How the Lytro Light-Field Camera Works

Light-field photography has the potential to revolutionize photography. Ren Ng, the founder of Lytro, explains how a camera can capture images that are never out of focus.

Lytro has big plans. This week the Mountain View, CA-based startup said it would soon bring to market a new kind of camera that's based on light-field photography. The result: photographs that you can focus after you take them. Simply click your mouse on the spot on the picture you want in focus, and it changes before your eyes (check out the effect in the photo below).

Best Microsoft MCTS Training, Microsoft MCITP Training at certkingdom.com



The idea of light-field photography isn't new. Other companies, including Adobe and another Mountain View startup called Pelican Imaging, have dabbled in the technology. But the Lytro camera, assuming it debuts later this year as planned, will mark the first time the tech makes an appearance in a consumer camera. Its creator, Ren Ng (pronounced "ing") was inspired to make the camera when he found he couldn't capture properly focused pictures of his friend's five-year-old daughter. If a camera existed that could capture all the light information—direction, intensity, and color—of the entire field of view, he thought, it would make getting the "perfect" shot incredibly easy.

"If you can shoot first, focus later, it's going to be the fastest camera you've ever used," Ng said in an interview with PCMag. "Because when you press the shutter button, it takes the shot instantly. It doesn't have to wait for the lens to move."

That's the promise of Lytro's upcoming light-field camera, which the company says will be on store shelves by the end of the year. Besides taking pictures extremely fast, the camera can use the light-field information to create 3D images, and with so much data being gathered, it has improved performance in low light.

How It Works

It all sounds a bit magical, but it opens up a lot of questions. How exactly does it work? What are the tradeoffs, and how will these so-called "living photographs" interact with existing viewing and sharing software?

The basic premise of the light-field camera is to gather all data about the visible light in the camera's field of view so that software can manipulate the photo later. While the concept has been used previously to create imagery like the "bullet time" special effects in the Matrix movies, it required a room full of cameras, Ng says, and the power of a supercomputer. With special optics and sensors, Lytro has built the technique into a single, portable device.

"Regular photographs just don't tell the whole story. If you think about all the light that enters that enters the lens of a camera, that's much more than a photo. The light-field is all the higher-dimensional information that's lost in a regular photo. When we record all this information, that provides us the opportunity in software after the fact."

What happens "after the fact" is the big breakthrough: once the light-field data is captured, Lytro's algorithms can do some impressive tricks. First and foremost is the ability to focus on any point the viewer wishes.

"When a regular camera focuses physically, what the regular camera is doing is adjusting the lens relative to the sensor to bring different parts of the scene into focus," says Ng. "So if we have the whole light field, what we can do what that physical lens would normally have done, but in computation."

Beyond Focus: 3D and More

3D pictures work similarly. With complete light-field information, software can discern how a scene would have looked to two separate cameras (which is how most 3D is shot). The algorithms separate the light from the left and right sides of the camera to create the 3D effect. Ng says the result has improvements over conventional ways of shooting 3D.

"It goes beyond the 3D you see in the movies because we can also change the perspective in the scene. It's what would happen if you were standing at that scene, and you were kind of moving your head from side to side.

"For low light, all the light rays participate. We're using all the light coming through a large aperture to make a picture with a large depth of field—totally impossible with a conventional camera. To give a sense of perspective, there can be 16 times less light for [a specific] depth of field."