iTaSC solver mixes up parent relationship #29041

New Issue

Tobias Oelgarte · 2011-10-26T18:13:46+02:00

Tobias Oelgarte commented

2011-10-26 18:13:46 +02:00

%%%The iTaSC solver mixes up the parent relationship while it is in posemode. Even so the bone has its correct parent in edit-mode iTaSC thinks that it is the child of another bone. This makes the bone-chain act as if it really was parented to the other bone. But since this isn't true it also misses it's goal by the offset of the "right" and "wrong" parent.

I assume that the structure that is passed to the solver is build incorrectly, while the solver does it's job depending on the wrong relationship input.

I appended an example file in which this error occurs. I marked the bones with colors. The red bone is the "wrong parent", the blue bone is the "right parent" and the green bone is the child that uses the "wrong parent" as a parent for IK in pose mode. If you switch between edit- and pose-mode you should clearly see that it suddenly changes it's parent, while the rest of the chain is missing it's goal by the offset between the right and wrong parent.

This bug occurred to me under Ubuntu 10.10. with the revisions r39294 and r41275 and makes it impossible to use iTaSC in this case, while the legacy solver can't handle the situation at all (very wrong looking results).%%%

%%%The iTaSC solver mixes up the parent relationship while it is in posemode. Even so the bone has its correct parent in edit-mode iTaSC thinks that it is the child of another bone. This makes the bone-chain act as if it really was parented to the other bone. But since this isn't true it also misses it's goal by the offset of the "right" and "wrong" parent. I assume that the structure that is passed to the solver is build incorrectly, while the solver does it's job depending on the wrong relationship input. I appended an example file in which this error occurs. I marked the bones with colors. The red bone is the "wrong parent", the blue bone is the "right parent" and the green bone is the child that uses the "wrong parent" as a parent for IK in pose mode. If you switch between edit- and pose-mode you should clearly see that it suddenly changes it's parent, while the rest of the chain is missing it's goal by the offset between the right and wrong parent. This bug occurred to me under Ubuntu 10.10. with the revisions r39294 and r41275 and makes it impossible to use iTaSC in this case, while the legacy solver can't handle the situation at all (very wrong looking results).%%%

Tobias Oelgarte commented

2011-10-26 18:13:46 +02:00

Changed status to: 'Open'

kanttori commented

2011-10-30 19:03:18 +01:00

%%%Hi,

At least in this case it seems depth in tree is used for parent index which omits branching. I'm attaching a patch that does an array search of the tree from parent's depth index onwards (the "default" case) to the end to find a match.%%%

%%%Hi, At least in this case it seems depth in tree is used for parent index which omits branching. I'm attaching a patch that does an array search of the tree from parent's depth index onwards (the "default" case) to the end to find a match.%%%

Brecht Van Lommel commented

2011-11-13 14:08:55 +01:00

%%%Fix in svn, thanks for the patch. Also fixed legacy solver.%%%

Brecht Van Lommel commented

2011-11-13 14:08:55 +01:00

Changed status from 'Open' to: 'Resolved'

Brecht Van Lommel closed this issue

2011-11-13 14:08:55 +01:00

Sign in to join this conversation.

No Label

Download

What's New

Blender Studio

Manual

Developers Blog

Documentation

Benchmark

Blender Conference

Development Fund

One-time Donations

iTaSC solver mixes up parent relationship #29041