Hi Mukund, On Fri, Mar 18, 2011 at 6:23 PM, Mukund Keshav <osgfo...@tevs.eu> wrote: > Wow! interesting. Could you please share how the stereo of the system was? i > mean was the -ve parallax really good?
It worked well enough to be considering the limitations of the computer hardware and the display technlogy. Getting everything calibrated accurate, ghosting and low performnce when doing stereo were the biggest challenge. > Also, could you please suggest some resources on stereo? the ones i read had > pretty good explanation. I derived everything from first principles, and would recommend doing the same. Some of resources I've seen since can be rather misleading. > One thing what i couldn't understand though was, the relation between the > distance of separation between the left and right eye images > and the distance from the screen the object appears to be. > > Any resources on this would be really helpful. Since I did everything was first principles, doing all the maths for coordinate frames, the projection and view matrices myself,, and all the code was left behind at the where I used to work, soo... I don't really haven't anything apart from memories to pass on. The key is not to think about stereo so much, but to think about how to render an accurate view for each individual eye point relative to the display system, then work out the required projection and view matrices that each this. Then for stereo you just need to compute the positions using the same maths but with the appropriate left and right eye positions. If you are thinking too much about +/-ive parallax then you are probably not thinking at a low enough level. Don't think about steroe, think about one eye at a time, get one eye view perfected and stereo will just drop out. Robert. _______________________________________________ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
