mirror of
https://github.com/overte-org/overte.git
synced 2025-04-20 04:24:07 +02:00
Fix for incorrect avatar bounds after changing avatars
This commit is contained in:
parent
868cf83b48
commit
97a2eb62d4
1 changed files with 9 additions and 12 deletions
|
@ -1027,7 +1027,6 @@ void Rig::buildAbsoluteRigPoses(const AnimPoseVec& relativePoses, AnimPoseVec& a
|
|||
|
||||
// transform all absolute poses into rig space.
|
||||
AnimPose geometryToRigTransform(_geometryToRigTransform);
|
||||
geometryToRigTransform = _modelOffset * _geometryOffset;
|
||||
for (int i = 0; i < (int)_absolutePoses.size(); i++) {
|
||||
absolutePosesOut[i] = geometryToRigTransform * absolutePosesOut[i];
|
||||
}
|
||||
|
@ -1134,6 +1133,12 @@ void Rig::computeAvatarBoundingCapsule(
|
|||
// compute bounding box that encloses all points
|
||||
Extents totalExtents;
|
||||
totalExtents.reset();
|
||||
|
||||
// HACK by convention our Avatars are always modeled such that y=0 is the ground plane.
|
||||
// add the zero point so that our avatars will always have bounding volumes that are flush with the ground
|
||||
// even if they do not have legs (default robot)
|
||||
totalExtents.addPoint(glm::vec3(0.0f));
|
||||
|
||||
int numPoses = finalPoses.size();
|
||||
for (int i = 0; i < numPoses; i++) {
|
||||
const FBXJointShapeInfo& shapeInfo = geometry.joints.at(i).shapeInfo;
|
||||
|
@ -1142,25 +1147,17 @@ void Rig::computeAvatarBoundingCapsule(
|
|||
for (int j = 0; j < shapeInfo.points.size(); ++j) {
|
||||
totalExtents.addPoint((pose * shapeInfo.points[j]));
|
||||
}
|
||||
} else {
|
||||
totalExtents.addPoint(pose.trans);
|
||||
}
|
||||
}
|
||||
// HACK so that default legless robot doesn't knuckle-drag
|
||||
if (rightFootIndex > -1) {
|
||||
totalExtents.addPoint(finalPoses[rightFootIndex].trans);
|
||||
}
|
||||
if (leftFootIndex > -1) {
|
||||
totalExtents.addPoint(finalPoses[leftFootIndex].trans);
|
||||
}
|
||||
|
||||
// compute bounding shape parameters
|
||||
// NOTE: we assume that the longest side of totalExtents is the yAxis...
|
||||
glm::vec3 diagonal = geometryToRig * (totalExtents.maximum - totalExtents.minimum);
|
||||
glm::vec3 diagonal = (geometryToRig * totalExtents.maximum) - (geometryToRig * totalExtents.minimum);
|
||||
// ... and assume the radiusOut is half the RMS of the X and Z sides:
|
||||
radiusOut = 0.5f * sqrtf(0.5f * (diagonal.x * diagonal.x + diagonal.z * diagonal.z));
|
||||
heightOut = diagonal.y - 2.0f * radiusOut;
|
||||
|
||||
glm::vec3 rootPosition = finalPoses[geometry.rootJointIndex].trans;
|
||||
localOffsetOut = geometryToRig * (0.5f * (totalExtents.maximum + totalExtents.minimum) - rootPosition);
|
||||
glm::vec3 rigCenter = (geometryToRig * (0.5f * (totalExtents.maximum + totalExtents.minimum)));
|
||||
localOffsetOut = rigCenter - (geometryToRig * rootPosition);
|
||||
}
|
||||
|
|
Loading…
Reference in a new issue