Hi Eric, >From your description of the problem is does sound like a depth buffer issue, however, debugging this remotely by me implementing somehting then you testing will be very inefficient - especially as you are on a different time zone.
Could you possible enable the same options in one of the OSG examples like osghud and tweak the example to load a model so that it minimics what you application does to see if you can reproduce the problem with it. Cheers, Robert. On Tue, Jun 14, 2011 at 1:52 PM, Eric Sokolowsky <[email protected]> wrote: > It will be hard for me to provide an example, but the behavior I get > (sometimes) is that the text does not appear at all perhaps because of depth > buffer. It is probably true that these problems would go away if I reworked > my program but it was easier just to select the "Stencil Buffer" method for > drawing the backdrop, and everything worked just fine. I'm using osgText to > draw text on top of my scene as a HUD, not drawing 3D text in the scene. > > That said I can certainly test what you implement. > > On Tue, Jun 14, 2011 at 5:26 AM, Robert Osfield <[email protected]> > wrote: >> >> Hi Eric, >> >> On Fri, Jun 10, 2011 at 2:49 PM, Eric Sokolowsky <[email protected]> >> wrote: >> > One last thought about the existing backdrop implementations: I have use >> > for >> > them. I was getting some artifacts using the now-default >> > DelayedDepthWrite >> > method, and reverting to the Stencil Buffer method fixed the problem. >> > Please >> > keep the alternatives around. >> >> Could you provide an example that illustrates the problems you see >> with the current default implementation? >> >> osgText::Text implementation is one area I'm planning to look at this >> week in prep for 3.0 feature freeze. I'd like to simplify the code >> this week. As part of this work I'm considering changing the backdrop >> implementation to improve the ability to handle overlapping >> writes - either by making largee quads and using texture matrix to >> create the offsets, or dividing the quad geometry up so that >> the various passes share vertices to prevent the quads end up with >> different depth values during write. >> >> Robert. >> _______________________________________________ >> osg-submissions mailing list >> [email protected] >> >> http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org > > > _______________________________________________ > osg-submissions mailing list > [email protected] > http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org > > _______________________________________________ osg-submissions mailing list [email protected] http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org
