Hi Curt, I am one week late but here are the pics that I hope will help you to help me ;-)
http://paxettepaxou.free.fr/fg/horizontal_fov.JPG [17 Ko] http://paxettepaxou.free.fr/fg/vertical_fov.JPG [19 Ko] http://paxettepaxou.free.fr/fg/horizontal_fov_front.JPG [17 Ko] http://paxettepaxou.free.fr/fg/vertical_view_front.JPG [14 Ko] http://paxettepaxou.free.fr/fg/horizontal_fov_side.JPG [19 Ko] http://paxettepaxou.free.fr/fg/vertical_fov_side.JPG [15 Ko] As I told you in my last mail, I have a cylindrical projection and the point of view is not centered in the cylinder. I have three asymetric frustums (vertically and horizontally asymetric). I really hope it will help you to help me. Thanks David > That's the basic idea. You specify a larger "virtual" screen that has a > symmetric frustum and then each display get's assigned a portion of the > larger display. I realize this is probably not an industry standard way > to do it, and this approach probably can't cover every possible screen > configuration, but I needed something quick a few months ago, and this > approach meshed pretty well with the existing code. > > I shold point out that there is some subtle wierdness depending on the > size of your display so for example, let's say you have 5 forward > displays. The center 3 are all parallel so they need to act as a single > larger fov display. If you want to assign 30 degrees field of view to > each of them, you would naturally pick a 90 degree field of view for the > center 3 and give each display 1/3 of that. However, you can't just > give the 2 edge displays an symmetric 30 degrees and have them match > up. There is some subtle aspect ratio stuff going on there that causes > problems. So what you'd need to do is setup the side channels as a 90 > degree fov virtual display and give them 1/3rd of that, then they should > match up with the center channels. > > At some point it would probably make sense to clean up the view pipeline > to handle this better, but time and priorities are always the big problem. > > Regards, > > Curt. > _______________________________________________ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
