I'm attempting to implement a multiple pass rendering algorithm in OSG, but I'm not sure what the best way is. I am just learning how to do this and how to use Frame Buffer Objects. I initially tried to implement an osgFX::Effect, since it supports multiple passes transparently. The problem is how to pass the data between the passes.
The algorithm is to render the model to a texture, then process that texture over 3 passes and composite it back into the scene. I used the osgprerender example to do the first part using osg::Camera::FRAME_BUFFER_OBJECT without problems. I then tried to use osg::FrameBufferObject to pass the data from the first processing pass to the second, but it doesn't work. It looks, from osgUtil::RenderStage's Camera support, like osg::FrameBufferObject can't be used simply by adding it to a StateSet, but it needs glCheckFrameBufferStatusEXT after the apply(). Is this a limitation of osg::FrameBufferObject, or am I missing something? My current attempt is to use nested osg::Cameras with osg::Camera::POST_RENDER for each pass, like so (hoping the ASCII art turns out...): Root |->Camera(PRE_RENDER) ||->Model |->Camera(POST_RENDER) ||->Pass 1 quad ||->Camera(POST_RENDER) |||->Pass 2 quad |||->Camera(POST_RENDER) ||||->Pass 3 quad ||||->Camera(POST_RENDER) |||||->Final pass quad At each pass, the texture target from the Camera's FBO is used to texture the quad for the next pass (with shaders attached to the next pass's StateSet to do the processing). Is this a recommended approach? Are there any requirements on the Cameras (Transform values, ABSOLUTE_RF, etc.) I should be configuring? Thanks, -- Bryan Thrall FlightSafety International [EMAIL PROTECTED] _______________________________________________ osg-users mailing list [email protected] http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
