 |
|
|
|
|
|
||
|
|
||
|
|
||
|
|
|
|
|
|
||
|
|
||
|
|
||
|
|
|
|
|
|||||||||||||||||
Add Forum to Favorites | Send Topic To a Friend | View Forum FAQ | Track this topic |
Last Thread Next Thread |
| Terrain Geomorphing in the Vertex Shader |
|
 |
d000hg Member since: 1/21/2002 From: Durham, United Kingdom |
|||
 Posted - 5/8/2003 5:50:49 AM |
||||
| Can someone give me a link to PVS for terrain - I can't think of a good way to test if a patch is visible. Cheers! Read about my game, project #1 NEW (18th December)2 new screenshots, one from the engine and one from the level editor John 3:16 |
||||
|
||||
|
BenjaPrieto Member since: 1/10/2001 From: Chile |
|||
|
Posted - 5/8/2003 3:33:17 PM |
||||
| The code does not work in my radeon 8500 |
||||
|
||||
|
DrGoldie Member since: 1/17/2002 From: Vienna, Austria |
|||
|
Posted - 5/9/2003 7:51:43 AM |
||||
| could you be more specific why it does not work? i developed it on a geforce3 but also tested it on a radeon8500... DANIEL |
||||
|
||||
|
BenjaPrieto Member since: 1/10/2001 From: Chile |
|||
|
Posted - 5/9/2003 11:25:17 PM |
||||
| Sorry for the lack of detail :D I think it was a driver problem, because today, when i was going to write waht the problem was, it worked flawless. Anyways, the previus error, seemed to be not accesing agp memory, so the game said that the hardware could not perform the game, so it had to go to reference software drivers. |
||||
|
||||
|
DrGoldie Member since: 1/17/2002 From: Vienna, Austria |
|||
|
Posted - 5/10/2003 2:48:38 AM |
||||
| BTW: the source code also includes a complete bsp (quake3) level renderer including collision detection... maybe somebody has usage for it, DANIEL |
||||
|
||||
|
d000hg Member since: 1/21/2002 From: Durham, United Kingdom |
|||
|
Posted - 5/11/2003 5:45:58 PM |
||||
quote: Please? I can't think how to do check what's behind a patch of terrain efficiently, and I've not come across such a topic in the terrain articles I've read upto this point... Thanks! |
||||
|
||||
|
DrGoldie Member since: 1/17/2002 From: Vienna, Austria |
|||
|
Posted - 5/13/2003 8:00:13 AM |
||||
| well, PVS in terrain rendering is a really simplified case of PVS in general 3D graphics (say 3D meshes) since in terrain it is more or less a 2D problem. i did describe the basic idea in my article. if you have any specific questions feel free to ask... bye, DANIEL |
||||
|
||||
|
Junwon Member since: 5/30/2003 From: Republic of Korea |
|||
|
Posted - 5/30/2003 3:49:07 AM |
||||
I'm in trouble in DirectX 9.. maybe, I have to find something instead of D3DVSDT_FLOAT4 in vertex declaration. // this is my code. // TerrainMorph.vsh vs 1.1 dcl_position0 v0 dcl_position1 v1 dcl_position2 v2 dcl_position3 v3 ...same ///// // verxtex declaration in dx 9.0 HRESULT hr; D3DVERTEXELEMENT9 decl[] = { { 0, 0, D3DDECLTYPE_FLOAT4, D3DDECLMETHOD_DEFAULT, D3DDECLUSAGE_POSITION, 0 }, { 0, 16, D3DDECLTYPE_FLOAT4, D3DDECLMETHOD_DEFAULT, D3DDECLUSAGE_POSITION, 1 }, { 0, 32, D3DDECLTYPE_FLOAT4, D3DDECLMETHOD_DEFAULT, D3DDECLUSAGE_POSITION, 2 }, { 0, 48, D3DDECLTYPE_FLOAT4, D3DDECLMETHOD_DEFAULT, D3DDECLUSAGE_POSITION, 3 }, // stream, offset, type, method, usage, usageindex D3DDECL_END() }; if( FAILED( hr = pd3dDevice->CreateVertexDeclaration( decl, &m_pVertexDeclaration ) ) ) { return false; } pd3dDevice->SetVertexDeclaration(m_pVertexDeclaration); but, it won't work.. give me some help, plz.. |
||||
|
||||
|
Anonymous Poster |
|||
Posted - 6/23/2003 6:43:38 AM |
||||
| It seems you're using static vertex buffers and rendering each patch individually with a drawprimitive call. <br> It doesn't sound very good way in case of you have many patches with the lowest LOD. So u end up calling rendering method every second triangle. <br> Therefore I think that using dynamic vertex buffer and software based tweening is MUCH faster way. <br> juhnu |
||||
|
||||
|
jmhewitt Member since: 8/1/2003 From: USA |
|||
|
Posted - 8/1/2003 11:48:56 AM |
||||
| I am about to start implementing a terrain engine of my own and I was curious why most geo-morphing techniques use quad trees as a basis? Are there advantages over the quad tree for geo-morphing over a triangular binary tree? Or are there other reasons for using a quad-tree for terrain visulization? I know triangle bin trees are specifically used for this purpose and was wondering if geo-morphing techniquest hav more problems wiht them or why this is so. Thanks, Jason |
||||
|
||||
|
RAZORUNREAL Member since: 1/18/2004 From: Auckland, New Zealand |
|||
|
Posted - 1/21/2004 6:12:29 AM |
||||
| I thought it would be a good idea to mention that whats visible (or at least what might be) can be worked out in realtime per vertex, keeping the memory usage down. I dont think many people know how (in fact I dont know if anyone does, I made it up) but I'm gonna see if it works first. It only works for terrain/grid based things though I think. And wierdly enough this is an offshoot of my shadowing method (which I'm gonna try out first, along with my lighting method, cause they are more visualy impressive). I never meant to invent all this stuff, just you invent one thing by accident and the rest come and knock on your door... |
||||
|
||||
|
dnaxx Member since: 3/4/2005 From: Austria |
|||
|
Posted - 6/30/2005 5:52:44 AM |
||||
| Hello, When I try to run the demos (gamescene, mountains,...) the scene jerks a bit while moving around although the framerate is about ~100 fps. When I set "Disable_HardwareVertexShader yes" the jerks get less but are still noticeable. Is this common and where could this problem come from? Is there a way to switch the demos to fullscreen? When I try to set to fullscreen with "F2" the demos end with an error. Thanks, |
||||
|
||||
|
vince35 Member since: 3/6/2007 From: Montreal, Canada |
|||
|
Posted - 4/13/2007 9:50:15 AM |
||||
| Hi, Maybe I'm a little late to talk about this article, but I want to implement something like what they talk about in this article. Unfortunately, the link to the source code doesn't work anymore. Does anybode has the source code, or another example using the same technique? Thanks. |
||||
|
||||
|
GraphicsBas Member since: 8/12/2004 From: Utrecht, Netherlands |
|||
|
Posted - 6/9/2007 5:14:57 AM |
||||
| The source(and images btw) for this article are no longer available. Does anyone have the source available for download? GBS |
||||
|
||||
|
chronos Member since: 6/2/2000 From: San Juan, PR / Mount Olympus |
|||
|
Posted - 8/7/2007 2:56:27 AM |
||||
| Hi, I'm trying to figure out the reasoning behind the author's choice of tweeningA = (1.0 - tweeningL) * tweeningIfor geomorphing. Consider this example: Assume we have a geomipmap that is currently at level 0.5 and that switching would introduce a large enough error that its detail level will not change much as the camera moves away from it. Furthermore, imagine that its left neighbor is changing rapidly, from 0.5 through 0.99. In that case: Wagner's formula places tweeningA below both tweeningL and tweeningI (e.g., when tweeningI and tweeningL are both 0.5, tweeningA will be 0.25). Furthermore, as the tweeningL factor increases, tweeningA decreases. This means that as the "L" vertices draw closer to the position where they'd be removed, the geomipmap's "A" vertices will shift away from their removal point, producing a kind of seesaw effect between the "L" vertices and the "A" vertices. Is that how it's supposed to work? I don't get it. |
||||
|
||||
| All times are ET (US) |
Last Thread Next Thread |
|
