Hi, Using many cameras is only a workround.
To get accurate rendering results you need a real distortion of the rendered image. Osgdistortion is a nice example. Basically you have two options: Distort you rtt pass on a per pixel basis with a shader, or apply the rtt result on a mesh and distort that. To get the correct settings how to distort you have again two options: A) modeling your project in an apropriate tool andvlet the tool calculate your camera setup, the distortion map and blending map B) create an interface that allows to drag image coordinates with the mouse to setup your distortion manually. To A): there is a ready to use solution provided by osgVisual. You do not need to use osgVisual, just use the sourcefiles of the distortion module. You can also download there a tool to setup your distortion, it is called projection designer. My experience: approach a) works technically great, but especially in multi channel setups it is useless: you can never setup your projectors and screen as you want, you have always a offset in angle and position, so the blending fails. So instead of generating the distortion by modeling the geometry, I'm currently working on approach B). I hope to finish it soon after my vacation and want to integrate it at least into osgVisual, but finally I will try to submitt it to OSG to enhance its distortion capabilities. Thank you! Cheers, Torben ------------------ Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=42011#42011 _______________________________________________ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
