Hi Ramy, I've done a little work on osgShadow since 2.8.3, so svn/trunk LispSM is now a little more stable for more view points.
More significant is that I've experimented with using viewer level depth partitioning to divide the scene into two halves - this improves the quality and the stability of the shadow significantly. Using depth partitioning is similar to ParallelSplitShadowMap (PSSM), but adds improvements in depth precision to the table as well. The depth partitioning is implementated via viewer slave cameras, and can work with all the current shadow techniques, so you can happily combinat LispSM with depth partitiioning, from my testing this results in the best results, but still not perfect. Try out svn/trunk, and try adding --dp to the osgsadhow command line to look at the results, i.e. osgshadow --dp --lispsm yourmodel.ive If you have a look at the osgshadow.cpp you'll see how to call osgViewer::View to create the required slave cameras to do the depth partitioning. Robert. _______________________________________________ osg-users mailing list [email protected] http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
