BobtheVila
Newcomer
The vu1 does not store the graphics, it is on the GS and that is why it's connected to the GS, Also 4 passes equals 300Mpixels(out of the fully textured and Gouraud shaded 1.2Gpixels) as shows from a developer's ps2 optimization paper. The ps2 is only ment to do one draw buffer like GC so multi-pass is 2d overlays, though they use alittle more fillrate then normal single pass render to texels. This is good too as 2 objects can't occupy the same space at the same time, so it should too in performance when doing texture effects.
If you think the ps2 version uses worse nm then you have the lighting detail only on the defalt 5, at with no frame rate cost the lights are more refined, therfore it's better looking and more appearent. It's basically on everything, they just have that game even on xbox to show the NM on specular surfaces. Also ps2 can't doing it is saying NM and BM is not made of code period, and actualy the pixel engines are operations too, the bump mapping and rest on xbox a merely a bunch of hardwired FPU's...literally. This is so the CPU is only for AI and gameplay.
PS2 is like this to by having VU1 seperate from fpu and VU0 the only vectors doing AI. Vu1 has 3.08Gflops and 2.4 when all the transformation and lighting is done. 10 to 80 Gflops of the GC and xbox is actually the whole console including pixel pipes, vertex shaders, multi-pass operations, and tons of hardware effects like filtering alpha, even cpu Gflops and other, so they have much less then that for the vertex, T&L and TEV on GC, I mean GC only does 12million triangle transformations and ps2 66 and xbox 115, so the vertex shaders which do the poly transformations are better in that....but they are hardwired and xbox can't do curved surfaces on everything like ps2 can with still 16mill transformations left(more then GC's #) There is some demos that show this, I seen the vids, they are pretty cool.
Also you guys are probobly bandwith confusers, you confuse the vram as something that stores models when it doesn't. The bandwith is that, it's how much information is going through the tube. The vram part is merely the part that stores the buffers, all that is is the game screen's resolution byte per pixel.
If you think the ps2 version uses worse nm then you have the lighting detail only on the defalt 5, at with no frame rate cost the lights are more refined, therfore it's better looking and more appearent. It's basically on everything, they just have that game even on xbox to show the NM on specular surfaces. Also ps2 can't doing it is saying NM and BM is not made of code period, and actualy the pixel engines are operations too, the bump mapping and rest on xbox a merely a bunch of hardwired FPU's...literally. This is so the CPU is only for AI and gameplay.
PS2 is like this to by having VU1 seperate from fpu and VU0 the only vectors doing AI. Vu1 has 3.08Gflops and 2.4 when all the transformation and lighting is done. 10 to 80 Gflops of the GC and xbox is actually the whole console including pixel pipes, vertex shaders, multi-pass operations, and tons of hardware effects like filtering alpha, even cpu Gflops and other, so they have much less then that for the vertex, T&L and TEV on GC, I mean GC only does 12million triangle transformations and ps2 66 and xbox 115, so the vertex shaders which do the poly transformations are better in that....but they are hardwired and xbox can't do curved surfaces on everything like ps2 can with still 16mill transformations left(more then GC's #) There is some demos that show this, I seen the vids, they are pretty cool.
Also you guys are probobly bandwith confusers, you confuse the vram as something that stores models when it doesn't. The bandwith is that, it's how much information is going through the tube. The vram part is merely the part that stores the buffers, all that is is the game screen's resolution byte per pixel.