I'd never heard of tilt-shift lenses until Cabel showed me the bit that showed up on boingboing Friday. A tilt-shift lens lets you move the lens elements off-axis for various purposes, the same effect you'd get by moving or tilting the film or CCD. This tilting lets you change the focusing distance across the film plane: to focus both near and far objects, you'd normally need a long depth of field and thus a narrow aperture, but a tilt-shift lens lets you keep the aperture wide open (so you can shoot with less light) and have the focusing distance long at the top of the frame and short at the bottom. Or, as in these photos, you can defocus everything but the middle of the image for an artificial depth of field effect at a long distance.

What's really surprising to me is that people have enough fluency with photographic images that they recognize that a short depth of field implies a short distance without even knowing why—or even knowing what depth of field is. We have an intuitive sense of the mechanics of photography just by being exposed to these images all our lives. If tilt-shift lenses were as common as they apparently used to be, we wouldn't have this reaction, and these images would just look blurry, not tiny.

To try to figure out what the cues are that make these pictures look small to us, I played with faking one up in Photoshop. It's easy enough to blur things out from a center line with the blur tool, but I found that that isn't quite enough: long perspectives give it away, so you have to crop out the horizon, and more. The tighter, the better. The Bittergirls images above work so well because there's almost no perspective at all—it's all parallel lines. I think the shift function of the lens helps there, as far as I understand how that works. It has to be from a high angle, for obvious reasons. Finally, the irony is that the bigger the object is, the smaller it looks. Because we judge the distance at thousands of feet, we know the image should have a uniform focus—if part of it's out of focus, it all must be. Since it isn't, the only way we can resolve the contradiction is to believe that the city is a tiny model. This all happens subconsciously and involuntarily. Even after we know that it's not a model, we can't stop seeing it that way.