Two years ago, Blaise Aguera y Arcas, a developer at Microsoft, unveiled an application called Photosynth. In a fantastic presentation, he showed how the software can assemble a collection of digital snapshots taken at a certain place—say, all the Flickr photos of the Notre Dame Cathedral —into a grand, three-dimensional environment. What's more, Photosynth lets you pan and zoom through the resulting scene as if you're a director scouting out locations. On a large video screen, Aguera y Arcas showed how Photosynth had stitched together hundreds of tourist photos into an image of the Western Facade of Notre Dame. Then he clicked on the snarling gargoyles hanging above the main archway. The software zoomed in on that section—Photosynth had quickly selected the best Flickr photo of the gargoyles and smoothly shifted the 3-D scene to show where that single snapshot appeared on the face of the cathedral.
The demo seemed magical, conjuring a whole new way of collecting the thousands of photos that each of us now produce during our lifetimes: Imagine being able to review your Grand Canyon vacation not as a static slide show but as a tour of the 3-D environment produced by your and everyone else's pictures. "This takes data from the entire collective memory of what the earth looks like," Aguera y Arcas told the crowd. "All of those photos become linked together, and they make something emergent—greater than the sum of their parts."
Microsoft launched Photosynth last summer, but the Web application had its real coming-out party at the inauguration. Though no one has really counted, you probably wouldn't be too off the mark in guessing that the 1 million or 2 million people who gathered to watch Barack Obama take the oath of office produced tens of millions of snapshots—making it one of the most photographed events ever. As part of their inauguration coverage, both CNN and MSNBC invited viewers to submit their photos to Photosynth. These contributions produced dozens of 3-D scenes—called "synths"—at various locations across Washington during inaugural weekend. They're absolutely mesmerizing.
For instance, here's a synththat combines more than 100 photos taken at the Lincoln Memorial concert on the Sunday before the inauguration. (You might need to download a special viewer to get these synths to show up in your browser; they worked fine for me on a PC using Internet Explorer 8 and Firefox 3.05, but your experience may vary.) You start off with a low-res picture captured by someone way in the back—the memorial can be seen as a tiny square in the distance, far beyond the reflecting pool and a sea of people. But you can move deeper into the picture—click around the center of the image, and you zoom into a shot taken by someone a few hundred yards closer to the action. Now you can see the memorial and the reflecting pool, but you're still pretty far away. Click once more, and you move into someone else's snapshot, in which you can just make out Lincoln's statue. And on and on you go: As you click, you fly deeper and deeper into the crowd, flowing through photos taken by folks successively closer to the stage. Eventually you come to a sharp, detailed picture of Obama and Lincoln alone on a vast platform; you're close enough now to read the memorial's inscription. But, of course, people who were this close didn't just snap photos of the stage—they also turned around to capture the throngs of people behind them. Photosynth lets you do so as well. Click to the left of the picture, and you see the other direction—the scene Obama saw from the platform.
There is something vaguely embarrassing—even narcissistic—about our new era of mass photography. Because we're always carrying cameras, we're moved to document every moment of our lives—sometimes to the exclusion of actually experiencing that moment. Take a look at this picture of Barack and Michelle Obama at one of the inaugural balls. Everyone in the audience has a hand up with a cell phone pointed at the stage, but nobody is actually looking at what's going on. The scene is puzzling: If the guy next to you is taking a picture—one that you can be reasonably sure will end up on a photo-sharing site somewhere—why do you need one, too? But we do this often these days. Win Butler, the lead singer of the band Arcade Fire, once told Terry Gross that he and his band mates have stopped going out into the crowd to perform because nobody pays attention to them—everyone's got their cell phones and cameras in front of their faces.
The genius of Photosynth is that it takes advantage of the psychosis of digital photography. The software collects similar pictures that, by themselves, aren't very interesting and builds an entirely new kind of image, what you might call a social photograph. Ordinary photos show you an event as one person captured it. A synth is a scene as seen by everyone. Look at this synth assembled by CNN, titled "The Moment." The network asked people to send in their snapshots of the moment Obama held up his hand and took the oath. There are now nearly 700 pictures in the synth, and together they give you an incredibly detailed picture of what took place that morning. It's not just one person's view—you can see the event as people in every part of the crowd saw it. There are sharp photos from CNN's professional photographers and blurry pictures from people using cell-phone cams, all shot from every conceivable angle around Obama—his left and right, below him, even behind him. Looking at the synth is even better than having been there live—this way you see everything.
Five years ago, I spotted a pair of images online that forever shook my faith in photographs. The pictures showed an American Marine, Lance Cpl. Ted Boudreaux, standing with two smiling Iraqi boys in the desert. The photos were identical, except for one detail. In one, the boys held up a cardboard sign on which someone had scrawled, "Lcpl Boudreaux saved my dad. then he rescued my sister!" In the other picture, the boy's sign read, "Lcpl Boudreaux killed my dad. then he knocked up my sister!" The military had investigated Boudreaux for misconduct—they suspected he'd been making fun of the kids by forcing them to hold up a sign they didn't understand—but Boudreaux proclaimed his innocence. I contacted a half-dozen photo analysts, and none could say definitively which photo was real and which was doctored.
Of course, people have been questioning the authenticity of certain photographs ever since photography began. But many professional photographers I spoke to saw the mystery of the Marine as a sign of the modern-day decline of photography as a journalistic pursuit—in the Photoshop age, you could dismiss any photo as false, which meant no photo would stand as proof that something actually occurred. Fred Ritchin, a former photo editor at the New York Times, told me at the time: "You can say Tiananmen Square happened—there was a video, there was a massacre. But if we typically disbelieve the evidence of a photograph, then when the Chinese government says there's no massacre, what are you going to hold up against that?"
But there was one person I spoke to who was more optimistic about photos entering the digital age. Pedro Meyer, a Mexican photographer who's known for his liberal use of Photoshop, pointed out that people around the world now snap billions of pictures every year. True, any single photo can be doctored—but in the future, Meyer believes, we'll get at the truth of what happened at any particular event not by looking at one photo, but by looking at the amalgam of hundreds. How will you know whether a massacre occurred if a tyrant Photoshops the evidence? Because we'll be able to combine the many undoctored pictures—even if they're all blurry cell-phone pictures—into a composite picture of the scene.
At the time, I was skeptical of Meyer's idea; it seemed a bit dreamy to think you could patch together small pictures into something resembling the truth. Now that I've seen Photosynth, I'm a believer.