Can 3D Be Saved? (or, It’s the Content, Stupid)

At the Consumer Electronics Show last January, 3D television was both the great hope of electronics manufacturers and a great novelty. By September’s IFA trade show in Berlin, it was still the hope of the CE industry, but novelty had been replaced by the queasy feeling that if this was as good as it gets, the industry is going to have to find another way to move its wares.

Photo: Sony Corp.

There are a lot of issues with 3D TV, the most significant of them being the need to wear special glasses to see the 3D effect (and if you try to watch without, everything just looks extremely blurry.) Some progress has been made on image quality. For example, Panasonic demoed sets that eliminated crosstalk—double images caused by part of the picture intended for one eye leaking into the other eye’s field of view. Sets that don’t require glasses could get out of the lab and the experts may find a way to eliminate the artifacts and image glitches. They might even find away to keep 3D images from breaking into discrete planes, an effect somewhere between and old View Master and a shoebox diorama. But even if they do, I fear 3D TV may never become a popular hit.

Increasingly, I am convinced that the real problem with 3D involves the psychology of vision and perception in ways that technology cannot solve. Human 3D vision is a very subtle thing. Try a simple experiment: Just block one eye. Not that much changes. In fact, you are more likely to notice the loss of peripheral vision than you are the loss of depth perception.

That’s because our ability to perceive the world in 3D depends on a lot more that our rather poor stereoscopic vision. We rely on perspective—ojbjects in the distance appear to converge even when seen through only one eye. Our ability to see the sides of objects and the shadows they case also helps. Most important, there’s the fact the far-away things appear smaller than nearby objects. In fact, probably all of us have experienced optical illusions in which messing with the sizes of nearby and distant objects completely overwhelms the clues provided by stereoscopic vision.

This leaves the producers of 3D content with two choices. They can keep the 3D natural, in which case the effects are subtle and sometimes barely noticeable. This is the route Pixar Animation has gone with recent movies, including Toy Story 3: The effect was pleasant enough, but I really didn’t think it was worth paying extra for. Or they can go the route followed by most of the meager assortment of  3D content now available and go for exaggerated and unnatural effects.

Pixar has another advantage. Animation is supposed to look artificial. The same was true to some extent with <em>Avatar</em>; the movie consisted almost entirely of computer-generated images and we have no idea of how Pandora or the Na’vi are supposed to look.

I have yet to see any #D video featuring ordinary human beings that seems natural. This was brought home forcefully during a Sony demonstration at IFA. I watched the live proceedings on the stage, including a performance by the pianist Lang Lang and a presentation by Sony CEO Howard Stringer while the same events were projected in 3D on a big screen at the back of the stage. The 3D had every advantage: Sony was using a properly lit set, professional 3D cameras (including, in what I think may be a trade show first, a crane-mounted camera), and cinema-quality projection gear.  Yet the contrast between what appeared on the stage and what I saw on the screen was disconcerting in the extreme. It was a long trip to the Uncanny Valley, that unhappy place where images that look nearly, but not quite, real play nasty tricks on your brain. What seems peculiar is that I have attended any number of presentations where people appear both live and on big screen 2D images and these seem perfectly natural. In generally, most of the 3D television that I have seen actually looks better and more natural in 2D.

I’m trying to think of applications where 3D would really help. Games are an obvious example, and, again, the unnatural look of video games works in their favor in the translation to 3D. I thought sports were a prime candidate. A 3D broadcast of the Master’s tournament last spring made it much easier to see the contours of the golf course, especially the greens. But golf, like most sports, relies on long telephoto shots that provide very little 3D effect. At IFA, Panasonic showed a 3D version of tennis from the French Open, and it too was disappointing. The main difference was that the fast-moving ball was blurrier in 3D than in 2D. Rapid motion, deep shots, and live broadcasts that leave no opportunity for post-production work make sports especially challenging.

The quality of 3D production will get better over time as both the engineers and the creative folks—cinematographers and directors—learn how to use it better. After all, artists have spent the last 700 years perfecting 2D representations of a 3D world that don’t leave us longing for 3D versions of, say, Raphael’s “The Transfiguration.” But the very fact that artists have achieved 2D perfection leaves me wondering if 3D is necessary—or whether it will be sufficient to ever sell large numbers of TVs.

Disclosure: I wan in Berlin to participate in a couple of panels at IFA. My travel expenses were paid by the organizers.