Re: [Flightgear-devel] Reasonable vertex count for ground static?
On Tuesday 21 Dec 2004 07:34, Innis Cunningham wrote: The vertex count is down from near 1800 to 1200 and the number of 512x512 textures is halved. The static 747 and 737 at KSFO are 1100 and 400 vertices respectivley.So that may give you an idea how complex you think the 172 should be. Dave Martin Cheers Innis I'm going to have to scratch around and see what I can loose off it then :-) I tried loading 15 of them into Flightgear and getting them all in frame; I recieved about a 1/3 drop in framerate so the chances are that older cards would throw their toys out of the pram. Cheers for the info :) ___ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] Reasonable vertex count for ground static?
On Tuesday 21 Dec 2004 07:34, Innis Cunningham wrote: The static 747 and 737 at KSFO are 1100 and 400 vertices respectivley.So that may give you an idea how complex you think the 172 should be. Cheers Innis ___ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d I've now got 702 vertices and still a reasonable looking model that you can taxi past 20 feet away or fly over at 100 feet and think yes, thats a 172 ;-) Very close inspection gives the game away but your unlikely to find an aircraft that can taxi up to it without catching the prop. :-P One thing I've noticed is that lightplanes are definitely more complex in shape than commercial jets. The model now uses just one texture (which needs to be change to represent a different a/c to Fox Sierra.) Screenshot: http://www.cyfinity.com/fgfs/c172p-static.jpg I'm hoping to 'do over' the pa28 also so I can place some static aircraft at Midland airstrips - I hope to start on Wellesbourne Mountford (EGBW) but I need to get the taxi-ways / disused runways sorted with Taxidraw first. ___ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] Reasonable vertex count for ground static?
Dave Martin wrote: I've now got 702 vertices and still a reasonable looking model that you can taxi past 20 feet away or fly over at 100 feet and think yes, thats a 172 ;-) Very close inspection gives the game away but your unlikely to find an aircraft that can taxi up to it without catching the prop. :-P One thing I've noticed is that lightplanes are definitely more complex in shape than commercial jets. The model now uses just one texture (which needs to be change to represent a different a/c to Fox Sierra.) Screenshot: http://www.cyfinity.com/fgfs/c172p-static.jpg I'm hoping to 'do over' the pa28 also so I can place some static aircraft at Midland airstrips - I hope to start on Wellesbourne Mountford (EGBW) but I need to get the taxi-ways / disused runways sorted with Taxidraw first. Dave, Are you familiar with Level of Detail or LOD? This is a technique where you can build multiple versions of your aircraft with different level of details. The system then automatically picks which version to display depending on how far away it is. This let's you do a super high res version, one or two medium res versions, and a super low resolution version (which might just be a single dot.) It's more work to set this up, but if you carefully balance your polygon counts with your transition ranges, you can get a result such that you can display hundreds of models on the screen pretty easily. In most cases, 99% of the aircraft will be drawn in the ultra-low res LOD, and only the few closest will be drawn in the highest LOD. Regards, Curt. -- Curtis Olsonhttp://www.flightgear.org/~curt HumanFIRST Program http://www.humanfirst.umn.edu/ FlightGear Project http://www.flightgear.org Unique text:2f585eeea02e2c79d7b1d8c4963bae2d ___ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] Reasonable vertex count for ground static?
On Tuesday 21 Dec 2004 19:28, Curtis L. Olson wrote: Dave, Are you familiar with Level of Detail or LOD? This is a technique where you can build multiple versions of your aircraft with different level of details. The system then automatically picks which version to display depending on how far away it is. This let's you do a super high res version, one or two medium res versions, and a super low resolution version (which might just be a single dot.) It's more work to set this up, but if you carefully balance your polygon counts with your transition ranges, you can get a result such that you can display hundreds of models on the screen pretty easily. In most cases, 99% of the aircraft will be drawn in the ultra-low res LOD, and only the few closest will be drawn in the highest LOD. Regards, Curt. Now thats interesting stuff; I *had* noticed that FG has lod settings and such but I don't know much about how it works in FG (Although I fully understand the principal). I'll have a go at making another version of the 172 with say, 2d wings, solid windows and less polys in the fusealage. How many levels of LOD (no. of models) do you think would be worthwhile? (I suppose you can first sight individual a/c at 6 miles out.) Cheers. ___ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] Reasonable vertex count for ground static?
Dave, you can check out the animation file (specifically: engine animation) for the MD11 if you want to learn how to set different LOD's. Curt, as I was working on the MD11, I get a feeling that FlightGear is loading every single polygon into the scene graph. Is my feeling correct? If so, does that mean not implementing LOD's may actually consume less resources? Ampere On December 21, 2004 02:42 pm, Dave Martin wrote: Now thats interesting stuff; I *had* noticed that FG has lod settings and such but I don't know much about how it works in FG (Although I fully understand the principal). I'll have a go at making another version of the 172 with say, 2d wings, solid windows and less polys in the fusealage. How many levels of LOD (no. of models) do you think would be worthwhile? (I suppose you can first sight individual a/c at 6 miles out.) Cheers. ___ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] Reasonable vertex count for ground static?
Ampere K. Hardraade wrote: Dave, you can check out the animation file (specifically: engine animation) for the MD11 if you want to learn how to set different LOD's. Curt, as I was working on the MD11, I get a feeling that FlightGear is loading every single polygon into the scene graph. Is my feeling correct? If so, does that mean not implementing LOD's may actually consume less resources? Right, all versions of your models will be loaded so the system can pick which to display at any particular instance. OSG has a neat feature that will demand load your LOD levels, but we aren't quite to that point yet. Curt. -- Curtis Olsonhttp://www.flightgear.org/~curt HumanFIRST Program http://www.humanfirst.umn.edu/ FlightGear Project http://www.flightgear.org Unique text:2f585eeea02e2c79d7b1d8c4963bae2d ___ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] Reasonable vertex count for ground static?
Dave Martin wrote: Now thats interesting stuff; I *had* noticed that FG has lod settings and such but I don't know much about how it works in FG (Although I fully understand the principal). I'll have a go at making another version of the 172 with say, 2d wings, solid windows and less polys in the fusealage. How many levels of LOD (no. of models) do you think would be worthwhile? (I suppose you can first sight individual a/c at 6 miles out.) I would think that a super detailed version would be nice, some sort of middle range version, and some super tiny version that might be a dot or just a couple polygons. From there you could tune or add if you thought you needed to. At some point it would be nice to support Fade-LOD (I think OSG does that) but plib certainly can't. Regards, Curt. -- Curtis Olsonhttp://www.flightgear.org/~curt HumanFIRST Program http://www.humanfirst.umn.edu/ FlightGear Project http://www.flightgear.org Unique text:2f585eeea02e2c79d7b1d8c4963bae2d ___ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] Reasonable vertex count for ground static?
On Tuesday 21 Dec 2004 20:17, Curtis L. Olson wrote: I would think that a super detailed version would be nice, some sort of middle range version, and some super tiny version that might be a dot or just a couple polygons. From there you could tune or add if you thought you needed to. At some point it would be nice to support Fade-LOD (I think OSG does that) but plib certainly can't. Regards, Curt. If I regard the earlier 702 vertex model as the 'super-high', this one has about 300 vertices: http://www.cyfinity.com/fgfs/c172p-static-lod.jpg and is viewed from about 100 feet. (Could be used up to 2 miles or so.) I'll make another that is just four or five polys and no textures and then try that one as a scenery placement viewed from 2miles. ___ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
RE: [Flightgear-devel] Reasonable vertex count for ground static?
Hi Dave Dave Martin writes I've made the model look near to the normal model in quality unless you go right up and look closely (interior textures, controls etc are gone) The seats are still in but I've cropped a lot of vertices from the fusealage and wings and re-orientated the model nose-up to simulate an empty cabin. (I made some 'line' tie-downs too just for show. The vertex count is down from near 1800 to 1200 and the number of 512x512 textures is halved. The static 747 and 737 at KSFO are 1100 and 400 vertices respectivley.So that may give you an idea how complex you think the 172 should be. Dave Martin Cheers Innis ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
