|
From: Norman V. <nh...@ca...> - 2001-06-14 19:14:28
|
Curtis L. Olson writes: >David Megginson writes: >> Norman Vine writes: >> >> > The only other 'natural' way I can think of to not have the Panel >> > block most of the scenery with out a major rewrite of the FGFS SSG >> > display pipeline is introduce a 'minor' deflection downwards into the >> > view vector when the Panel is displayed. >> > >> > This, IMHO, would be perceived as natural in that when looking at the >> > Panel in real life your look vector is probably slightly downwards. >> >> Perhaps the deflection angle is what we should make adjustable. > >I used to understand this code since I originally wrote it, but it's >been optimized quite a bit based on certain assumptions so I'm not >sure how easy it would be to defelect the view forward direction down >by some value. Hey .. All the older code is still there albeit inside of #ifdefs. I think that all of my 'optimizations' are done on a 'matrix by matrix' way so that the overall logic is unchanged. :-) >Conceptually, you would take the view_forward vector and rotate about >the pitch axis. Yep just adding the appropriate rotation matrix into the pipeline that creates VIEW should do the job. >It would be nice to have a less brittle view frustum setup >approach... I think that we should take a good look at this AFTER the release as part of a general overhaul of the viewing pipeline. i.e off the top of my head we will shortly need 1) fixed camera locations < tower view > 2) better external view 3) possibly multiple aircraft, each with their own view 4) support for 3D Panels 5) .... Note that SSG has some newer features to help with this that we should be able to finally use :-)) Cheers Norman |