Go Back   nV News Forums > Software Forums > Gaming Central

Newegg Daily Deals

Reply
 
Thread Tools
Old 10-12-03, 01:34 AM   #61
Zenikase
Registered User
 
Join Date: May 2003
Posts: 99
Default

About that HL2 screenshot, it looks more like debris than just plain dirt. In that case, it could be scattered with shiny metallic materials that have been assigned a reflective metal texture. The only thing that looks odd is how the wall next to the ground is not lit.
Zenikase is offline   Reply With Quote
Old 10-12-03, 01:36 AM   #62
GlowStick
CoD4!
 
GlowStick's Avatar
 
Join Date: Feb 2003
Location: Florida
Posts: 5,786
Send a message via AIM to GlowStick
Default

doom 3 is not pickey choosey of what can be per pixel, in hl2 and every other game ever made some effects can only be cast on certin objects, and once that effect has to interact with a diffrent type of object the **** hits the fan. an example of it is the light thats shineing on the floor, it jumped from the wall to the floor and it was too much for it to handle, while doom3 dosent have that problem.
__________________
Intel i7-2600K, Corsair 8Gig, Corsair H100, Corsair 650D, Corsair HX750, ATi 6970, WD Caviar Black 2TB
Sony Vaio SB: i7, 8Gig, Intel 320 300gig
GlowStick is offline   Reply With Quote
Old 10-12-03, 01:39 AM   #63
Zenikase
Registered User
 
Join Date: May 2003
Posts: 99
Default

It may be true that all objects in Doom3 are lit per-pixel. But I don't understand why using a light on both a PP and non-PP surface at the same time would result in an inaccurate result like that.

Last edited by Zenikase; 10-12-03 at 01:43 AM.
Zenikase is offline   Reply With Quote
Old 10-12-03, 05:35 AM   #64
Pafet
GeForce 7900GT
 
Pafet's Avatar
 
Join Date: Sep 2002
Posts: 430
Default

Quote:
Originally posted by Zenikase
I can't list them ALL off the top of my head, but I certainly remember (from the glacier level) that UT2003 used per-pixel effects to bump map the walls and give the effect of jagged edges of ice. Consequently, I'm sure Splinter Cell uses per-pixel lighting extensively in conjunction with its other lighting/shadowing effects.

Halo also used a lot of bump-mapping, the one I remember most is in the outdoor levels when looking out at the ocean. The whole surface is glimmering and full of tiny dimples.

I think that new Chrome game uses it too.

Again, I can't give a lot of references because I haven't played a lot of recent FPS games at all. You need to consider the fact that there are still games out that use the Q3 engine like Call of Duty and Jedi Academy. I also know that I'm referencing only to bump-mapping, but that's because it's the most noticeable effect that per-pixel lighting gives. Aside from that, you also get a nice round specular highlight as opposed to the geometry-dependent ones that are warped and jagged.


You sure those use per-pixel lightning and not just some pixel shaders on certain surfaces?
per-pixel lightning by itself is pretty subtle. its more noticable where polygons meet.
Pafet is offline   Reply With Quote
Old 10-12-03, 08:30 AM   #65
MrNasty
MrNasty
 
Join Date: Dec 2002
Location: Deli
Posts: 213
Default

Quote:
Originally posted by Toaster
Either iluminated or not...Like this you mean?



Take a good look at the nearest 'spider' shadowS..


Nutty's earlier post summed it all up very nicly From a technical point of view D3 is superior to HL2.


[Edit] link fixed.
When I say "illuminated" or not that pic proves my point perfectly... Notice how the ground is either shadowed or it is not shadowed by character shadows... there is no happy medium IE gradients between the two as you would find in real life.

Also if you look closely the legs of the spider thing are sooooo low detail. Low polly and you can actually see the leg is slightly pixelated and the pixelation has been blured.

It looks sooooo fake.

The doom3 engine is using a more advanced lighting system but it sacrifices realisim for dynamicacy of lighting.

It's very obvious that lightmaps and texture projection provide a more realistic looking lighting enviroment compared to stencil shadows in senarios where lighting is static.

Simply look at Splinter cell to understand what i'm talking about.
__________________
You are not your system specs.
MrNasty is offline   Reply With Quote
Old 10-12-03, 08:39 AM   #66
Nutty
Sittin in the Sun
 
Nutty's Avatar
 
Join Date: Jul 2002
Location: United Kingdom
Posts: 1,835
Send a message via MSN to Nutty
Default

Quote:
Also, is it correct to assume that the game uses mostly fixed-function T&L and texturing,
It dont make a difference if you use fixed function, register combiners, or advanced PS2.0 shaders. They all use textures. Textures are here to stay. There seems to be this misnomer that pixel shaders are gonna replace multi-texturing. I dont know where it comes from, it couldn't be more further than the truth.

NV3X has the ability to access 16 textures in a single pass. Multi-texturing and pixel shaders are not mutually exclusive.

On GF1/2 and equivelent ATI's it will use fixed function yes. On GF3/4 it uses vertex programs, and register combiners. On NV3X and R3XX, it uses vertex programs, and pixel shaders for a bit of added precision over what GF3/4 is capable off.

Quote:
that's great. how are bullets handled? do they have sizes and shapes to decide whether they'll pass through fingers? can they ricochet?
They probably dont have size. Most likely a ray. (Infinitely thin). Because ray->triangle intersections are faster than capsule->triangle tests. Ricocheting is very possible, and surprisingly easy too.. You just reflect the ray, of the surfaces normal, with a bit of randomness thrown in, when you hit a wall, or you could specify which surfaces ricochet bullets.. wood for example would just absorb the bullet.

Quote:
Please list the games we can see per pixel lighitng in for the last 2 years.
Halo! Giants Citizen Kabuto has per pixel lighting on the gf3 patch. Kabuto was very bump-mapped.

Quote:
It may be true that all objects in Doom3 are lit per-pixel. But I don't understand why using a light on both a PP and non-PP surface at the same time would result in an inaccurate result like that.
If its not per pixel lit, then its goraud (sp?) shaded. If the polys are big. In relation to the distance to the light source, they wont shade very accurately at all..

Imagine a big square, with a light very close to the surface in the middle of it. With per vertex lighting, the lighting contribution at each vertex is interpolated over the surface. But because the angle between the vertices' normals, and vertex to light vectors is soo large, you will get very low lighting contribution on the corners. Interpolate this, and you get a dark quad all over.

With per pixel lighting, the normals are taken at every pixel, so you correctly get the right lighting contributions at every pixel.
Nutty is offline   Reply With Quote
Old 10-12-03, 08:41 AM   #67
Nutty
Sittin in the Sun
 
Nutty's Avatar
 
Join Date: Jul 2002
Location: United Kingdom
Posts: 1,835
Send a message via MSN to Nutty
Default

Quote:
Also if you look closely the legs of the spider thing are sooooo low detail. Low polly and you can actually see the leg is slightly pixelated and the pixelation has been blured.
I'm pretty sure thats the JPEG causing that. Go look at the uncompressed 4000x3000 TGA's. They dont have that problem.
Nutty is offline   Reply With Quote
Old 10-12-03, 08:44 AM   #68
mrsabidji
I am the cheese...
 
mrsabidji's Avatar
 
Join Date: Jul 2003
Posts: 506
Send a message via ICQ to mrsabidji
Default

Gouraud.

mrsabidji
__________________
Signing posts since 1427.
mrsabidji is offline   Reply With Quote

Old 10-12-03, 09:30 AM   #69
Toaster
Registered User
 
Toaster's Avatar
 
Join Date: Sep 2003
Location: Antartica
Posts: 143
Default

Quote:
Originally posted by MrNasty
When I say "illuminated" or not that pic proves my point perfectly... Notice how the ground is either shadowed or it is not shadowed by character shadows... there is no happy medium IE gradients between the two as you would find in real life.
Yeah, as you said:"HL2 is a lot better cause the shadows look more real and soft like in real life" and you are correct.
The shadows in HL2 are way more realistic. Characters/objects don't selfshadow ( as in real life ), weapons don't cast shadows at all( as in real life ), shadows have ugly jaggies ( as in real life )...

Besides, even HL2 'soft'shadow edges are nothing like in real life.

You simply can't say that the HL2 shadows are better than D3's. Period.

Quote:
Also if you look closely the legs of the spider thing are sooooo low detail. Low polly and you can actually see the leg is slightly pixelated and the pixelation has been blured.

It looks sooooo fake.
The shot is a jpeg compressed jpeg ( the original jpg was to big to upload here, so I had to downsize and recompress it )

Quote:
The doom3 engine is using a more advanced lighting system but it sacrifices realisim for dynamicacy of lighting.

It's very obvious that lightmaps and texture projection provide a more realistic looking lighting enviroment compared to stencil shadows in senarios where lighting is static.

Simply look at Splinter cell to understand what i'm talking about.
How can you even claim that projected textures give higher quality shadows than stencil based ones?

You have to see D3 in motion to really apriciate it.
Toaster is offline   Reply With Quote
Old 10-12-03, 09:37 AM   #70
Zenikase
Registered User
 
Join Date: May 2003
Posts: 99
Default

Quote:
Originally posted by Nutty
It dont make a difference if you use fixed function, register combiners, or advanced PS2.0 shaders. They all use textures. Textures are here to stay. There seems to be this misnomer that pixel shaders are gonna replace multi-texturing. I dont know where it comes from, it couldn't be more further than the truth.

NV3X has the ability to access 16 textures in a single pass. Multi-texturing and pixel shaders are not mutually exclusive.

On GF1/2 and equivelent ATI's it will use fixed function yes. On GF3/4 it uses vertex programs, and register combiners. On NV3X and R3XX, it uses vertex programs, and pixel shaders for a bit of added precision over what GF3/4 is capable off.


I think you misunderstood my statement. What I meant is that it uses fixed-function texture shading for features like bump-mapping and per-pixel speculars, instead of relying on custom pixel shaders to achieve the same effects (although, from reading Carmack's latest .plan file entry, it seems that specular effects on this type of hardware are still dependent on triangles). On the other hand, it still may need to depend on these pixel/vertex shaders for the reverse stenciled shadow volumes, which Carmack developed after that generation of video cards had been released and thus could not be used on the GF2 series.


Quote:
If its not per pixel lit, then its goraud (sp?) shaded. If the polys are big. In relation to the distance to the light source, they wont shade very accurately at all..

Imagine a big square, with a light very close to the surface in the middle of it. With per vertex lighting, the lighting contribution at each vertex is interpolated over the surface. But because the angle between the vertices' normals, and vertex to light vectors is soo large, you will get very low lighting contribution on the corners. Interpolate this, and you get a dark quad all over.

With per pixel lighting, the normals are taken at every pixel, so you correctly get the right lighting contributions at every pixel.
Right, but the problem isn't that the gouraud-shaded wall has inaccurate lighting...it's not lit AT ALL.

Last edited by Zenikase; 10-12-03 at 09:41 AM.
Zenikase is offline   Reply With Quote
Old 10-12-03, 09:46 AM   #71
Zenikase
Registered User
 
Join Date: May 2003
Posts: 99
Default

Quote:
Originally posted by Pafet
You sure those use per-pixel lightning and not just some pixel shaders on certain surfaces?
per-pixel lightning by itself is pretty subtle. its more noticable where polygons meet.
Per-pixel is a concept, which is later translated into an algorithm. This algorithm first appeared on the GF2's Shading Rasterizer and, later on, in programs that used pixel shaders 1.x/2.0.

And the point of per-pixel lighting is to ignore the lack of complex geometry in order to achieve better lighting results. So I would think that it's not noticeable at all where polygons meet. That would be a more accurate description of per-vertex (gouraud) shading.

Of course, I may just be misunderstanding your statement. Please let me know if I've done so.
Zenikase is offline   Reply With Quote
Old 10-12-03, 09:49 AM   #72
rth
Registered User
 
Join Date: Sep 2003
Posts: 152
Default

Quote:
Originally posted by GlowStick
Its not a bug, it cannot be changed.
lol you've got some serious mental problems.

on the one hand you believe FX's bad performance due to problems with its hardware is somehow fixable.

on the other hand you believe a lighting anomaly due to problems with software is impossible to be fixed.

wow
__________________
Ifnvidiaturnedoffcheatsforonedayit'sworldwouldfall apart
rth is offline   Reply With Quote
Reply


Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT -5. The time now is 06:49 PM.


Powered by vBulletin® Version 3.7.1
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Copyright 1998 - 2014, nV News.