People are now learning how to look at stereoscopic moving images just as there was a time, a century ago, when people learned how to look at movies. I’ve read that when motion pictures were first projected, if the composition didn’t include the entire actor – say they were cut off at the knees – people were taken aback by the effect, thinking that this was an image of an amputee. It took a bit of time to get away from composing a shot to look like the entire theater stage. There are stories about the Lumières projecting a movie of a locomotive headed towards the camera with the audience fleeing the projection in terror.
Perhaps such stories are apocryphal but I have lived through a somewhat similar change from the Edison aspect radio, of about 1.3:1 – to Scope which is about 2.4:1, and I remember what happened to our collective perception of the new cinema. This transition took place in the early fifties when it was believed that people wouldn’t be able tolerate a wide-aspect-ratio image in which the normal cutting pattern was used – which in those days was much slower than it is today. So we had shots of much longer duration, to give the audience time it was supposed it needed to take in the composition, and closeups were avoided. I guess they thought that a fifty foot mug shot would be unacceptable. Compositions were set up to allow the audience to take in the panoramic nature of the medium. Forward motion was big because it involved a three-dimensional-like effect heightened by peripheral vision. The Robe, the first CinemaScope movie, is of interest because of these visible changes in filmmaking style.
I think we have the same kind of thing going these days with the stereoscopic cinema. It is conventional wisdom that stereoscopic images can’t have rapid cutting. That’s the same thing people said about CinemaScope movies when they were introduced. This concern is based on the assumption that rapid changes in parallax can’t be taken in comfortably, and if these differences are large enough, they are going to cause discomfort.
The basis for this is the idea that it takes time for the eyes to re-converge and for people make sense of a stereoscopic image. There is some validity to this. Bela Julesz’s book Foundations of Cyclopean Perception has examples of some remarkable stereoscopic images (Julesz figures) -complicated CG textural images lacking in planar depth cues – that take time to build. (They are also called random-dot stereograms.) If you stare at some of the images for 30 seconds or a more, gradually depth will emerge from a flat image. It’s a trip to experience this time-plastic phenomenon, and it’s proof that it takes time for certain kinds of stereoscopic images to “build”.
What many people in the field recommend these days is to place the principal object at the plane of the screen in successive shots. That seems like a safe bet, since, for one thing, the breakdown of convergence and accommodation doesn’t occur at the plane of the screen because at zero parallax there’s no difference between viewing a stereoscopic image and a planar image. That would be all well and good if the breakdown of convergence and accommodation was an important consideration when looking at projected theatrical stereo images; but typically (this is well documented) it doesn’t apply for well-shot images on big screens viewed from the usual distances.
Another point to consider is this: If this problem was of such concern, how can it be comfortable to look at any stereoscopic image since there are different values of parallax spread all over the screen. Wouldn’t the perception of such images, with very rapidly changing values of parallax, as the eye darts about the composition, also cause discomfort? After all, what’s different about this from what goes on when we are cutting to another shot?
When you look around the visual field your eyes don’t see when they’re in motion. I know this may be hard to believe, but it’s a well established physiological fact that as your eyes move – when you have gross movements of the eye looking from one point to the other in the visual field- there is no vision. You’re blind for that instant. Your mind creates an effigy of the world, of the room or wherever you are seeing, so that when you look around things don’t appear to be discontinuous. In this case your eyes must instantaneously reconverge on whatever you’re looking at. If this is the case, why should rapid cutting be forbidden for stereoscopic films?
There are certain interesting psychophysical comparisons to be made between that which occurs in the visual world and in motion pictures. Pans don’t exist in the real world. You can move your eyes but when you do you don’t see a pan. Try it. When you look from here to there, it’s more like a cut; so cutting makes perfectly good sense since cutting in a film is a lot like how you perceive the visual world. When you look around the world, you’re actually cutting. The point I am trying to make is that you obviously don’t go through a period of confusion every time you look from here to there, which goes on constantly. You’re not suffering because your eyes are reconverging.
So if you take my points together – that the breakdown of convergence and accommodation doesn’t matter worth a hill of beans when looking at images on the big screen, coupled with the fact that when you look around you can’t see, and your eyes must reconverge instantaneously – you may agree that maybe the idea of keeping parallax values for the main portion of the composition, like an actor, may be questionable advice. Same thing goes for the rapid cutting of a stereoscopic film. Rapid cuts are what your eyes do ten thousand times a day.
I think when people learn how to look at stereoscopic images they will be able to enjoy effects that are now forbidden by conventional wisdom. Today when you see a movie in ‘scope you’ll see cutting that can be any which way. You can cut fast, you can cut slowly, you can even have closeups. Cutting is no longer deemed to be aspect ratio dependent.
We are in a new age of filmmaking, and when things change conventional wisdom gets battered; it’s part of the process of human learning for the forces of change and the forces of conventional wisdom to do battle. I’m not saying that filmmakers shouldn’t be conservative when they’re composing or cutting stereoscopic movies, because the audience doesn’t have much experience looking at 3D movies. But sooner or later there will emerge, by means of a collaborative effort between filmmaker and audience, a new style of cinema, a stereoscopic cinema.