candid camera

Shortly after our wedding last fall, I was in Seattle for a conference, and naturally made time to visit with family. Somehow, I got into a debate with my mother about how she looked in some of the wedding photographs. Specifically, she felt she looked heavier than she does in real life (my mother is quite petite and slim). It's something she'd noticed in other photos of herself, too, and friends assured her she looked nothing like that. In fact, I'd wager that most of us have been dismayed at times by the fact that we look heavier -- or at least wider -- in certain photographs than we believe ourselves to be in the all-too-solid flesh. Our hips can't possibly be that wide, we insist, and loyal friends are quick to reassure us that the offending photo simply doesn't do us justice. "The camera adds 10 pounds!" they exclaim, and we are all too happy to agree.

So my mom wanted to know: why can't camera manufacturers build a camera to make her look like she "really" is? Well, at least one manufacturer has tried. Hewlett-Packard offers a digital camera with a built-in "slimming" feature that magically "takes off" a good 10-20 pounds off the subjects in the photographs, so that everyone looks thinner (see photo below). It caused a bit of a ruckus when it first came out -- how is it any different than Photoshopping (or, as some call it, "lying")? For some, it smacks of dishonesty, of "cheating," while defenders claim that it merely offsets the "fact" that the camera adds 10 pounds.

Anecdotally, this "fact" would appear to be at least partially true, although photographs are wildly inconsistent. For instance, I can personally attest that I, myself, appeared rail-thin in some of the wedding photos, and quite fleshy/voluptuous in others, whereas the "truth" probably lies somewhere in between. And I've noticed the same thing about certain celebrity photographs. But is this true from a scientific standpoint? Does the camera really "add 10 pounds"? Are those tiresome laws of physics (i.e., optics) conspiring against my mother to make her look unattractive in photographs? Inquiring minds need to know! So I set out on a haphazard quest to find the answer. Honestly? I was expecting to uncover some complicated optical secrets involving angles of refraction/reflection, lens shapes, and so forth. In retrospect, I was bound to be disappointed.

First I asked Michael Richmond, a physics professor at the Rochester Institute of Technology, while I was visiting RIT for a colloquium last fall. He was skeptical at first that any such effect existed, but I assured him that there were definitely wedding photos where the same people looked lighter and/or heavier -- in some cases, considerably so -- in different pix, even though all had been taken the same night. But he didn't think there was anything particularly optical going on -- at least not with the actual camera lens. Apparently there are several factors which might influence how "heavy" a person appears in a photograph, other than the lens:  the lighting, what a particular person is wearing, or how they happen to be standing.

Richmond's main hypothesis, however, was that the effect stems from the fact that the camera only has one "eye" (i.e., the lens), whereas human beings have two eyes, roughly 7 to 8 centimeters apart. The camera, it seems, lacks depth perception. The result is a kind of "flattening" effect that can make objects seem wider in photographs.

A true scientist, Richmond wasn't satisfied to merely hypothesize. He designed a simple experiment to test that hypothesis, involving a simple coffee mug and a background pattern featuring regular columns of numbers taped to the wall a few feet behind the mug. Then he took a series of photographs from different vantage points: straight on, directly in front of the mug (the "single eye" of the camera lens); 4 centimeters to the left (the position of  a human's left eye if said human's nose was directly at the center), and 4 centimeters to the right of center (corresponding to a human's right eye).

You can see his results, complete with photos/animations, here. Basically, the left eye captures a slightly different "view" of the background behind the left edge of the right eye, which in turn captures the "view" of the background behind the right edge of the left eye. The end result is that even though the mug in each photograph stretches across exactly the same number of pixels, the background turned out very different. When Richmond combined two photographs taken from the perspective of each "eye" and then fused the images, the background appeared to be much wider than in the photograph taken from the straight-on "single-eyed" view. Richmond's conclusion: "Even though the mug is really the same width in each, it will look 'fatter' relative to the background in the camera view." (Incidentally, this might explain why actors/objects in those 3D IMAX movies appear to be quite thin when viewed in 3D, but "flatten" into wider girths when one briefly removes the special 3D glasses.)

Richmond made a pretty convincing case, but every good journalist will tell you that it's essential to have a confirming source. So I emailed Charles Falco, a physics professor at the University of Arizona, who studies all kinds of things, including optics and human perception. You might recognize the name from his collaboration with David Hockney on the use of optical tools by the Dutch masters  (a controversial theory, albeit an intriguing one), but that's really just a favorite sideline. The bulk of Falco's research lies in artificially constructing multilayered and metallic superlattice materials and studying their x-ray optical, magnetic, magneto-optic and (in some cases) superconductive properties. Suffice to say, the dude knows his optics.

Anyway, Falco was even more unyielding than Richmond when it came to dispelling my fanciful notions about complicated optics conspiring to make us look chunky in photographs. He believes the effect is entirely in our heads. "What we see owes at least as much to psychology as it does to the simple optics of our eyeballs," he told me, because we actually "see" with our brains. Our eyes merely collect the data as light reflects off the objects around us. The brain interprets the data, fusing the input from our two eyes into a single image.

And therein lies the origin of "the camera adds 10 pounds" myth: the imperfections of human perception. Our eyes really do play tricks on us. According to Falco, most common optical illusion puzzles play off the fact that our brains tend to fill in the blanks when it comes to visual information, which means we can see different things depending on how we look at a photograph or a painting. (I'm sure the folks at Cognitive Daily could verify this fact with loads of examples from their online surveys.) Check out this famous optical illusion of Leonardo da Vinci painting a man on a burro -- an image that is reflected (if one adjusts one's gaze a bit) in the face of the depicted painter:

Pretty nifty, huh? And so it is with photographs. If we see a photo of a woman in a dress, and part of that dress is in a shadow, chances are, she will appear to be slimmer than in another photograph where her dress is fully lit. That's because we only "see" the shape that is brightly lit. Similarly, if we see a picture of a large boulder, but part of the boulder is outside the frame, we automatically fill in the missing piece. In fact, we may well later recall having seen a photograph of the complete boulder, not just one part of it.

So are our loyal friends lying when they insist a photograph looks nothing like us? Not necessarily. We also store mental images of the people we know, including our own reflections in the mirror, and those mental images might not match the particular 1/1000 of a second captured in the camera's frame. We form impressions of people when we first meet them, and "Our neurons seem very reluctant to let go of those initial images and 'update' themselves with new visual information," says Falco.

So there you have it. If you think your thighs look way too big in that photograph from your high school reunion, don't smash your digital camera in disgust. It's not the camera's fault. It's your brain's. Maybe it wasn't the answer I was looking for, but the "evidence" is pretty solid. That's science for you: it's all about ignoring one's wishes and biases to uncover the truth. Even if the "truth" might not be what we want to hear.

Comments

[drew]

A good way to test this 'eye parallax' hypothesis would be take two pictures of the same person against the same background. One would use a normal lens, and the other would use a telephoto lens from a much longer distance. The telephoto lens should make the background appear wider with respect to a fixed distance. I took the following two photos to make a different point awhile back:

http://forums.howwhatwhy.com/showthreaded.php?Cat=&Board=space&Number=186473
http://forums.howwhatwhy.com/showthreaded.php?Cat=&Board=space&Number=186476

It still boils down to human perception, so you'll have to determine subjectively if the subject looks slimmer in one picture than the other.

[Shardith]

So it's not the camera that adds 10 pounds, it's my brain!

[Jam]

I'd like to know if the physics behind hp's slimming feature is same as what is discussed here at the end. (btw, the link to "hp's built in slimming feature" doesn't seems to be working!)