Well there are many different ways to achieve 3D in games. Some are easy, some are cheap to implement, some are perfect and therefore cost a lot more performance to implement and some are crude and cost even more performance whilst not delivering the best image quality.

The Nvidia method is the crudest and requires the most performance. It can yield 3D on games not originally designed to output like that but it also has quite a few artifaces. Its simply forcing the game to render at a different camera angle at its most basic.

The Killzone 3 method renders the best 3D effect however it costs almost as much performance as the Nvidia method. It requires a scene to be rendered twice as often so the compromise is that they lower the rendering resolution by 1/2 along the horizontal axis. So 1280 by 720 becomes 640 by 720 which makes it the lowest rendering 'HD' console game im aware of in this mode.

The Crysis 2 method is an anaglyph method. It uses the depth buffer and its more efficient in that it costs ~1.4% performance to implement as the game is simply reusing the original 2D image. However it suffers the greatest number of artifacts as one eye can see parts of the scene which are not needed to be rendered in the original 2D version. For example you might only be able to see one side of a fence in 2D however in 3D you might be able to see the other side of the fence so the game must be designed to minimise these artifacts.

So much for that "free lunch" Cervat Yerli has been hyping.

Still, it's impressive that Crytek can get the effect they have for so little overhead. He confirmed that this 3D effect could in fact run on an Xbox 360 and has during testing.

Basically they both can give you the same result when you take into account the actual 3D technology and capabilites of the consoles.

I still think people are stupid to consider jumping into this tech now. Its just a poor implementation and will cost WAY to much for a meager visual enhancement on a tiny amount of media.

This same idea is why I waited until this year to buy a HDTV.

I'll wait until 2012 or 2015 when 3D tv's either use a passive system with cheap glasses or have a full glassless option.

.

The Crysis 2 method is an anaglyph method. It uses the depth buffer and its more efficient in that it costs ~1.4% performance to implement as the game is simply reusing the original 2D image. However it suffers the greatest number of artifacts as one eye can see parts of the scene which are not needed to be rendered in the original 2D version. For example you might only be able to see one side of a fence in 2D however in 3D you might be able to see the other side of the fence so the game must be designed to minimise these artifacts.

MS can say what ever they want, we will see, Also its not that the PS3 can't handle, or has a hard time handeling 1080p games, its the cost, developers don't want to spend the extra money, thats all it is.

Great post and Im learning new stuff lol but it still dont answer the framrate drops.

So lets concentrate at this part of your post for simplicity. Let's pretend the 2-D version of Crysis 2 on consoles is 30 frames per second. Now, what happens after the ~1.4% performance loss (which is insignificant, pretty much 0%)? Will the PS3 or X360 still output 30 fps in the 3-D version or is it 60fps?

It means that each eye would get 29.6 frames per second from 30FPS (note often the framerate is higher than the capped 30FPS) because they use the data from the Z buffer (distance into the screeen) to compute what the other eye can see as two eyes will see roughly the same picture.

Say frame A is left and frame B is right with each number in sequence being the next whole frame.

What Killzone 3 does is roughly: Compute A, Compute B, A1, B1 etc. in sequence as if they are original frames.

What Crysis 2 does is roughly: Compute A -> Infer B, Compute A1, Infer B1 so it only computes frames at the normal rate and not the double rate of other implementations like Killzone 3. This is possible because what A sees is pretty much nearly identical and often is identical to what B sees. However its not perfect and it does cause artifacts.

I am trying to draw conclusions. Good.

In the KZ2 (or KZ3) example. Lets say the 2-d version is 30fps. WIll the end result be that the PS3 outputs 30 frames per second (every other frame is meant for separate eyes, but that's something for the shutter glasses to sort out)? And the gamer will experience it as 15 frames per second since every other frame is blocked for one eye?

Now Disolutiude mentioned the framebuffer and double resolution in the 3-D version and that could explain why he talked about 8 frames per second per eye, but in my eyes that's practically the same thing because the end result is the same - the viewer gets his framerate halved due to 3-D.

I think its wise for Microsoft not to invest in 3D right off the bat. I have a feeling consumers are not ready for 3D. The adoption rate will be low just think about it a second I for one will not be buying a 3D TV just got my HDTV last year and with 3DS I have to ask the question will glasses 3D be around too much longer?

Not to mention the cost fact at least 1,300$ TV plus addition 200$ glasses for everyone wanting to watch. Geez how many consumers can afford this? I don't think the adoption rate will be high enough for at least 2-3 years!