[Bf-funboard] Euler for animator, Quaternions for Blender

Roland Hess rolandh at reed-witting.com
Thu May 10 13:55:40 CEST 2007


Glad the animation system is finally getting some love and attention!

My initial thoughts and research lead me to think that what we do is to 
store and maintain rotations as eulers in the Ipos, and convert to quats 
behind the scenes for interpolation. You're right that quat->eul is not 
discreet. Personally, I don't think this would be that hard to do.

One question, though -- is this desirable for all objects, or just 
bones? As you know, right now only bones use quats. One place to look in 
the code (and I guess that's more for me, than you guys :) ) is the 
joint limiting in the IK code. It uses Euler constraints on Quat 
rotations, so there almost certainly has to be some kind of process 
already in place there to deal with it.

One other question from a user standpoint -- if we did this and let the 
animator work with Euler curves and values, would you ever need direct 
access to the quats? I'm thinking not. Opinions?

Roland -- harkyman

  > in my recollection, animation master allowed selection between multiple
> curve types for display ( quats, eulers, axis angle maybe or spherical)
> and independently for interpolation.. so you could select euler
> editing/display and quaternion interpolation. Don't quiz me on how this
> works though- I don't think that quat to euler is one to one, so I've no
> idea how am or other apps handle it internally.
> 
>>         Following up a discussion in Comitters list:
>>         
>>         I also found myself struggling with Quaternions when
>>         retouching 
>>         animation and agree that having a more understandable
>>         interface for 
>>         
>>         animators (let's say, using Eulers instead) and let Blender to
>>         manage 
>>         internally Quaternions would be a great improvement in
>>         workflow.
>>


More information about the Bf-funboard mailing list