Provocative comment by Id member about PS2 (and Gamecube)!
- Thread starter bbot
- Start date
Panajev2001a
Veteran
notAFanB said:
Well, you can notice this quite nicely almost everywhere ( small rooms are the best places to observe this IMHO ) when the flash light is pointed in a cedrtain direction and certain items that the light is not hitting directly are still casting a shadow: try to play Silent Hill 1 and then Silent Hill 3, the difference in lighting is quite obvious.
Silent Hill 1 did really use a single light source.
Gubbi, point taken.
As Silent Hill 3 showed to me, shadows can be other than pitch black and still allow the game to creep you out
everything is portable to an extenct. but the quality that's the difference off course.
remember those arcade games they ported to snes, megadrive, amiga, gameboy..etcc.. for example Mortal kombat.
ok , it was mortal kombat but damn what an ugly game it was on Amiga (lets not talk about the gamboy version )
SNES version was pretty close
if porting and publishing games was as cheap and ease as in the 8/16bit times, we would see alot more games on every console including 25 versions of DOOM3
remember those arcade games they ported to snes, megadrive, amiga, gameboy..etcc.. for example Mortal kombat.
ok , it was mortal kombat but damn what an ugly game it was on Amiga (lets not talk about the gamboy version
)SNES version was pretty close
if porting and publishing games was as cheap and ease as in the 8/16bit times, we would see alot more games on every console including 25 versions of DOOM3
Wrong._phil_ said:
_phil_ said:
No. It is not. I'm sorry to say, but you obviously don't have a clue how the lightning in Doom3 works.
The old way to draw shadows in games is to draw everything lit, then darken the things suposed to be in shadow. That is a hack.
The way doom3 does it is that it adds light to everything that is not in shadow. Things in shadow are not affected (They are not darkened or anything, they are left completely untuched). This is the correct way to do it, except the game doesn't add ambient lightning. It's completely possible to add ambient lightnig together with this, it was a design choice not to use it in D3.
Black shadows are only black because no light are added to the area it's covering, not because black polygons has been drawn there. If there is more than 1 light source then the shadow will not be black. Most places in D3 has more than 1 light source.
DeathKnight
Regular
The speed of the system! A GeForce2 won't be running Doom 3 similar to how you see it now. Neither will a GF1. The whole tie-in with the GeForce stuff is that the game's technology base is centered around everything that can be done with an original GeForceSDR. To actually pump out all of these things at playable framerates you need an insanely fast graphics processor.
The PS2 probably wouln't have that much trouble with the vertex stuff, although it would probably have to be scaled back somewhat to fit in with the rest of the system limitations. What's standing in the PS2's way is the pixel processing side of things and the amount of ram. The Cube on the other hand probably wouldn't have that much trouble with the pixel stuff, although it would probably have to be scaled back somewhat to conform with the speed of the pixel processing. What's standing in the Cube's way is the vertex processing side of things as well as the amount of ram.
The Xbox doesn't have the vertex processing limitation that the Cube has. It doesn't have the pixel processing limitation that the PS2 has. It also doesn't have the ram size limitation that both the PS2 and Cube have.
http://www.doomworld.com/php/screenie.php?dir=/doom3_051303/&number=1
Quite a nice screenshot demonstrating D3's ability to have several light-sources. Which creates nice non-black shadows except in areas where all the light-sources are blocked at once...
Quite a nice screenshot demonstrating D3's ability to have several light-sources. Which creates nice non-black shadows except in areas where all the light-sources are blocked at once...
Panajev2001a
Veteran
DeathKnight said:
Replace DOT3 per-pixel lighting with vertex lighting, decrease a bit the triangle size in the environments, segment the levels more ( like Quake II for PSOne and Halo do ).
PSOne had ~2 MB of main RAM, how many people expected to see Quake II's level in there ?
We would have to say bye-bye to normal mapping, butg that can be partially taken care of by better modelling: people estimate just about or less than 5K polygons per car in GT4 and the cars look as good or better than cars in games like PGR2 which supposely use 10K polygons per car.
It would not be easy to make a decent conversion, not at all, but I do not beleieve it would be impossible.
_phil_ said:ok,ok ,i admit you can vertex-alpha the shadow volumes (that doesn't goes far
And I admit that you still don't have a clue. D3 does not make shadow by drawing black polygons. Didn't I allready say that? (And no, you can not vertex-alpha the shadow volumes, nor would you want to, since they are drawn to the stencil buffer.)
D3 only adds light to a scene, it never subtracts light. Do you understand?
How to add multiple lights:
1. Clear the screen to black.
Consequence of step 1: The screen will be black
2. add light from light1 to every pixel that can 'see' light1.
Consequence of step 2: The only black pixels are those who can not see light1.
The pixels that can see light1 will have the brightness of the light received from light1.
3. add light from light2 to every pixel that can 'see' light2.
Consequence of step 3: The only black pixels are those who can neither see light1 nor light2.
The pixels that can see light1 but not light2 will have the brightness of the light received from light1.
The pixels that can see light2 but not light1 will have the brightness of the light received from light2.
Those pixels that can see both light1 and light2 will have a brightnes totaling the light received from light1 plus the light received from light2
4. add light from light3 to every pixel that can 'see' light3.
Consequence of step 4: The only black pixels are those who can neither see light1, light2 or light3.
The pixels that can see light1 but not light2 or light3 will have the brightness of the light received from light1.
The pixels that can see light2 but not light1 or light3 will have the brightness of the light received from light2.
The pixels that can see light3 but not light1 or light2 will have the brightness of the light received from light3.
Those pixels that can see both light1 and light2 but not light3 will have a brightnes totaling the light received from light1 plus the light received from light2
Those pixels that can see both light1 and light3 but not light2 will have a brightnes totaling the light received from light1 plus the light received from light3
Those pixels that can see both light2 and light3 but not light1 will have a brightnes totaling the light received from light2 plus the light received from light3
Those pixels that can see all three lights will have a brightnes totaling the light received from light1 plus the light received from light2 plus the light received from light3.
and so on until all lights are added. Notice that all that the engine has to do is what is written right after each number ("add light from lightx to every pixel that can 'see' lightx"). The complicated interplay between the shadows from the different light sources happens completely automatically.
Thowllly said:_phil_ said:
Hm, shadows, being stencils and an entirely different technique compared to lights, would have to be added separately from the rest of the scene. I'm pretty sure they darken things down rather than light everything else down, and that they are drawn as polygons...
*G*
stooop !...
Ok. what i mean : the volumes that define light and shadow (r_showtris 0)
Prohibits anything other than a straight hard edge for what appears to be the said shadow (can't be blured). also you won't see many multiple sources together very often in game because the price is exponential.
Ok. what i mean : the volumes that define light and shadow (r_showtris 0)
Prohibits anything other than a straight hard edge for what appears to be the said shadow (can't be blured). also you won't see many multiple sources together very often in game because the price is exponential.
Taking a wild guess, I figure it's probably because Doom3 has the same claustrophobic short distances and is therefore quite comparable in that respect.Qroach said:
Please. Splinter Cell for instance has shadow aliasing a dozen times worse then any of the above mentioned games and I never saw Chap ragging on it for it (granted it's on the wrong platform for him to do that but it doesn't change the facts).notAFanB said:
And stencil volumes have stairstepping period - just like everything else before you use AA to reduce it.
In case of SH3 the volumes are filtered instead - which will increase aliasing in some situations - but also reduce it in many others as well as give shadows a considerably more natural look then keeping them sharp.
We could debate whether they are overfiltered (imo ICO does the same thing nicer, although they also have to deal with much long draw distance which forces you to filter less), but the shadow aliasing relative to most games out there is minor.
Anyway, I was talking about something else - I personally find the low poly look of characters&shadows in D3 distracting, and SH3 for all its faults(if we really wanted to point and show what SH3 simplifies over D3 I could easily do it much better then Chap right now) doesn't suffer from that issue.
I suspect this is a rhetorical question?
I know your stance, but the tech argument may have "some" relevance regardless. Namely without using ambient, the first Z-filling pass can be textureless black pixels and therefore a fair bit easier on bandwith, if nothing else (on some chips you could do tricks to fill it with AA fillrate too).Gubbi said:
The art choice I can't argue with, as I don't know the background well enough. ID used to be hallmarked for pushing photorealism (which current stuff actually detracts from), but it's nice to know if they are really choosing stylistic approach (I've always been more of a fan of more NPR-ish stuff
).And why don't they do this, instead of making like D3 PC does.
1. Begin drawing the shadows that are the furthest from the light source.
2. While you go nearer the light source, increase the level of transparency of the shadow (the shadow could be a texture created by projecting the polygonal information of the object, 100% procedural).
1. Begin drawing the shadows that are the furthest from the light source.
2. While you go nearer the light source, increase the level of transparency of the shadow (the shadow could be a texture created by projecting the polygonal information of the object, 100% procedural).
marconelly!
Veteran
64 megabytes not a limit? Let's not be kidding ourselves.
Similar threads
