PS3 3D Gaming NOT to require Special T.V.

[famousringo]

Slimebeast said:

famousringo said:

Slimebeast said:

What are these two different 3D techs called?

I mean one is like in the video in the OP, head tracking 3-D or something.
- this requires a huge increase in processing power.

And the other is with glasses where objects appear to pop out from the screen (like in some movie cinemas today, and the Nvidia 3-D glasses Desloitude has).
- this requires double the processing power

I want both techs immeditaly.

Head tracking doesn't require a huge amount of processing. If you're making the user wear infrared LEDs, it should take very little processing at all. You just interpret the position of the user's head and move the in-game camera appropriately to achieve the 3D illusion. Using a conventional camera, like the PS Eye, might take a little more processing power to interpret which part of the image represents the user's head, but I doubt it would take very much at all.

I agree with you. Somebody needs to get off their duff and take head tracking out of tech demos and onto the market.

But when u watch the OP, it seems like with head tracking on the scene must all of a sudden be rendered from every little angle in addition to the normal render of the scene. And u can even see the fram rate stutters because it's such a heavy scene.

So what I mean is, that the part u describe where the camera detects our head position may not take much processing power at all, but on the other side - what we see on the screen due to movements of our head position - needs big additional processing power.

I just watched the video, and I don't get why they're using 6 teraflops of processing power. Perhaps because they're using clustered CPUs instead of a GPU with a properly optimized 3D engine. All they're doing is shifting the viewpoint of the scene approriately as they move the camera, it shouldn't chew up much more processing power than walking around in an FPS. How the camera shifts would be a little tricky with this implementation, since it has to zoom out and pan in at the same time so that objects look like they're getting closer while more of the scene is revealed in the 'window,' but it shouldn't take a cluster to pull off.

I'm betting these guys simply aren't doing it the easy way because they're just hobbyists who lack either the software or the expertise, perhaps both.