| Saturday, October 29, 2005 |
  The quest for high frequencies.. | 
Posted - 10/29/2005 10:44:07 AM | 
I'm still working on the terrain integration and texturing.
It's progressing well but i'm still not happy with the atmosphere haze. It looks pretty good in the screenshots, but when animated it's not as nice. The haze seems to become "red-ish" ( it's very sublte ) in the distance, rather than blue.
I've adjusted the heightmap generation algorithm with the diamond-square noise to give more interesting and varied landscapes.
The texturing and lighting is not yet good though; seams are still very visible between adjacent textures, and the lighting seems to have lost its high frequencies. I'm not sure why, but i'm investigating..
| |
| Wednesday, October 26, 2005 |
| Quake 4 vs FEAR | |
Posted - 10/26/2005 3:04:10 PM | So, i finished both Quake 4 and FEAR this week end. I'll only be speaking of the solo, since i haven't got time to get far into the multiplayer.
Quake 4: a good surprise. Very similar to Doom 3, but that was expected. More action, more scripts, but as linear as its brother. The game is dark, but a bit better than Doom 3. The level design is a bit repetitive, but still okay. The vehicles sequences are very badly done, and pretty boring. The weapons and monsters are pretty varied. Overall, i'd give it a 14/20.
FEAR: a bad surprise. It's not a bad game at all, but i was expecting it to be so much better than Quake 4 than i could only be disapointed. To sum it up: the action is great, the graphics are great, the scenario is nice (although it suffers from the Half Life 2 syndrom, where nothing is really explained at the end), the special effects are fantastic (especially the slow-motion), the AI pretty smart.
But.. wow. Whoever made the level design deserves to be shot down.
You'll spend 2/3 of the game in a building made of generic rooms and empty corridors. Seriously, not only it's as linear as Quake 4, but there's virtually nothing interesting to see during a large part of the game. The maps do not make sense either, you feel like you're in a huge maze mapped with a building-style set. Also, FEAR is the king of video games regarding crates, barrels and various boxes. I want to cry.
The weapons and the opponents are not very varied. The scenario is a bit disapointing. The apparitions of the little girl are well placed, but often too cliché, and you quickly understand that you have nothing to fear from them, which killed the mood for me. Too bad for a game with that name.
The ending levels are a bit more inspired, but it does not save the game IMO.
Overall i'd give it a 14/20.
| |
| Tuesday, October 25, 2005 |
| The 3D artist quest | |
Posted - 10/25/2005 2:17:42 PM | So, i am now looking for a 3D Artist who can model/texture space ships or space stations. This is a freelance position with a small pay, not enough to get rich, mind you, but better than nothing. If you know any talented artist who might be interested, please ask him to contact me. Thank you.
| |
| Terrain integration, step 2 | |
Posted - 10/25/2005 5:54:19 AM | Been a while since i last posted some screenshots:
Things are becoming messy, to say the least. It looks nice and clean in the screenies, but i still have to face a high amount of bugs or problems. I've recently been working on matching the lighting with the sun direction as seen through the atmosphere (yeah, it was incorrect before!), but i think i still have a bug somewhere.. and it's not easy to fix, since the sun has to be transformed by the whole objects hierarchy, which means lots of object-to-world and world-to-object space matrices calculations.
Edges are still visible between adjcent terrain patches, sometimes it's very ugly, sometimes almost unnoticeable. In order to fix it, i'll have to replicate the normals on the bounaries between adjacent patches, but i've yet to write the code.
Another problem is with the heightfield. In the transition between my previous tech demo (which was a standard, flat terrain) to the current one (which is curved on the planet's surface), i feel the heightfield has lots some style. Hard to explain, but now, everything looks very repetitive and similar, at any place on the planet. It's probably just a matter of finding good parameters for the noise functions, but again another thing to fix.
I've switched the high frequencies noise generation to diamond-square - that is much, much faster than 3D Perlin noise, by an order of magnitude at least, and it's important when the camera is near the ground level, since if you're moving quickly, a lot of terrain patches have to be regenerated ( and at that altitude the noise is at high frequencies ). Unfortunately, the same scale is used everywhere on the planet, so it's impossible to mix flat areas noise (for valleys) and mountainous areas noise. Another point in my TODO list.
| |
| Misc. things | |
Posted - 10/21/2005 5:35:05 AM | For a while, i've been unhappy with the colors of my atmosphere. When the camera was positionned on the planet surface, perfectly under the sun direction (hence at noon), it didn't really look like daylight. There was a large band of white at the horizon, and the blue color high in the sky was too dark.
Sean O Neil suggested that i use an exposure function like 1 - exp(-cst * color) in order to fix these colors. I tested it, and i must say, the results are much better now! I had tried a lot of tricks but most of them changed the relative brightness or the colors, or desaturated them too much..
I'm now starting to generate the textures for the terrain, but i'm getting some ugly artifacts and seams. Hopefully i can fix them, and post more screenies..
I bought F.E.A.R this morning. Haven't tried it yet, but i think i'm not going to be very productive this week end.
| |
| Tuesday, October 18, 2005 |
| Two new wallpapers | |
Posted - 10/18/2005 8:28:27 AM | By popular request, i have posted two new wallpapers (1600x1200) on my website. You can also get them directly here:
Sunset
Moonrise
| |
| Terrain integration, step 1 | |
Posted - 10/17/2005 8:47:40 AM | 
I'm progressing well on the terrain integration. I have fixed a problem in the frustum culling code, which caused some planets to disapear when they were close to the edge of the screen. This actually wasn't a bug, but a restriction that i was unaware of, to the standard frustum culling algorithm ( the bbox vs 6 planes one ) regarding on how close the znear plane can be, which i fixed by adjusting the distance to the znear plane.
The two screens above are again at sunset ( or close to ), to "hide" the lack of detail texture on the terrain ( yeah i'm cheating ). The terrain is still very simple, until i add all the details and noise back.
| |
| Starfield | |
Posted - 10/14/2005 9:19:56 AM | I've been working on a new version of the starfield with darker colors. You can see two versions:
- pure black space:
- dark space with shades of blue dust:
The clouds and some more planet details are missing for now.
Another atmospheric sunset:
| |
| Wednesday, October 12, 2005 |
| Z-Buffering issues | |
Posted - 10/12/2005 8:50:01 AM | I partially solved the Z-Buffering issues. One of the problems when rendering a planet, is that (unless you're cheating), its dimensions are so big that you get Z-fighting very quickly. Today, i tackled the problem by reworking the piping system, and adjusting the ZNear/ZFar values for each planet. It looks like this:
- sort planets from back to front
- for each planet:
- duplicate the camera, set znear/zfar to match planet
- clear the z and stencil buffers, but not the color buffer
- render the ground
- render the atmosphere
- render the clouds and the rest
Remarks:
1. The znear value is set to the closest distance between the camera and the atmosphere. If the camera is inside the atmosphere, it is set to a constant.
2. The zfar value is set to the longest distance between the camera and the atmosphere (basically, the symetrical point on the other side of the planet).
3. Since the znear/zfar change for every planet, the zbuffer values become obsolete between many planets, and rendering has to be done from back to front.
4. By rendering the ground first, early-rejection of the atmosphere or cloud pixels can happen.
| |
| ATI, NVidia | |
Posted - 10/10/2005 4:51:21 AM | I've been pretty disapointed by ATI in the past year. It looks like the golden age of ATI is done. Since the Radeon 9700, ATI has improved their drivers a lot, and the developer support was excellent and responsive. But since a year, i've submited driver bugs twice and have been ignored - a simple email stating "we're looking into this" would have been enough. But no, nothing. And i'm a registered developer.
I wonder how ATI or NVidia are testing their drivers. Recently i saw a post on opengl.org, where somebody was reporting a driver bug in one of the latest NVidia drivers.. in a demo that was posted on their own developer site! You'd think that before releasing a new driver, they'd test it with all their available samples, demos and games, but it does not look like it is the case.
The last problem i had on my ATI ended up being a driver bug.. introduced in the Catalyst 5.4 driver. Hello Mac Fly ? Aren't you supposed to fix bugs instead of introducing new ones ?
Nowadays, it looks like the competition between NVidia and ATI is raging. They are rushing their drivers to be more up-to-date than the competitor. Seriously, how often are new drivers released ? Every month or so.
Yesterday i was reading the technical description of the future X1800, in the Radeon SDK. I'm not very impressed. ATI is catching up. In addition, in all the benchmarks i've seen, it doesn't look much faster (and even often slower) than the current Geforce 7800. And the 7800 still has more features.
So, we'll be missing floating-point textures filtering. According to ATI, this is no big deal because it can be implemented in a pixel shader.. in addition, vertex texturing is missing in VS 3.0, because "anyway it's too slow, nobody's gonna use it". Hello again ? Who's the developer ? Me or you ? So please don't tell me what i need or not. I can imagine quite a few uses of vertex texturing, even if it's slower than standard textures. If your hardware does not support it, that's one thing, but don't try to convince me that it doesn't matter or that it's not important, especially with the propaganda and marketing bullshit that has contaminated even your technical documentations.
Anyway, i'm currently working on the integrated terrain engine (no i'm not sleeping), but i have no new screenshot to post.
| |
|
| S | M | T | W | T | F | S | | | | | | | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | | 11 | | 13 | | 15 | 16 | | | 19 | 20 | | 22 | 23 | 24 | | | 27 | 28 | | 30 | | | | | | |
OPTIONS
Track this Journal
ARCHIVES
October, 2009
August, 2009
July, 2009
May, 2009
April, 2009
March, 2009
February, 2009
January, 2009
November, 2008
October, 2008
July, 2008
June, 2008
May, 2008
April, 2008
March, 2008
January, 2008
December, 2007
November, 2007
October, 2007
September, 2007
August, 2007
July, 2007
June, 2007
May, 2007
April, 2007
March, 2007
February, 2007
January, 2007
December, 2006
November, 2006
October, 2006
September, 2006
August, 2006
July, 2006
June, 2006
May, 2006
April, 2006
March, 2006
February, 2006
January, 2006
December, 2005
November, 2005
October, 2005
September, 2005
August, 2005
July, 2005
June, 2005
May, 2005
April, 2005
March, 2005
February, 2005
January, 2005
December, 2004
October, 2004
September, 2004
August, 2004
|
