Re: Reducing airports polygon count

Dale E. Edmons wrote:

> I don't consider it insane at all, more like terribly inconvienent due 
> to the horrible things I'm trying to make it do--display approx 200 
> polygons per display per frame!  I've actually found a way to exceed 
> that limit for the general terrain.  I'll be better able to judge once 
> I get my Terragear version debugged.  I'm judging based on FGSD output 
> right now.  Though very similar, they aren't exactly the same.


Do you have any screen shots of the sort of output quality/complexity 
that your system can currently achieve with it's original databases?  
... general terrain ... on approach to a runway/airport, etc.?

> Two would be nearly equivalent to  modeling the  runway flat.  I'm 
> just trying to avoid what we'll actually end up doing.  Use the 
> TerraGear version as a rough-in and re-model each airport by hand.   
> The RAS system coming up is going to tax us and the equipment to the 
> very limit (hopefully not beyong) of what is required in terms of 
> accurately modeled airports.  RAS, as I understand it, is to be tied 
> into the EGPWS.  The  EGPWS gives terrain elevation map on the radar 
> display.  The RAS will be kind of like one of the city street maps 
> tied to a GPS in a car.  RAS will be for airports and taxiways.


Remind me what RAS and EGPWS stand for?


> That's why I'm using FlightGear.  Quality is the important part!  Once 
> I understand the code better and finish debugging my current stuff, 
> I'll take a serious look at this.  Especially if a user selectable 
> parameter could be added to affect the desired end count similar to 
> what can be done for the terrain, that would be a very good solution.
>
> 1) Certification.  I believe that FGFS as been through certification 
> on "Training Equipment" or maybe even for FTD's, but I don't think its 
> been certified for a Level C/D Full Flight simulator--even a 
> grandfathered one.  On my system at home I notice that FGFS display 
> will stop at certain intervals and continue on.  Granted its only 
> momentary (maybe 1/4 sec or less) but noticable.  When we put an 
> O-scope on the signal and do the step response tests as required by 
> the FAA, I reasonably expect FGFS would fail--if only intermittantly.


What frame rate do you need to sustain?  You can throttle FG back to 
30hz or 20hz with a command line option ... it's possible to have 
hardware that is a bit iffy for sustaining 60hz, but will solidly hold 
30hz.  Also, you can throw additional hardware at the system.  A scsi 
disk subsystem will improve load performance over ide because scsi does 
direct dma transfers to memory, where ide reads need to involve the cpu 
rather heavily.  There's no difference when you load your MS word 
document, but there can be a big difference if you are trying to load 
scenery in a threaded scenery loader and render the display at the same 
time.

For Level3 FTD they don't take a close look at the display rate so we 
got off pretty easily there ourselves.

> 2) Hooking a PC to a caligraphic projector is non-trivial.  Converting 
> the projection system to a raster only display is non-trivial.  There 
> are some vendors doing exactly that and we're looking into equivalent 
> FAA certified systems, but that brings up item 3).


Hey, that's an area I don't know too much about.  If you figure 
something out or can explain the issues I'd love to hear about it.

> 3) Cost.  If we can purchase an equivalent system with new IG then 
> I'll still be doing the same thing except that polygon count will no 
> longer be a major problem.  But, cost-cutting has become an absolute 
> imparative in the Airline industry in its current form.  Even then 
> we're only looking at replacing one of the two systems--unless things 
> change.  If management gives us the funds I'll be very happy indeed.


Yah, these high end display systems are not cheap ... especially when 
you get into systems that can warp themselves to project on a dome, and 
systems that do edge blending and such.  Well good luck with all that!  
You are definitely operating on a slightly higher plane.  I'd love to 
get there someday, and am happy to benefit from your experience. :-)

Regards,

Curt.

-- 
Curtis Olson        http://www.flightgear.org/~curt
HumanFIRST Program  http://www.humanfirst.umn.edu/
FlightGear Project  http://www.flightgear.org
Unique text:        2f585eeea02e2c79d7b1d8c4963bae2d


_______________________________________________
Terragear-devel mailing list
[email protected]
http://mail.flightgear.org/mailman/listinfo/terragear-devel
2f585eeea02e2c79d7b1d8c4963bae2d