Nicely organised, but you get weapon pics with the above thumbnails aswell.

Sorry about that, the bug has been fixed.

You really need to stop commenting on technical GPU stuff.

1. NV2A in the Xbox is DX8.1 actually which is a solid featureset above DX8.0.

2. The PS2 graphics chip (the GS = Graphic Synthesizer) is mainly designed for output and is closely tied to the 4MB of eDRAM. Most of the graphics for the PS2 are done on the V0 and V1 processors on the EE. In a nutshell the PS2 has no shader model and has no equivalent to Microsofts DX7 API. It is a totally separate animal and there is no comparison. End of story.

3. HDR only means "High Dynamic Range" and is a reference to lighting. There are a LOT of ways to do this. For example Far Cry, SC:CT, Serious Same, etc use OpenEXR which requires Floating Point blending (namely FP16, although the 360 also introduces a penalty free FP10 using a mantesia variable and is 10-10-10-2). On the PC side the GeForce 6 series, GeForce 7 series, and Radeon X1000 series are all capable of this; previous chips are not.

But there are other ways to do HDR. One such method is to use Integer (I8). Another way is to use Pixel Shaders; indeed this is the method Half-Life 2 is now using.

Ultimately HDR is a group of technologies that include bloom, HDR render targets, adjustable aperature, and so forth. The Xbox *is* capable of a number of these and could do HDR. So it is incorrect to say, "it is impossible". It is possible, but much like the GF6 series--which is slow with most advanced HDR techniques--it is not necessarily worth the effort. BUT some elements of HDR (again, HDR is a group of technologies) are possible on the Xbox and are used.

Please, stop commenting on GPU technology. My head is still spinning from your comments that DX10 is in "early development". Your comments are misleading and not founding in facts. Sadly people are building on this misinformation.Both Sony and MS have been releasing high-AA shots the entire generation. The devkits frequently have features to take screen grabs that can either super sample the data or add extra effects. Almost EVERY Sony PS2 PR shot is like this. They look REALLY smooth and clean, yet the real game NEVER looks so good.

Basically you have to imaging the Aliasing :lol: Ironically this gen things are going a level deeper. We first saw pre-rendered footage called "game footage" and we are seeing stuff like DoA4 where in the screenshots you can add blur and really glossy up a nice picture. And they can LEGITLY call it "in-engine" since it is done in engine in a screen grab.

We need a new term for REAL footage, something like: Realtime Typical Gameplay Joe Consumer will experience out of the box with absolutely no deviation on a 720p TV.Hello? DX and OpenGL are only APIs. Application Protocol Interfaces. If an IC does not support a Shader Model or a specific API does not mean it cannot do certain tasks.

This is a totally "PC" way of thinking of hardware, and outside of Xbox1 and the 360 NO console EVER conformed to such methods. Heck, the first PC GPUs were not OpenGL or DX either! The first commercial success was the Voodoo1 which was GLIDE (3Dfx proprietary interface).

Bottom line is the PS2 is a completely different method of doing graphics. The GS is fairly limited, with its main resource being an extremely high bandwidth connection to the eDRAM for texturing and basic effects/lighting. Almost everything else is done on the processor (Emotion Engine).

In some ways this is sort of similar to way back when the GPUs were ONLY for rasturizing and the CPU did all your polygons, texturing, and lighting. Ironically that was less than 10 years ago.

In a nutshell a dedicated GPU with Pixel Shaders, Vertex Shaders, etc... is not the only way to skin the cat. Even more the PC bottlenecks (memory bandwidth; CPU<>GPU bandwidth and latency, piss poor CPU floating point performance) has resulted in a new feature in DX10 known as a Geometry Shader. This will allow the GPU to recycle information, do particle effects all GPU side, allow the control and deformation of models at the primative and higher level instead of working on a per pixel/vertex level, etc. BIG IMPORTANT stuff currently severely limited due to the above limitations. In 5 years we will wonder how in the world we ever lived with a Geometry Shader on our PC GPUs. Yet at this point most of you have no clue what it is, let alone what it does.

The point being is consoles are not limited by such things. Why would the PS2 need Vertex Shaders when the CPU does 6GFLOPs (about as much as a modern x86 chip)? Ditto why does the Xbox 360 need a Geometry Shader when the Xenon processor can write DIRECTLY from the L2 cache to the GPU without ever touching the memory bandwidth or FSB? Why would they need to dedicate the realestate to a GS when Xenon has more than enough floating point performance to work on the procedural data at a high primitive level and then send it to Xenos to tesselate? This allows more flexibility--the less hardwired limited function silicon the better the price/versatility of the hardware.

Same ideas apply to the PS2. It does NOT need to be OpenGL or DX to do the same things. It just does it DIFFERENTLY. This is why the PS2 has some killer looking games even though it is not "DX".

The bottomline is DX and OpenGL are only APIs. They do not dictate the only ways to end up with an end product. Ultimately to the consumer the only thing that matters is what ends up on the screen, not the API. And traditionally, until NV and ATI entered the scene everything was hardwired and the end product was a result of the hardware abilities or features, not any pre-fabbed API. And even now we are finding devs going backdoor to exploit the hardware in ways the API does not allow to get effects more advanced than the API itself allows.

Really, all PC knowledge needs to take a backseat when discussing consoles and other IC setups because the PC is a totally different animal.

Another great post by Acert93 :)