Originally posted by seven
Oh well, I was bored and this is after all a good reply on your side, so I might aswell reply:
As Fast pointed out, those 62\'208\'000 pixels is a flat image. Triangles are there though to give it depth, thus making a 3d image (hence the name 3d games).
*Suddenly, the images from my monitor pops out of the screen and starts dancing*:rolleyes:
Fastson is wrong on this point. It may be 3d gaming, but
TV screens are flat. They only look 3d because 3d games use polygons to give a 3d perspective. What he is right about is that polygons do get hidden. This is called overdraw, something that is best avoided and games are usually designed to have as little of it as possible. With Xbox\'s early Z detection, overdraw can be greatly reduced, more so than PS2, so Xbox isn\'t at any big disadvantage. Judging from how the Xbox seems to be designed, I\'d say that the extra polygons that the PS2 can draw simply don\'t matter. Xbox is more of a balance between textures and and polygons, where PS2 is almost pure polygons.
And my calculations above are correct and even if 32 pixels triangles might be unrealistic for in-game situations, it\'s there to make a comparasment where both systems are compared equally. You can\'t argue which system is better by comparing both using different values. If the game will use an average of 32 pixel triangles or more - who cares? The math is still the same and the result proportional.[/b]
Read above.
That means that no matter what polygons you use, the larger draw fillrate on the PS2\'s hardware will allow for more polygons using 0 and 1 texture layers. How many they\'ll use is entirely up to the developer and in real in-game situations, I believe it will be a mix between 1 texture layer and up to many more. As said in a reply further up, objects futher away don\'t need more than 1 texture layer because you won\'t see them as good anyway. Objects up close however will benefit from more texture layers, so on average PS2 might have a slight edge (example: Jak and Daxter, more polygons, but worse textures up close than Xbox games). And I also believe that the developer will have to program their engine to exclude overdraw situations.
This may get a bit confusing, but here goes: Any decent game programmer will put LOD (Level of Detail) methods on their games, so as objects get farther away, the # of polygons and possibly the # of texture will decrease. I don\'t think objects in the distance will mean more at all to either PS2 or Xbox, since they should have very low amounts of detail. Close up however, the most important thing is probably textures. For PS2 I\'m afraid, textures are a crash and burn matter. I\'m not that clear on the details, but AFAIK, for every texture layer you add, fillrate halves. From a start of 1.2 Gtexels/sec with 1, 0.6 with 2, and so on. Xbox on the other hand, has two TMUs per pipeline, so in fact, it has 2 Gtexels/sec effectively. Plus, it can do it twice in a pass, or produce 4 texels w/o redoing geometry (I\'m guess a bit here I\'m afraid). So basically, with 4 texture layers, texel fillrate is 1.33 Gt/s, which the PS2\'s drops to 0.15:eek:. I believe there was another problem with how the Graphics synthizer(sp?) is connected. I\'ll look into that later.
