Hi all, I have a problem in which I have to do multipass rendering. When I say multipass, it's in fact rendering to a texture which is the input for the next rendering, etc... and that about a hundreds time... I could explain why, but only if necessary, but let's say that no there is no better way than doing a cascaded rendering with that amount of steps, and in fact, that's sphecifically why I'm doing it on GPU...
Moving on to OSG related problem, it seems to me that every time you want to render to a texture you have to add a Camera node. I've heard that you could maybe do a little bit better, by having only one Camera node, but several paths leading to it with differents statesets. But in my case, I can't have only one camera, and several statesets, because each time, I have to change the fbo to render to, that can only be done by adding a new camera, with a different texture to render to ... So my question is, is there a better way to do this kind of multipass rendering, than adding a camera for each step ? (better in the sense of less costly ...) I'm asking this because when I look at the stats, drawing takes about 60 ms, and gpu time is 80ms. I expected the value for the gpu, because for each camera, there is a fragment shader associated. But I didn't expect the draw time to be so high, and I really think that is because of the way I'm doing it (too many nodes maybe ... :( ) Thank you all _______________________________________________ osg-users mailing list [email protected] http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
