Dear all, I have been trying to apply osgPPU for stereo rendering and that appeared to work. However, when I took a closer look at the anaglyph rendering of the osgppu_dof example (run it with --stereo), I noticed that it does not work as expected (running the osgPPU pipeline on both views, then applying anaglyph), even though it kind of looks plausible. What happens instead is that the "left" render pass executes the pipeline on a red cow (osg uses colormasks for rendering the anaglyph). In the second pass, osg additionally renders the cyan parts into the same input texture and runs the "right" pipeline pass based on it. For the second frame, the input texture is still the previous anaglyph with updated red-channel. Both passes are using the correct depth buffer for their view.
This problem is bugging me, as the same stereo pipeline works correctly with quad-buffered stereo. I think the fix would be to render both left and right views in full colour and then combine them in a shader to compute the final anaglyph image. This way it would also be possible to support other types of anaglyph, such as red-blue, green-magenta [1] or blue-amber [2], which may not be easily implemented using RGB colour masks. My question is: would this be a useful modification of osg? What do other people think? Thanks, Christian. [1] <http://www.3dglassesonline.com/trioscopics/> [2] <http://www.colorcode3d.com/Home.html> _______________________________________________ osg-users mailing list [email protected] http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
