Core77

Lytro's "Light Field" Camera Adds Fabled Third Dimension to Photography

By Ray Hu - June 23, 2011

in Tech

Well, almost: Lytro, a Silicon Valley startup, has been getting quite a bit of buzz for its new "light field" camera, which allows photographers to manipulate the focal length for any given image after the photo has been taken. While the technology has existed since the mid-90's, Dr. Ren Ng has taken his lauded dissertation research from the hallowed grounds of Stanford to nearby Mountain Valley, where he's managed to adapt the imaging technique—"light field" once meant some 100 cameras in a room— for consumer use under the Lytro moniker.

And what, exactly, is a "light field," anyway? The short description is that it captures every ray of light, deflecting off every object at every angle, in any given image. Where traditional camera lenses "simply add up all the light rays and record them as a single amount of light," the "light field sensor captures the color, intensity and vector direction of the rays of light."

The light field is a core concept in imaging science, representing fundamentally more powerful data than in regular photographs. The light field fully defines how a scene appears. It is the amount of light traveling in every direction through every point in space—it's all the light rays in a scene. Conventional cameras cannot record the light field.

via Ng's dissertation (PDF)

Frankly, I'm not much of a photographer myself—though I do like to play around with manual focus from time to time—so I'm curious as to whether Lytro truly represents a step towards leveling the playing field for the art of photography: light field aside, one must still compose an image. Or is it the case that Lytro is introducing a much-needed element of composition to the point-and-shoot mentality that previously lacked this sense of refinement?

Michael Soo for Lytro

So is it a game-changer? We'll find out later this year, when Lytro releases more details about the what is already being hailed as the most significant advance in photography since the shift from film to digital (if not the 1800s). In the meantime—while Ng & co. put $50 million of fresh venture capital towards commercializing Lytro—check out some of the "living photographs" on their site (above and below):

Eric Cheng for Lytro

o

Favorite This
Q
8

Comment

Comments

Mark

13 years ago

Z

Z

Reply

I'm skeptical so far, and lots of the marketing material sounds like marketing weasel doublespeak. I have a 12mm lens which, at infinity, keeps everything outside of 1.5 feet from the camera in focus. I can selectively blur things in photoshop to simulate a focal plane.... not sure if this is what they're doing, but it could be that simple...

The question I have not seen answered is how close can the camera focus and what is the effective aperture of the "lens"? Absence of these details make me suspect the technology isn't that great...
!Report as spam
Mark

13 years ago

Z

Z

Reply

I'm skeptical so far, and lots of the marketing material sounds like marketing weasel doublespeak. I have a 12mm lens which, at infinity, keeps everything outside of 1.5 feet from the camera in focus. I can selectively blur things in photoshop to simulate a focal plane.... not sure if this is what they're doing, but it could be that simple...

The question I have not seen answered is how close can the camera focus and what is the effective aperture of the "lens"? Absence of these details make me suspect the technology isn't that great...
!Report as spam
CD

13 years ago

Z

Z

Reply

This is amazing. This will change everything. Give it some time.
!Report as spam
Berkana

13 years ago

Z

Z

Reply

I would love to see the eventual application of this light-field technology to video cameras. Imagine capturing video you could re-focus at will long after the video has been captured.

It would probably consume memory and processing power like nothing else, but it would be awesome.
!Report as spam
Dusty

13 years ago

Z

Z

Reply

You guys ought to read his dissertation where he describes in detail how it works. Nothing altogether revolutionary in the hardware, it's just an array of microlenses in front of a single conventional imaging chip. The easy part is that it forms a conventional image, and the processing then selects certain pixels from each micro-image to achieve the final image. The cost is that your final image resolution is just a small fraction of the raw sensor resolution.

For instance, their prototype used a 10MP medium format sensor, which produced a 296x296 pixel processed image (each microlens site created a 12x12 pixel "image"). Scale that up and you'd need a 120MP sensor to create a 1MP image (1024x1024) with the same information per pixel (12x12). They propose manufacturing medium/large format chips with the pixel density of compact point-n-shoots to get up to 250MP on a single chip.
!Report as spam
Mike

13 years ago

Z

Z

Reply

So the company is going to basically reduce 100 cameras to a single camera? In think version 1 of this effort will result in a new camera the size of a 15 in. monitor. I'm waiting for at least version 2.
!Report as spam
Andy

13 years ago

Z

Z

Reply

So a conventional sensor captures 24 to 42 bits of information per pixle, and this technology is going to.... Capture an entire image resolution worth of infomation per pixel so that it can be post-processed to simulated lense specifications? That sounds ludicrist.

So instead of doing the impossible, what's left is lytros implementation which is going to what? use optics? and capture a much more limited field?
!Report as spam
J~

13 years ago

Z

Z

Reply

I can't wait to see how much memory each photo takes and other specs like the size of the camera, flash or no flash etc.
j~
!Report as spam