Hi Robert,
2013/4/3 Robert Osfield <robert.osfi...@gmail.com> > > > There new approach should be able to encompass novel displays that require > multiple camera and then a second pass to render the result. In essence it > won't be any different to how stereo will be done - a master Camera on the > view that controls the main view and projection matrices, then slave > Camera's to render the left/right or n views with the appropriate view and > projection offsets, then if required a final slave Camera that does the > composition. > > The question for me is how best to wire up the interface for the stereo > support and any other schemes. One approach is to leave it low level > viewer configuration where the user or a configuration file describe > exactly what slave Camera are required, this will require just what we have > right now in terms of the osgViewer API and implementation, the only > addition would be a fully functioning viewer configuration file. Providing > easier to use helper functions in osgViewer::View in addition to the viewer > configuration file is what I'd like to do so the user can use either route > or a combination of these - prior art to these helper functions are the > View::setUpView*() methods. > > Yes that is the idea, but I think a simple offset is not enough, as depending on the scene and/or the fusionDistance the "offset" should be a callback which computes the correct displacement for each View. If I recall correctly SceneView has a StereoMatricesCallback to achieve this computations. Maybe exposing this callbacks into the addSlave API will allow better stereo support. Maybe for a "simple" configuration file one can use a default stereo settings and set up default callbacks or a simple offset depending on what do you want to achieve, and of course the grid of cameras configuration, for instance, 2 columns, 1 row is horizontal split, 3 columns 3 rows is a multiview, but in the end it behaves the same. Then if you want to add a second, third or whatever passes for getting mesh distorsion and/or even a special interleave perhaps the configuration file could only have a "filter" per pass and a set of params for the filter. Not sure if these things fits on your idea. > BTW, if you can create the appropriate viewer slave Camera and shader > combination then I'd be able to wrap this up in a helper function. > Yes, I will do that, but I never did it because it was not possible to change the "offsets" per frame, maybe I could prepare the code with a simple offset and then improve it when something more configurable is possible. Thoughts? Rafa. > > > Robert. > > _______________________________________________ > osg-users mailing list > osg-users@lists.openscenegraph.org > http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org > > -- Rafael Gaitán Linares CTO at Mirage Technologies S.L - http://www.mirage-tech.com gvSIG3D Developer - http://gvsig3d.blogspot.com
_______________________________________________ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
