Hi Daniel,
Well, the MRT is the performance part, without it, multiple render passes have to occur, and deferred shading wouldn't be possible for highend games.
The original question was about post-render RTT passes, not deferred shading. Obviously without MRT, deferred shading would not be feasible at acceptable speeds. But that wasn't the question - Jeremy was asking how to make post-render RTT passes fast, and the answer is that they are fast as long as the shader you use is fast.
What is costly is RTT passes through your whole scene. For that, MRT is useful for doing deferred shading because you can do a single render pass of your scene and render to multiple textures to fill them with data (G-Buffers). But fullscreen RTT passes onto a fullscreen quad can be used in other contexts. For example, for light bloom, you'll generally do one RTT pass of the whole scene (no MRT needed), and then multiple fullscreen blur and combine passes, which are very cheap.
You can of course do many such effects easily once you have a deferred shading pipeline in place, but it is not required, which is why I said they are independent (but linked) concepts.
J-S -- ______________________________________________________ Jean-Sébastien Guay [email protected] http://www.cm-labs.com/ http://whitestar02.dyndns-web.com/ _______________________________________________ osg-users mailing list [email protected] http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
