, 2013 by sanostol

[MrGAG]

Hi,

Yes indeed, it works better now . I am still a little confused by quaternions, but I guess I'll have to play around with them to get more confortable with their use.

Anyway, thanks a LOT for your help