3-D Movies Are Missing The Point...Of View
    By Mark Changizi | December 8th 2010 01:36 PM | 6 comments | Print | E-mail | Track Comments
    About Mark

    Mark Changizi is Director of Human Cognition at 2AI, and the author of The Vision Revolution (Benbella 2009) and Harnessed: How...

    View Mark's Profile

    Many of the films we love manage to put us in someone else’s shoes, whether it be the shoes of a social network tycoon or a zombie killer. After all, we don’t pay $15 to see on screen what we do all day. Writers and directors get us into the protagonist’s head in a variety of ways, including letting us hear his or her inner thoughts.

    But a more direct route into the shoes of the protagonist is to make it appear as if the viewer is actually in the body of the protagonist. That sounds simple enough: film from the protagonist’s vantage point and have the actors pretend that the camera is the protagonist.

    Achieving a first-person perspective is, however, considerably more subtle than this. The question with which one must grapple is this: “What does a human view of the world actually look like?” Just because you are, at every open-eyed moment of your life, seeing such a view doesn’t mean the question is trivial. What is immediately before our eyes is often the stuff most easily missed.

    First-person video games get one step closer to realism by putting the character’s arm and weapon into the player’s view, providing the visuo-motor feedback of one’s arm that is needed for ably pulling levers or chainsawing demons.

    But the view from ourselves includes more of ourselves than just our arms and whatever weapon we’re wielding. As the father of “ecological psychology,” James J. Gibson, pointed out nearly fifty years ago, we can also see our own nose, cheeks and brow. That’s why American football players place black streaks on their cheeks -- their eyes see their cheeks, and are otherwise partially blinded by them on bright days.

    Neither video games nor movies have gone so far as putting the protagonist’s view of his or her own face into the camera view. And there is a very good reason for not putting the character’s face into the first-person viewpoint: it blocks the world that the viewer is trying to see.   If you close one eye, you’ll see that your nose alone takes up a considerable portion of your view. That’s a lot of movie screen to fill with a schnoz! And the same difficulty applies to the arm and wielded weapon, which also block one’s view of the actors and objects beyond the protagonist’s face and arms. No wonder the protagonist’s view of him or herself has been avoided in movies. It blocks too much of the world.

    And, yet… When we look out at our world with both eyes open, we do not find a nose occluding our view. Instead, our nose is visible to us, but it is rendered as transparent -- we see our nose, but also see through it to the world beyond.

    (If you can’t see your nose, try wiggling it.)

    In fact, as you attend to your transparent nose in your visual field, you’ll notice that you see not one nose, but two.  The nose on your left is the right eye’s view of it, and the nose on your right is the left eye’s view of it. Also, if you hold up your hand in front of you and look at something far in the distance, you will see two transparent copies of your hand. The amount of the world your hand occludes with both eyes open is small compared to how much is occluded when only one eye is open.

    As I discuss in my recent book, The Vision Revolution (Benbella, 2009), one of the central functional advantages of binocular vision concerns “seeing through” our own body parts. How else to put “gangly parts” (e.g., nose and hands) out in front of our eyes, thereby getting visuo-motor feedback, without occluding the world beyond?

    A transparent-double-view of our gangly parts is the result. That is, the answer to the question, “What does a human view of the world actually look like?” is roughly this: two partially transparent copies of our nose and hands, through which we see the world.

    We can, then, mimic the true human vantage point in movies and video games. And we can do so without occluding the viewer’s view of the world beyond.

    But…we can only do so by having both eyes open.

    For movies and video games that means using “3-D”. I put “3-D” in scare quotes, because now the intent of having different left- and right-eye images is for accurately putting ourselves into the protagonist’s shoes, not for getting 3-D pop-out effects at all.

    And that’s the point that 3-D movies are currently missing. What everyone appreciates about 3-D movies is the more realistic view of the world it provides. What is not appreciated is that 3-D movies are capable of simulating for viewers what it really looks like to view out of the eyes of the story’s hero.

    The greatest experiential advance 3-D movies might provide may not be the 3-D depth at all, but, I submit, the body-transport powers that bring us to stories of all kind in the first place.


    Ladislav Kocbach
    I have a question to this: is it not so that objects and structures more than say 5 meters away are perfectly well "depicted" by the usual perspective, combined with our "processing" and interpreting the relative sizes of various objects? It seems to me that only the nearest structures and objects really give us the spatial feeling, and such (is is called that) binocular pictures can give us new and exciting experience. I have lived through two demo 3-dim films, and they were really amazing. Monkeys were throwing bananas after me, birds were flying on me etc. In the second one we were siting on a sledge - nearly hitting trees - and moving faster and faster. Such scenes are really special in 3-d versions.
    I would also like to be able to have a permanent 3-d display for my molecular and atomic structure models.
    But I must completely agree that the 3-dim vision would not add anything to a conversation scene where two persons discuss in a dark room.  I suppose that when this somehow becomes mainstream, it will be without glasses and it will be sort of switched on and off, just like the light exposure and zoom 
    (But then there is also volumetric 3d in real spatial volume displays, but that is another story).

    3-dim imaging has been here for at least 100 years, movies perhaps for 50 years. Now some teams have found easier and more flexible variations of the traditional techniques - and suddenly everything must be 3-dimmed. 

    Excellent post Mark.

    And the other point of view of the matter, is the point of view itself.

    I refer to the angular aperture of the camera lens used, the distance to the objects, the distance between the two cameras lens to produce the stereoscopic effect, and the scale relation between them and the final viewer.

    To get a real visual perception you'd need two borderless screens (or a superimposed one), to project the images to each eye (nose and cheeks included) and use only a fixed camera with the two lenses placed at the same distance, aperture and focal angle than human eyes, which of course would terribly limit the usual tools of movie makers to get special or spectacular views and effects.

    But in addition to the problem referred to the size of the image and specially its borders, there´s a  very different effect in the image if you use an great angular or a zoom, and the distance between "eyes" also determine the perception of distance from objets in a 3D image. 

    As extreme examples I'd take the two tele-photos taken from the opposite point of the earth's orbit around the sun to make possible an stereoscopic view of the sky, the displacement of a satellite between the two photos needed to make a 3D perceptible effect on single picture of the earth surface, or the very little (micro) distance to produce a stereoscopic pict using macro lens on an insect's eyes.

    So, It also seems to me that, 3D technology (equipment, knowledge and techniques) have to experiment and evolved through all this difficulties along time to get good enough 3D movies. I intuit possibilities are unlimited.
    Though the first person view does lend a subjective quality to imagery that can be effective at immersing the audience in the character's world, there are limitations to just how true to life a stereoscopic image can be. 3D, as it currently exists, is still a technical cheat. In the real wold, our eyes converge, that is, "toe-in" on objects of focus as well as focus on them. In a theatrical presentation, our focus remains fixed at the screen plane, which for most of the audience will be around infinity. Though it is not what we experience in everyday life, most people's brains are able to recognize the visual cues as being somewhat familiar and will properly extract perceptual depth. It is, however far enough removed from reality that around 13% of people cannot see 3D at all in stereoscopic movies. The introduction of extreme foreground objects appearing in only one eye further "breaks" the illusion since the brain, now having no stereoscopic cues to extract depth, falls to the focal plane to determine the distance to the object which, as you remember, in a theater, is almost at infinity. Unable to resolve the scene, your brain will collapse the objects perceptually to the screen plane, causing discomfort and compromising the illusion.

    ..."toe-in" on objects of focus as well as focus on them. In a theatrical presentation, our focus remains fixed at the screen plane ...
    That's also right Tyler. Focus is another aspect of the visual perception. But I don't really know if focusing is a so unconscious, automated and periferal process that is independient from perception itself (object programation in software speaking). The way I see it, distance perception have a lot more to do with stereoscopism than focus, but it surely may at least cause some discomfort or visual stress.
    Mark Changizi
    Thanks all, for the comments!
    "It is, however far enough removed from reality that around 13% of people cannot see 3D at all in stereoscopic movies. "

    That's an interesting figure. i wonder if there are cultural differences, and if people can get better at seeing 3-d? I've done some animation of 3-d effects and was taught that as a general rule, objects flying right at the audience should never be more than 1/3 the screen width apart, otherwise the viewer cannot resolve the two images.

    As far as taking photos at different scales, as mentioned above, the most natural human perspective will result from the distance between the two vantage points being 1/30th the distance from the eyes to the focus point. It's fun to play around with those effects, for example looking at stereo pics taken from an airplane hundreds of feet to a mile apart makes you feel like a giant.