Why does tilt-shift photography make things look tiny?
April 10, 2008 12:16 PM   Subscribe

Why does our brain perceive tilt-shift photography as tiny little toy models? I'd love to know the science behind this popular effect.
posted by waxpancake to Science & Nature (19 answers total) 39 users marked this as a favorite
 
I think it has to do with the narrow depth of field. Both the foreground and background are rendered out-of-focus and the true "sense of scale" is obliterated. It approximates the same visual image as if you are crouching on the floor with you eyes macro focused on toy cars, trains, etc. You see them centered and any periperal elements, above, below and on the sides are blurred.
posted by ericb at 12:27 PM on April 10, 2008


The ultra-short depth of field tricks the eye into thinking that the shot was taken with a macro lens, or close-up filters. If you look at actual shots of model train sets and the like, you'll notice a similarly short DOF.

My thought is that this is a learned response - it's not like people are born knowing that short DOF = a certain type of lens.
posted by god hates math at 12:28 PM on April 10, 2008


When you focus your own eye (or a camera lens) at something far away, the depth of field - the range of distances over which other items are in focus - is almost infinite, and everything that isn't much closer to you is sharp. When you focus on something close to you, everything even a little closer or farther away is out of focus. So by simulating that effect, tilt-shift photographs make their subject seem close, and therefore small.
posted by nicwolff at 12:30 PM on April 10, 2008


"By distorting the focus of the photo, the artist simulates the shallow depth of field normally encountered with macro lenses making the scene seem much smaller than it actually is. Many miniature faked photographs are taken from a high angle to further simulate the effect of looking down on a miniature."*
posted by ericb at 12:30 PM on April 10, 2008


Because that's how your eyes work.

Simple example. Hold two objects in you hand, bring the first object in front of your cheek, about 3cm away from your face. Hold the second object about 10cm in front of your eye and focus on that object.

The first object will be very blurred and also the background will be very blurred-- tilt-shift simply copies that effect.
posted by Static Vagabond at 12:31 PM on April 10, 2008 [1 favorite]


I don't really know, but since nobody has hazarded a guess yet...

I have a feeling that part of the key is that we're not perceiving them as tiny little toy models. We're perceiving them as photographs of tiny little toy models. And that's a big part of it, I think. Normally, with a fairly long shot photograph like that, you wouldn't have such a narrow range of focus. So you wouldn't normally have such extreme demarcations of blurring both in the foreground, and in the background. But you often would if it's a closer up shot of a much smaller subject. And that same issue arises with the color saturation maybe...that because the colors are over-saturated, it appears to be a photo of a model, in a well lit indoor location, rather than outdoors.

But that's all idle speculation by someone who knows very little about photography...
posted by Ziggurat at 12:33 PM on April 10, 2008


When you look at something really close (like a small model), your depth of field is very narrow (watching your finger makes everything behind be out of focus). However, when you look farther away, your depth of field expands and eventually becomes infinite (that is, it goes as far as your eyes can see). It goes the same with your camera optics.

Therefore, by reproducing this effect through lense displacement (as in tiltshift photos), a large space is "reduced" as the proposed perspective has been modified through the change in depth of field.

on preview, like everybody said
posted by ddaavviidd at 12:33 PM on April 10, 2008


Oops. I guess people did hazard guesses...
posted by Ziggurat at 12:33 PM on April 10, 2008


It's interesting to compare the original of an image with the tilt-shifted version. The objects in focus are unchanged by the tilt-shift. Yet your eye perceives them as tiny because the objects on the periphery are blurred.
posted by SPrintF at 1:03 PM on April 10, 2008


Best answer: Andy, I think it helps to look at pre- and post-processed photos. Check out this Photoshop tutorial for mimicking tilt-shift. Here's the original image. The same image after Gaussian blur is applied to the foreground and distant background. The final step is to brighten the central, unblurred portion of the image. This last step is an important one as well because brightening the center of the image gives the impression that there is overhead lighting selectively directed on the object - something that you wouldn't see in a normal outdoor setting photo.

At the eye-brain processing level, we perceive it as a photo of miniatures because we've only seen images with this effect in photos with miniatures. Our brain has been set up to match this type of image pattern to images taken of miniatures.

If you've ever gone up to the top of a tall tower or building, looked out, and squinted hard at a small portion of the scene, you'd get the same effect. Your eyes would be selectively focusing on a narrow plane which would cause the foreground and distant background to blur. Suddenly, it doesn't feel like you're looking down at a city, you're looking down at a miniature of a city.

Now think about this: if you were to show these images to someone who only recently acquired sight (was previously blind), they probably wouldn't draw any associations to miniatures. Rather, they would just see a photograph of a scene with the tops and bottoms blurred.
posted by junesix at 1:12 PM on April 10, 2008 [7 favorites]


The brain doesn't perceive the use of if the image was toy models. Ziggurat has the best explanation, the photography of models is usually taken from a higher than real vantage point and the physics of lenses prevents them from having a normal depth of field. Some people used TS lenses to replicate that effect to toy with people's preconceptions about how toys are normally photographed. The bigger part of the effect is the apparent height from which the images are made.

The effect isn't really a psychological illusion, it's more just a linguistic illusion, so to speak. It uses the language of a particular type of photography(model photography) with a different context.
posted by JJ86 at 1:16 PM on April 10, 2008 [2 favorites]


BTW, although the eye sees things with narrow depth of field, the brain doesn't.
posted by JJ86 at 1:18 PM on April 10, 2008


Response by poster: There seem to be some differing points of view here... If I'd never seen a photograph of something small before, tilt-shift photography wouldn't have the same effect on me?
posted by waxpancake at 2:32 PM on April 10, 2008


Because that's how your eyes work.

Huh?

Are you a scientist? Do you study vision? This thread is meaningless speculation.

I know there was a paper about this effect at VSS last year, and possibly SIGGRAPH, too. You can search around for it and find some scientific inquiry about the matter.

I will try to find the reference when I have a little more time. In the meantime, please, stop making uninformed statements about "how your eyes work".
posted by fake at 2:36 PM on April 10, 2008 [1 favorite]


Best answer: JJ86: ...although the eye sees things with narrow depth of field, the brain doesn't.
I'm not sure what you mean by that. The eye doesn't really see narrow depth of field, it just records "video" based on the settings given by the brain and then sends those images to the brain. As far as your eyes are concerned, what you see is just how the light hit the rods and cones. That you're able to recognize an image has narrow depth of field is the brain's work. Empirically, the fact that all disorders affecting depth perception occur due to defects or damage to the muscles around the eye, optic nerve, and brain, and none from defects/damage to the eye itself supports that.

waxpancake: If I'd never seen a photograph of something small before, tilt-shift photography wouldn't have the same effect on me?
Not just a photograph but if you had never spent any time looking at small objects and made the connection between the object's size and what you were seeing, you wouldn't make that tilt-shift-miniature comparison. A lot of this comes down to depth perception and depth perception relies on a number of cues. From infancy, you brain starts to catalog these cues, steadily improving its "algorithm." That's why babies run in to things - it's not that their eyes are bad, it's that their brain hasn't yet learned to properly process what the eyes see and intelligently control the eyes. Along the way, the algorithm set up a set of rules that basically say "IF (depth_of_field='narrow' AND object.align=center AND object.lighting=10 AND lighting.valign=top) THEN 'miniature' " after receiving repeated inputs like a Bayesian algorithm. Google for Bayesian, depth perception, object perception - it's well-documented.

Oliver Sacks had a great case study in An Anthropologist on Mars of a man who was blind since birth due to excessive cataracts. Other than that his eyes were in decent shape though so once they removed his eyes, it was like taking off a blindfold off a near-perfect set of eyes. As expected, he had poor depth and object perception because he didn't have the processing for the visual cues. But the more interesting aspect was that he didn't have any "visual memories," that is he couldn't strongly relate what he saw to his knowledge of that object, rather he had to hold it or feel it to make that connection.

So what I'm saying is that it's my understanding the process goes like this:
tilt-shift photo -> tilt-shift aspects of photo recognized -> fitted to miniature "rules" -> miniature

rather than:
tilt-shift photo -> tilt-shift aspects of photo recognized -> miniature
posted by junesix at 3:35 PM on April 10, 2008


junesix, the eye is a mechanical device, it is a lens. It works pretty damn much the way a camera lens does with the iris as an aperture that modifies the DOF as the light gets brighter. The brain can negate a small depth of field by quickly refocusing the eye on anything else within its range of vision thus making it appear that the DOF is limitless. Most people don't notice that their eye is even focusing over a narrower area because your brain really only pays attention to that which it is focusing on! In a bright situation where the eye would focus on a wide DOF, your brain still wouldn't be able to read things it wasn't paying attention to. It works selectively in a bizarre way.
posted by JJ86 at 9:38 PM on April 10, 2008


Best answer: Thing one: Depth of field. (What they said.)

Thing two: High angle. (Again, what they said.)

Thing three: Perspective. If you take a photograph of a building with a regular (let's say 50mm) lens from street level, in the resulting photograph the building will be narrower at the top (which is farther from the camera) than at the bottom. The wider the lens, the more prominent the perspectival distortion. With a tilt-shift lens (which mimics the adjustable bellows of a large-format camera), it's possible to compensate for the perspective distortion so that the top of that building is the same width as the bottom. This is precisely what tilt-shift lenses were invented for. The narrower depth-of-field is just a fun trade-off. Add the high angle, and there you go: Big stuff looks little.
posted by Sys Rq at 1:55 PM on April 15, 2008 [3 favorites]


Sys Rq nails it. Its all about perspective, not depth of field.
posted by doctor_negative at 2:40 PM on April 15, 2008


My thought is that this is a learned response - it's not like people are born knowing that short DOF = a certain type of lens.

I can report that the effect works on six-year-olds. Well, at least one: I asked our six-year-old son to take a look at the large version of this photo--which doesn't use a high angle--and tell me what it was. He said, "Houses." A couple seconds later, he asked, "Is it a model?"

Seconding Sys Rq's comment about the lack of perspective distortion.
posted by russilwvong at 9:11 PM on April 18, 2008 [1 favorite]


« Older Where do you go to escape?   |   "Big hello to Refrigerator Jim, on the half-eaten... Newer »
This thread is closed to new comments.