In a comment to the previous post Nick asks ‘how did the 3d actually look?’
Tricky to describe, in large part because I haven’t yet seen a 3D flick at the cinema in 3D, so I can’t compare it to that, but only to reality and to other TVs.
I haven’t seen the Samsung yet, but I have seen the Sony which works in a similar way as I understand it. I saw it some weeks ago, so the passage of time has necessarily attenuated the accuracy of my observation. So with that proviso, I’d say they look pretty much the same (for reasons to do with light polarisation, the Panasonic retains its effect better as you tilt your head, but the Pansonic is also more subject to making external light sources appear to flicker).
Compared to reality, it doesn’t really look much like 3D in my opinion for several reasons. But it at the same time it does.
Here’s why I think that is. There are several mechanisms in our visual systems to give depth information to us. As you can imagine, knowing how far away a predator or prey is can be a rather useful survival skill. Just as much, being able to accurately reach for a berry on a bush half a metre away is a useful survival skill.
The obvious way for depth perception to work would be by triangulation: your eyes are apart from each other by a certain fixed distance: the baseline. By comparing their slightly different views of an object, the brain is able to integrate a position for the object, or a point on the object.
It conveys this information to your consciousness not by a string of 3D coordinates (as you see in The Terminator) but in a composite image created from the two discrete images generated by the eyes. Remember, what you ‘see’ is never what your eyes capture. What you ‘see’ is your brain’s visual processing mechanisms’ interpretation of what’s out there, drawing both on the captures made by the eyes and certain pre-programmed assumptions about the behaviour of the universe. That’s why optical illusions work.
Nonetheless, the two separate fields captured by the eye are all important for this purpose. Close one eye and then try to grasp between finger and thumb a small object about half a metre away from you. If you do it at a customary speed, you will probably miss. Generally you can do it only slowly and with uncertainty.
As it happens, the ‘stereoscopic’ effect resulting from different viewing angles from two eyes, leading to a depth perception, results in what we perceive as 3D imaging. And that’s what all these TVs (and cinema systems) pull off rather nicely.
So, yes, there is a clear sense of objects at different distances. One of the clips I saw yesterday was a beach volleyball game, in which the players were etched so cleanly and clearly, standing forward from the background, that it almost looked like the sharpness control had its afterburners turned on.
Except there were no ‘sharpness control’ artefacts. It wasn’t processing that was doing this, it was the sense of depth perception that made the players stand out.
Truth be told, it was pretty damned thrilling.
But …
How else does your depth perception work? There are two categories of these mechanisms: those that can’t be replicated by these 3D TV systems, and those that are replicated by both 3D and regular 2D TV systems.
Of the latter, there is far distance perception. Triangulation doesn’t work very far at all: the baseline between your eyes is too short. So far distance perception is in part generated by scale. Your brain knows that mountains are big, but when they occupy only a small part of your visual field, your brain interprets them as far away. It also uses relative scan (a man standing next to an unknown tree gives the tree size, and therefore distance). It also uses clarity: the further the distance the greater the atmospheric haze. These can be done using regular TV systems, and often are.
Another device that allows us to have a sense of depth, and one that is used very frequently in cinematography, is focus. As with cameras, the same with the eye: the lower the amount of light or the closer the object in focus, then the narrower the depth of field. If you focus on your thumb at arm’s length (close one eye to avoid stereoscopic complications), the background will be out of focus. Bring your arm closer and the background becomes even fuzzier. Pay attention to the background, and your thumb goes fuzzy.
This strength of the traditional cinematic arts actually works against 3D TV (and 3D cinema). And it is where the major departures from reality commence.
Jez Ford (after watching Avatar at the cinema) and Jennifer Dudley (from The Courier Mail) have mentioned this to me as well. To illustrate, one of the clips yesterday had a distant view of the Grand Canyon, and in the close foreground at the bottom right of the screen was a bush. My guess is that the clip-maker framed it this way specifically to give an enormous sense of depth.
The stereoscopic effect has this bush clearly hovering well forward of the background, as it should, but it was also out of focus, as it should be. Here’s where things went wrong: in the real world you might glance at the bush, and as you did so it would come into focus and the rest of the scene would go fuzzy. With 3D TV, you glance at it and … it remains fuzzy! This is completely unexpected to our automatic visual processing mechanisms and it doesn’t fit in with the model of real-world physics on which their processing assumptions are based.
As I mentioned, optical illusions are similar: some unexpected (usually unnatural) visual phenomenon tricks these processes into producing some non-existent effect (straight lines looking curved, same sized boxes looking different in size, static score-marks looking like they are moving, non-existent black blobs are the intersections of white lines on a black background).
However in this case — the fuzzy bush that persists in remaining out of focus no matter how much your eyes strain — no artefact or alternative processing mode (which would generate an optical illusion) appears to be available. Instead it just looks totally wrong.
I found myself wanting to look away from it, yet my eyes kept being drawn back as though it were insisting on trying to resolve the problem.
There is one last depth perception mechanism that I should briefly address before coming back to the implications of the last point. And that is movement. Of course, a moving 2D camera generates a sense of depth as it makes closer objects move across the faces of further-away objects. It also discloses a different view of objects, as it increasingly reveals the sides instead of just the fronts of objects.
I have a strong suspicion that at a subconscious level we assist our depth perception by moving our heads, usually slightly, to vary the angle of view. For close objects a tiny movement gives a significant new information about it and its location.
In a (real) hologram you can do this, and move slightly to a side to see a little further around the edge of something. With 3D TV and cinema you can’t. Your head moves but the relationship of the objects on the screen remain unchanged.
Implications
Unless someone with a mighty big brain comes up with some extraordinary solutions, 3D won’t be getting better than it is. I don’t think it can. So that suggests that how well 3D works will lay entirely in the hands of content providers, and how well they work within its limitations. Given that 2D isn’t going anywhere soon, it could also means that we may end up with different 3D and 2D cuts of movies, with framing and focus and camera technique optimised to deliver the best results in the two formats.
So what should filmmakers consider with 3D?
First, they should avoid showing off. My guess is that the Grand Canyon scene had that bush composed just so to emphasise the sense of depth. It did, but at the cost of making the whole thing uncomfortable to look at. The early stereo Beatles’ albums had silly things, like a guitar locked firmly into one channel and the voice into the other. That’s what happens with a new technology: its capabilities are overused. Heck, pop groups and their engineers — and movie makers — are human, so this is only natural. But ideally over time the technology will become part of making the work of art, instead of the art being twisted to show off the technology.
And second, they should keep everything in focus. Not having the background go out of focus when I moved my attention to the bush didn’t cause problems. It was the failure of the bush to come into focus. I don’t imagine focus will be important for chase scenes because the appearance of the individual out-of-focus elements is too fleeting to attract attention. But for lingering shots, all ought to sharp.
Which is why I’m talking about different versions of movies for the 3D and 2D formats. In the latter the depth of the field of focus is an important part of the story telling process.
This brings up a point that probably requires clarification. If an out-of-focus bush is so off-putting in a 3D movie, why isn’t it so in the 2D version? Once again we come back to the visual processing systems in our brains. They don’t expect everything to be sharply etched. Lots of stuff in nature has graduations of colour and tone. Think again of those distant mountains behind many kilometres of atmospheric haze, or many natural objects. My guess is that our eyes don’t treat the out-of-focus bush in a 2D movie as being out of focus, but a soft, graduated object.
But in the 3D movie, it is clearly standing well forward of the stuff behind it. It ought to look sharp when we look at it, and it just doesn’t.
Finally, they should consider avoiding static camera angles. I suspect that while the angle of view on the screen is changing, any loss of 3D effect due to the ineffectiveness of micro-head movements is minimised.
Conclusion
I find myself hoping that 3D is here to stay. As film makers learn to use it with discernment it will open up a whole new palette of creative opportunities. It need not be merely for splashy action movies. Often, the whole point of cinematography is to make some element of the picture — often a character — stand out from the rest. 3D is a new way to do this.
It just needs to be done with discretion.