Slashdot Mirror
3D Cinema Doesn't Work and Never Will
circletimessquare writes "Walter Murch, one of the most technically knowledgeable film editors and sound designers in the film industry today, argues, via Rogert Ebert's journal in the Chicago Sun-Times, that 3D cinema can't work, ever. Not just today's technology, but even theoretically. Nothing but true holographic images will do. The crux of his argument is simple: 600 million years of evolution has designed eyes that focus and converge in parallel, at the same distance. Look far away at a mountain, and your eyes focus and converge far away, at the same distance. Look closely at a book, and your eyes focus and converge close, at the same distance. But the problem is that 3D cinema technology asks our eyes to converge at one distance, and focus at another, in order for the illusion to work, and this becomes very taxing, if not downright debilitating, and even, for the eyes of the very young, potentially developmentally dangerous. Other problems (but these may be fixable) include the dimness of the image, and the fact that the image tends to 'gather in,' even on Imax screens, ruining the immersive experience."
Vision scientist here ... sorry to have to disagree with you, but actually they are linked ... mostly for very near objects though, so the problems mentioned would be worst for handheld video games like the 3DS.
From this reference:
According to Prof. Martin Banks, Professor of Optometry and Vision Science at U.C. Berkeley, the vergence-accommodation conflict should be kept at less than ½ to 1/3 diopters for the majority of a 3D viewing experience to avoid discomfort and fatigue.
Which means if you are sitting ~16 feet from the screen, things can come ~10 feet out of the screen without you having any discomfort or fatigue. That is plenty of depth budget for most 3D movies. Thus, focus/vergence mismatch is not a real problem for stereoscopic 3D cinema.
Now if you are ~20 inches from the screen, things can only come out ~3 inches out of the screen before potential discomfort or fatigue, so vergence/focus mismatch is a real problem for small screens. Thus personal gaming devices, computers, and televisions will need careful depth budgeting in stereoscopic 3D.
"Super multiview" non-glasses 3D displays (generally with >32 views) where more than one parallax image is projected into your pupil at a time can force you to focus on the virtual 3D image where your eyes converge (this is how a hologram or the real world works, only they have nearly infinite number of parallax views).
Lack of closed caption support doesn't help either.
Subtitling for foreign languages are done on international 3D cinema distribution - carefully placed by hand for now (see the Nav'i English translations in Avatar, for example).
There is an expectation that Closed Captions for the hearing impaired will be delivered on 3D optical media with appropriate depth metadata as well.
That is factually incorrect. The Wikipedia article on depth perception lists clues used in depth perception, and the vast majority of them are monocular, not binocular.
The two most important things are convergence and the physical sensation of focus. People with one eye (or with one eye closed) looking around at a scene can tell the difference between close and distant objects quite easily up to a certain distance, and less so the farther out you get. It is this aspect of 3D movies that causes problems. They are projected on a screen at a fixed distance, and as such, one of those depth clues (and arguably the more important one at the distances we're talking about) is missing entirely. You can get get away with using distance to create a perceived difference in size when you're using a camera because you can't physically feel the difference between a camera focusing at twenty feet and focusing at thirty. With a physical room and the human eye, you'll only get away with that illusion for a few seconds before your brain figures out something is wrong.
Secondarily, the microscopic motions of your eye are enough to create a limited amount of motion parallax even with just one eye looking at unmoving objects. The natural motion of your head contributes to this. And so on. That, too, is missing from 3D projection.
Finally, the human eye does not perceive things as a perfectly flat image in the first place. The rods in your eyes are much more sensitive than the cones, which means that they tend to pick up scattered light, whereas the cones basically only detect direct light. This means that a single human eye can perceive a difference in focal distance in a way that cameras cannot. This difference results in subtle fringing around real-world objects of differing depth that can provide further depth clues.
So is 3D useless? No. Is it likely to fool someone into thinking it is real? Also no. There are too many visual clues that simply cannot be simulated through projection on a flat screen.
Check out my sci-fi/humor trilogy at PatriotsBooks.
Another problem with any fake 3D (i.e. dual images projected on a flat surface with binocular separation) is the fact that the parallax is fixed. When you view a truly 3D scene, your head doesn't stay still; it moves, even if just a little. You're not just sampling the scene from two angles, but from multiple angles. If you want a better look at a background object, you move your head to one side, and the image shifts a little. That's part of how your perception of depth works. Its much the same as why you need more than two speakers to create a realistically 3D soundscape (because we judge the direction of a sound in part by moving our heads imperceptibly). Even the most perfect flat-3D projection system cannot simulate that.
This doesn't mean that 3D "doesn't work", of course. It simulates an approximation of a scene, just as 24fps 2D images approximate it, and B&W 2D images approximate it less realistically, etc. But it will always fall sort of a true three-dimensional viewing experience. And kind of like a CGI rendering that doesn't quite look real (the Uncanny Valley), it'll always fall sort of satisfying.
Until we get real 3D projections. :)
http://alternatives.rzero.com/
Most use stereo because we have 2 ears... and therefore only 2 channels are necessary for 3D sound. 5.1 is a gimmick. Stereo forever.
That's false. We only have 2 ears, but each of those ears can distinguish sound coming from many different directions because your head and earlobes alter the sound differently depending on where it comes from. You can only try to generate realistic soundfields with 2 channels if you use headphones. You can record sounds with microphones in the ears of a dummy head (binaural recording), or you can try to simulate these effects through headphones. Both of those methods have problems, including the fact that the sound stays the same when you turn your head.
By your reasoning, we should only need 2 pixels on a TV since we only have 2 eyes.
has been in 3D for centuries.
on my computer
i started editing it. one problem: it sucks
not story wise or acting wise. but technically. the sound is awful, scratching, wind-blowing, the lighting is obviously amateurish. i used wireless mics and you pick up odd hums and rf ghosts. a nightmare
so there it will lie, forever, unreleased, until such time that i get over my perhaps too high self-standards about releasing a technically super-crappy movie in my name. but its embarrassing. i just don't want to edit it and release it. too depressing
someday i may finish editing it, perhaps drunk, to get over the depression of how much it technically sucks, just for laughs. so someday, you will have your laugh at how much i suck at the technical aspects of filmmaking
intellectual property law is philosophically incoherent. it is your moral duty to ignore it or sabotage it