On Wednesday 23 July 2003 20:27, Curtis L. Olson wrote: > David Megginson writes: > > It's in docs/Model/fgfs-model-howto.html. You have to check out the > > docs module of FlightGear to get it, though, since it's not part of > > the main source distro. When I have time, I'd like to get it into the > > Wiki. > > Ok, I have added a section to this document for the scale animation > type. (note, it really should be expanded since most of the animation > types didn't exist when it was originally written.) > > One possible work around for plib's non-support of the scaling > operation is to build the object at it's maximum size and always make > it smaller. > > Plib's view frustum culling scheme works by building bounding spheres > around each object. It isn't able to compute how a scale operation > will change the effective size of the object (tough problem to be > fair) and so it will do the culling based on the original object size. > > This means that if you scale you object larger and the space occupied > by the original object goes out of view, your entire object will > disappear (this is bad.) > > However if you start with a big object and scale it smaller, the > scaled version will always live inside the original bounding sphere, > so you will not have the same sort of drop outs ... only a possible > inefficiency since the object could be drawn even when it's smaller > version is entirely outside the view frustum (because the originally > sized object could still be inside the view frustum.) > > So a work around at this point would be to build your object at it's > maximum size and then always scale it smaller. > > What do you think Lee, can you work with this? > > Regards, > > Curt. > -- > Curtis Olson IVLab / HumanFIRST Program FlightGear Project > Twin Cities curt 'at' me.umn.edu curt 'at' flightgear.org > Minnesota http://www.menet.umn.edu/~curt http://www.flightgear.org
I think it should work ok for the landing light scenario where you're not doing aerobatics and you're over fairly flat ground. I would've thought that a smaller bounding sphere would be better than a big bounding sphere for this - would there not be a greater chance of drop outs with a big bounding sphere due to the greater volume increasing the likely hood of intersecting something? I might be misunderstanding this though. Another posibility to try is to increase the poly count and design of the object. With the shadow I just used a simple rectangle (after I triangulated it). For a landing light, perhaps a polar type mesh might be better. Too much speculation and not enough trying out:) LeeE _______________________________________________ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
