colorcurvature opened this issue on Jun 16, 2014 · 4 posts
colorcurvature posted Mon, 16 June 2014 at 2:38 AM
Hi, I've somehow broken a basic technical property on .cr2 level, maybe some of you know how to repair that. I've managed that actor X and its direct parent actor Y rotate around different center points when X is using a bend (which of course means distortion) . Appearently, in the .cr2 in Y's section there is a descriptor that describes Y's reaction with regards to X's bends, and this section also has a "center" directive.
Why are there two center points for the same bend, this is really odd :(.
I somehow managed to get these out-of-sync when trying to animate a joint with the teaching tool. I was manually able to fix that by aligning the "center" in the .cr2, so if I do not animate I get the two actors back in sync. However, once again trying to animate the joint the two centers will once again split.
This is an unsual reaction and I think I've another problem on .cr2 level but I do not understand how to fix it.
Does anybody know about these redundant "center" directives and how they are technically kept in sync during joint animation on .cr2 level?
thanks :)
ockham posted Mon, 16 June 2014 at 5:12 AM
In those TwistY and JointX paragraphs, the Center and StartPt and EndPt are referring to the zone of deformation, the area that bends when you move the joint in those directions.
They aren't the center of the joint itself. These deformation centers aren't meant to be in sync. I believe they start out identical by default, but they can easily be different after twiddling with joint zones.
You wouldn't always need them to be different, but for something like a Shoulder joint, the area of deformation will generally have different centers for the different rotations.
PhilC posted Mon, 16 June 2014 at 5:23 AM
Just so that we can be completely clear would you be able to copy/paste an example of the two paragraphs in the CR2 file that you feel are causing the problem?
colorcurvature posted Mon, 16 June 2014 at 5:41 AM
In my actual "problem case" I was using only weightmaps, not zones.
Don't the connected actors have to rotate around the same reference point? Otherwise the bend transformation cannot be steady?
If the "center" in the twistX is not referring to the joint center - how can weightmapped actors rotate around different centers, how can that be possible :(
I had the connected actors entirely weightmapped to 1.0, and the bend ripped them apart. I observed one was rotating around the morphed joint center, the other was rotating around the unmorphed joint center.