I am using OccluderNode:
for each "building" in config file{
auto &pnt_pair = vecs[i];
osg::OccluderNode* onode = new osg::OccluderNode();
osg::ConvexPlanarOccluder* occl = new osg::ConvexPlanarOccluder();
osg::ConvexPlanarPolygon& polly = occl->getOccluder();
vector<osg::Vec3> pnts;
... get pnts from current building in custom config file
... most often 4 points
for (unsigned int j =0; j < pnts.size(); j++){
polly.add(pnts[j]);
}
onode->setOccluder(occl);
m_group->addChild(onode);
}
The performance without the occluders is bad. Basically the rendering of
buildings is not bad, its pretty quick even without the occluders. Where we get
in trouble is with the parked cars, pedestrians and various scene clutter.
These are relatively small objects, and lots of them have transparency.
We want to add some form of OIT which should help us to avoid having some many
depth sorted objects, but we are not there yet.
------------------
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=73903#73903
_______________________________________________
osg-users mailing list
[email protected]
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
