Please take whatever Charlie Demerjian says with a grain of salt. He doesn't like NVIDIA, and doesn't miss any chance to criticize them. For an extensive - and unbiased - overview of 3D technologies, I can recommend the hardware guide at MTBS3D.com.
That aside, most of the criticism in the Inquirer article holds true when considering mass market home cinema solutions, and I can see why Sony chose a solution using polarization: the glasses are pretty cheap and all the tech resides in the TVs - which Sony just so happens to sell as well. I'm just curious whether they will use interlaced polarization like Zalman and Hyundai do - thus losing half the resolution in 3D mode - or a full-blown two panel solution like iZ3D does.
The NVIDIA 3D Vision technology, however, focuses solely on the PC market, both for professionals (quad-buffered OpenGL) and gamers. In these scenarios, the shutter-based system actually has some advantages: Wider viewing angles, the ability to use any display supporting sufficient refresh rates (CRTs, DLP projectors, plasma screens) aside from special 120 Hz LCDs and - at least in my experience - better stereo separation. On the downside, at 60 Hz per eye, flickering is noticeable, though nowhere near as headache-inducing as pictured in the article. Oh, and the glasses are rather expensive.
Considering performance: As a rule of thumb, expect about half of what you get in 2D mode at the same resolution no matter whether shuttering or polarization is used, especially if the game is limited by the GPU - which should be quite common on the PS3 with its rather weak RSX. If interlaced polarization is used, the performance hit is somewhat alleviated by only having to render half the resolution for each image, but it is probably still noticeable. It is no suprise that Sony soon went back on the "compatible with every game" statement made by an employee at IFA.
In addition to the performance hit, there are also some rendering techniques that cause trouble with stereoscopic 3D, mostly postprocessing effects and sometimes shadows, some games also render certain objects at the wrong depth (skyboxes, interface, light sources). In these cases, some more or less extensive software patches would be necessary. Being able to adjust the stereo depth might be necessary as well, since the 3D effect is perceived quite differently by different people, some can enjoy really high levels of separation, while others only like a slight effect.
Depending on how ardent Sony is on selling its own "3D ready" displays, we might even see compatibility with existing systems. There is no reason why the PS3 shouldn't work with displays from Samsung and Mitsubishi which are already available and using shutter glasses.
Would it require more processing power to render a 3-D image?
Yes. Actual 3D --no matter how you produce the effect from screen to viewer (two screens on an HMD, polarized display, red/green, etc)-- requires rendering two complete, but slightly different views of the frame.
Rendering tasks are doubled. It's not trivial. It's never 'free'.
Yes. Actual 3D --no matter how you produce the effect from screen to viewer (two screens on an HMD, polarized display, red/green, etc)-- requires rendering two complete, but slightly different views of the frame.
Rendering tasks are doubled. It's not trivial. It's never 'free'.
Rendering tasks aren't necessarily doubled if the resolution of each image is halved. But it's not exactly the same either.
In addition to the performance hit, there are also some rendering techniques that cause trouble with stereoscopic 3D, mostly postprocessing effects and sometimes shadows, some games also render certain objects at the wrong depth (skyboxes, interface, light sources). In these cases, some more or less extensive software patches would be necessary. Being able to adjust the stereo depth might be necessary as well, since the 3D effect is perceived quite differently by different people, some can enjoy really high levels of separation, while others only like a slight effect.
- This is a good point. I think those rendering techniques will be very quickly solved when 3D becomes a blip on the developer's radar, so I'm not worried about them. Perhaps we'll see voxels make a comeback as well, turning volumetric clouds and the like into geometry in their own right rather than just the pixel effects they currently rely on.
I wonder what percentage of the population can't see a 3D image using any combination of these techniques? I know I have a very slightly wandering eye which syncs back up every time I open my eyes, but it doesn't affect me otherwise.
Perhaps they'll eventually establish categories that work for different people groups.
The major problem being that the brain uses more than one aspect of vision to take depth cues, one of which being focal length via the lens of the eye. This can cause mixed signals in some people and interfere with the perception of stereoscopic vision.
In any case, the only way to really solve it is to go to true holographic display, which does present actual depth the way the eye sees real objects in the real world. Then everyone would see 3D in a monitor as in the real world, assuming they're capable of it normally.
But, holographic displays are probably a few decades out at least :P I just love the fact that we're discussing 3D gaming and 3D displays as an actual real-world issue soon to enter out living room :) My first tv was a black and white 10 inch CRT with bunny ears that I played atari and NES on :P
/le sigh
Here's a thought: Using circular polarization, you can easily wear a polarized contact lens in each eye in order to achieve the 3D effect from any 3D system that requires circular-polarized glasses. But, if you use a system like iZ3D that uses only angular polarization, the contact lens must be properly oriented in order to get the 3D effect, and even being off a few degrees results in degradation.
So, while I think iZ3D's double-monitor system is the absolute best solution I've seen so far, I'd like to see them use circular polarization rather than opposite-angle polarization.
With sufficiently advanced head tracking, why couldn't Natal just do something to achieve the same level of immersive 3D as Johnny Lee did with his Wiimote + IR glasses, sans the glasses?
__________________ There are three kinds of lies: lies, damned lies, and "to spec" gameplay footage.
- Benjamin Disraeli (paraphrased)
I wonder what percentage of the population can't see a 3D image using any combination of these techniques?
According to c't, a German computer magazine, somewhere between 3 and 15% of the population (depending on which study you believe) cannot perceive stereoscopic 3D. For those, the current 3D solutions are of course rather cumbersome, having to wear glasses just to achieve the status quo.
They are missing out on something though. I've been using the NVIDIA 3D Vision set for a little over a month now and despite some teething problems (ghosting, artefacts in some games), I'm enjoing it immensly.
I really hope we're finally seeing a breakthrough for 3D technology in the mass market after so many failed attempts. But it does seem to be for real this time: Support from major Hollywood studios, game studios and hardware manufacturers (even to the point where it has been included in the HDMI 1.4 specifications).
With sufficiently advanced head tracking, why couldn't Natal just do something to achieve the same level of immersive 3D as Johnny Lee did with his Wiimote + IR glasses, sans the glasses?
Head-tracking never provides the illusion of 3D, only the illusion of real 3D space. The ultimate example in that video is the stadium-window, which shows more as you get closer, just like a real window.
What Johnny Lee showed is best used in conjunction with a 3D system. So yes, as a kind of stepping stone, you can hook-up head-tracking with Natal and get a new sort of mechanic that would open up new gaming vistas, and it would be cool and excellent. And the next logical step is true stereoscopic 3D, with head-tracking. With head-tracking and 3D you can actually look around an image--true 3D.
