There are also some comparison pics of the PS2 game vs. the XBOX version over at MSXBOX forums.
I am surprised nobody mentioned this sooner:
On the PlayStation 2…Specialize
"It is not at all the same working with the PlayStation 2 and PlayStation. The main thing was just with the simplest thing of just getting triangles up on the screen. Generally this is not a difficult process, but in order to get a high performance triangle (getting the clipping right and getting it to work at the maximum throughput) everybody (other developers) had to go through that work. We didn\'t have the same base to start as we had with the PSOne. You\'re really starting at ground zero with the PlayStation 2-expecially if you were a PlayStation developer because there were so many parts of the hardware to master"
"[The PS2 development library] had a very good low-level description of all the hardware, and then they had a very general high-level description. But in terms of the middle level (as to how all the thing interact)is really where the challenge has been"
"The biggest problem we had (and you\'ve already seen it in other games) is with the aliasing of the polygons. Right now you can really get a lot of polygons, but if you just put them next to each other you get kind of the stepped edges. Its really a challenge to author your assets to make them look good, especially when you\'re not used to dealing with so many polygons. Our artists get excited because they think they can do a car in 10,000 polygons, then you show it onscreen and it\'s all this flickering mess"
"Then they make one more [revision] and it looks a little bit better, another rev and it gets better. Every time they go through they learn how to make things look better and use the polygons more wisely."
"We went through a huge change in performance from where we\'ve started [on the NASCAR engine] until now in terms of trying to get all the processors running in parallel and doing useful work all the time-without it all locking up (laughs)."
"The PlayStation 2\'s graphics unit has this little programmer (a vector unit)inside that can run a little program independent of the main CPU. Maximizing the use of that vector unit is really what it\'s all about. You want to have that guy start generating polygons on its own and this is the sort of optimizations that we definitely can\'t get in the first generation, but in the second generation it greatly improves not only the look, but the speed of the game as well."
"We can do it, but there is a lot of basic work that everyone has to do to get things running. With the PSOne, it was very easy to get things running and then you crank it up to get things faster by writing specialized code. But with the PlayStation 2, you have to specialize from day one and get the architecture right."
On the Xbox…All about the Pixel
"Early on [Microsoft] didn\'t give us a lot of the low-level stuff and they gave us all of the high-level stuff. Unfortunately in this case, we wanted the low-level stuff because the transition from PC to Xbox is fairly easy, so I don\'t need a lot of help. (laughs) It was only recently when we went to the latest Xbox developers conference where they gave us a lot more of the detailed low-level information."
"Microsoft is really adamant about (bringing PC ports to Xbox), and said that they would not approve any title that looks like a PC game."
"The big changeover with Xbox is the NVIDA graphics processor is so much more advanced than the PS2 in terms of how much work it can do on one pixel. The PlayStation 2 is really fast, but it can\'t do a lot of work on a pixel. With the Xbox, you can run a little program to calculate what each little pixel is going to look like. So we write programs that say this little pixel is going to look like this because I can take a part, multiply it by another…you can do a lot of work. Really the Xbox is about trying to create high quality pixels."
"For instance if you have a car on PS2, you draw the car, and if you want an environment map, you draw the environment map. So essentially, you are drawing the car a second time. If you want a damage texture on it, you would draw it a third time. If you want to put a gloss map in it to remove some of the shine, you draw it a fourth time. So you are essentially drawing a car four times."
"On the Xbox, you have the ability to draw these four textures simultaneously, blend them all, and draw the car one time. In other words, you take the base texture, add in the environment map texture, subtract out the gloss texture, and then blend it with the damage texture. So you\'ve got really high-quality pixels that you can do. That\'s the part where some really interesting stuff is going to happen because this is the power that was previously available in the PC, but no one was able to use it yet."
"I must say tat this NVIDIA graphics chip is really amazing. The GeForce II Ultra runs at a similar speed, but this one will have to four textures, and you can run a program that is executed on every vertex and one that runs on every pixel. The power of the pixel is really where the difference in quality is going to come up."
"[Microsoft] is basing it all on the PC API (application programming interface)-the DirectX stuff-so we\'ve seen it all before on PC. Most of our code uses most of this stuff already and then we can really focus on the specialized stuff that the Xbox can do."
Hmmmmmmmmm...interesting stuff here. Good to see xbox is very powerful system, but why can\'t any of there games keep a decent framerate(nfl fever, auzik, project gothem, etc.). Just makes me wonder. What do you guys think of this?