Hi Daniel,
You are hitting up against small feature culling. If for you
application its inappropriate just switch it off. See the FAQ for
details.
Robert.
On 6/7/06, Daniel Trstenjak <[EMAIL PROTECTED]> wrote:
Hi all,
if I have a osg::Geometry with one point (GL_POINT) at 0,0,0 than
I get a bounding box with min: 0,0,0 and max: 0,0,0.
The bounding sphere of the osg::Geode, which only contains
this geometry, is center: 0,0,0 and radius: 0.
The first thing which irritates me are the valid methods
of osg::BoudingBox and osg::BoudingSphere.
bool BoundingBox::valid() const
{
return _max.x()>=_min.x() && _max.y()>=_min.y() && _max.z()>=_min.z();
}
bool BoundingSphere::valid() const { return _radius>=0.0f; }
What irritates me are the "=". So the above computed bounding
volumes are both valid.
So what happens during culling of a osg::Geode with such
a bounding sphere.
The first method called in osgUtil::CullVisitor::apply(osg::Geode&)
is osg::CullStack::isCulled(osg::Node&) which calls
osg::CullingSet::isCulled(osg::BoundingSphere&):
bool CullingSet::isCulled(const BoundingSphere& bs)
{
if (_mask&VIEW_FRUSTUM_CULLING)
{
// is it outside the view frustum...
if (!_frustum.contains(bs)) return true;
}
if (_mask&SMALL_FEATURE_CULLING)
{
if
(((bs.center()*_pixelSizeVector)*_smallFeatureCullingPixelSize)>bs.radius())
return true;
}
#ifdef COMPILE_WITH_SHADOW_OCCLUSION_CULLING
if (_mask&SHADOW_OCCLUSION_CULLING)
{
// is it in one of the shadow occluder volumes.
if (!_occluderList.empty())
{
for(OccluderList::iterator itr=_occluderList.begin();
itr!=_occluderList.end();
++itr)
{
if (itr->contains(bs)) return true;
}
}
}
#endif
return false;
}
In the case of the bounding sphere (center: 0,0,0 radius: 0)
_frustum.contains(bs) returns true, so the culling of this
"valid" bounding sphere depends only on SMALL_FEATURE_CULLING,
which will return false.
Is this complete behaviour accurate ?
Shouldn't there be a special case in osg::Drawable::computeBound
for a one point geometry ?
Daniel
_______________________________________________
osg-users mailing list
[email protected]
http://openscenegraph.net/mailman/listinfo/osg-users
http://www.openscenegraph.org/
_______________________________________________
osg-users mailing list
[email protected]
http://openscenegraph.net/mailman/listinfo/osg-users
http://www.openscenegraph.org/
