## Re: [Algorithms] Huge world, little precision

 Re: [Algorithms] Huge world, little precision From: Milos Tosic \(Shanghai\) - 2007-03-02 02:48:44 ```I would advise against using any ROAM-like algorithm with modern HW. You really don't want to be updating vertices all the time, among other = things! Milos Tosic=20 =CC=E8=EB=EE=F8 =D2=EE=F8=E8=9E -----Original Message----- From: gdalgorithms-list-bounces@... = [mailto:gdalgorithms-list-bounces@...] On Behalf Of = Mark Duchaineau Sent: Friday, March 02, 2007 5:27 AM To: Game Development Algorithms Subject: Re: [Algorithms] Huge world, little precision A thread that went over much of this ground on this list was "Terrain = Engine" -- see the posts by myself and Jon Watte around Jan 30th. Bottom line: there is a way to avoid the skirts and get the re-origin = operation to occur over a few frames to avoid hiccups. This is easy in = the chunked ROAM context. Cheers, --Mark D. Jon Watte wrote: > George van Venrooij wrote: >=20 >>It's true you don't need 200K vertices at a large distance from the = planet.=20 >>But it is our goal to allow fluent transition to ground level. If you=20 >>plug in very high detail elevation data and you're really close to the = >>surface you need a decent triangle budget to. We're aiming at about=20 >>100K triangles >> >> =20 >=20 >=20 > As I said before: the only way to do a large world on 32-bit hardware=20 > is to store world chunks in localized coordinate spaces, and then make = > up the difference in a combined worldview transform. You may get=20 > invariance problems between the blocks, which you'll have to fix up=20 > using some technique such as skirts (or just live with). >=20 >=20 >>We have read the geometry clipmapping paper and understand how it=20 >>works wonders for "flat" terrain systems, although it's hard to wrap=20 >>it around a sphere without doing some serious work on the polar = singularities. >> >>We decided to go for the ROAM-based solution instead to have a more=20 >>accurate view-based mesh without popping artifacts, true spherical=20 >>terrain (no toroidial mapping) and de-coupling from underlying data = structures. >> =20 >=20 >=20 >=20 > We express the sphere as an inflated cube. This gives us six separate=20 > projections, so the warping isn't that bad (1.41 : 1 at worst IAICR). >=20 > Cheers, >=20 > / h+ >=20 >=20 >=20 > ---------------------------------------------------------------------- > --- Take Surveys. Earn Cash. Influence the Future of IT Join=20 > SourceForge.net's Techsay panel and you'll get the chance to share=20 > your opinions on IT & business topics through brief surveys-and earn=20 > cash=20 > = http://www.techsay.com/default.php?page=3Djoin.php&p=3Dsourceforge&CID=3D= DEV > DEV _______________________________________________ > GDAlgorithms-list mailing list > GDAlgorithms-list@... > https://lists.sourceforge.net/lists/listinfo/gdalgorithms-list > Archives: > http://sourceforge.net/mailarchive/forum.php?forum_id=3D6188 >=20 -------------------------------------------------------------------------= Take Surveys. Earn Cash. Influence the Future of IT Join = SourceForge.net's Techsay panel and you'll get the chance to share your = opinions on IT & business topics through brief surveys-and earn cash = http://www.techsay.com/default.php?page=3Djoin.php&p=3Dsourceforge&CID=3D= DEVDEV _______________________________________________ GDAlgorithms-list mailing list GDAlgorithms-list@... https://lists.sourceforge.net/lists/listinfo/gdalgorithms-list Archives: http://sourceforge.net/mailarchive/forum.php?forum_id=3D6188 ```