Maybe this might help provide some insight: https://www.blender.org/forum/viewtopic.php?t=26417
On Sun, Mar 25, 2018 at 7:05 AM, Recep Aslantas <[email protected]> wrote: > Thanks Gaia for reply, > > For testing angles; > > In Blender I'm adding rotations to a cube with specific euler order e.g. > XYZ, XZY... then, > I'm exporting it as COLLADA file to see matrix produced with this euler > angles > > Then I'm testing it with cglm library to produce same matrix with same > euler angles like this: > > mat4 rot; > glm_euler_xzy((vec3){Xangle, Yangle, Zangle}, rot); > glm_mat4_print(rot, stderr); > > this produces same matrix as Blender. > > The problem is that I was thought that I was implemented intrinsic (pitch, > yaw and roll) rotations > but now I realized that they are extrinsic (elevation, heading and bank) > rotations > which are reverse order of intrinsic (intrinsic XZY == extrinsic YZX) > > So before fix that but I wanted to know what convention > does Blender use, extrinsic or intrinsic rotations? > Because after changed the orders, the result will not be same :/ > > Recently glm added new euler functions: https://github.com/g-truc/glm/ > pull/744 > My guess is that glm::eulerAngleYZX will not produce same rotation matrix > as Blender > > FWIW, I'm using euler angles in AssetKit ( https://github.com/recp/ > assetkit ) which is COLLADA/glTF importer > to convert coordinate systems, for instance Z_UP to Y_UP ... > > Thanks > > - Recep > > _______________________________________________ Bf-committers mailing list [email protected] https://lists.blender.org/mailman/listinfo/bf-committers
