Poor focus often ruins otherwise great snapshots, but a new kind of camera from a startup called Lytro could change that. Its light field camera is able to take photographs, save them as data, then allow the user to change the image’s focus long after the photo was taken. The same technology also allows the production of sharp photos in low-light conditions without a flash, Lytro says.
[…]
The camera uses light field sensors to capture the color, intensity and vector direction of all the light rays in a scene, then uses algorithms to process them and let users change the focus as desired after taking the shot.
[…]
The Lytro light field camera “is the first work” that makes light field photography practical within the context of a real hand-held camera, and as such is a “significant development in photography,” associate professor Ravi Ramamoorthi from the University of California at Berkeley told TechNewsWorld.Photographs taken with a Lytro camera can be refocused even years after they were taken, company spokesperson Dugan said.
The focus of the photos can be switched seamlessly between 2D and 3D views, Lytro stated.
Notice that the web examples have only 4 focus depth zones, and that the largest image quality is about 300 x 400 pixels.
Get out the bullshitometer. Holding your breath for this one is going to be like holding your breath for Holographic Movies. You know; next year for the past 40.
It reminds me of the concept that “the universe is nothing but “information.” and that notion gets applied in all sorts of weird ways. This one actually makes sense to me. The “information” is there, unfocused. Just do some math. Same idea applies to just about everything in ways I don’t begin to understand–black hole evaporation for instance…..huh? Doesn’t photoshop do something like this for only slightly out of focus pics? Or does it just call it that and human suggestibility supplies the rest?
I use preview and multiple takes. Much easier.
A actual tech story? No way!
interesting technology not ready for prime time because of the resolution hit but as a proof of concept cool
It’s all about parallax over the diameter of the lens (or at least the iris stop), and it can’t be gotten rid of. If you believe this, I have some swampland, er, highly desirable estates for sale.
#5,
I’ve been mentally wrestling with that since last night. However, I’ll give this a chance to prove itself.
Remember, never say never.
I clicked through to the guy’s dissertation that explains in detail how it works. Unfortunately my rusty college level maths isn’t up to the job of understanding how he gets the final image, but his experimental camera (that apparently works) seems a simple enough setup.
He has a standard camera lens and a standard high resolution CCD sensor. The ‘trick’ is an array of tiny micro-lenses placed a few millimeters above the CCD.
Each of these micro-lens focuses a small image on the CCD, so the image captured is an array of tiny sub-images of the subject. The raw pictures the camera takes looks like a collage of a normal image (kind of like those collages that used to be popular of George Bush made from smaller pictures of dead soldiers).
How the tiny images are then recombined is completely beyond me and fills most of the guys thesis.
Anyway, to my simple mind, in a very round-about and overly-complex way all he has done is effectively taken hundreds of small images at various focal lengths. You can then select/recombine these tiny images into your desired final image. Hence the poor resolution.
Interesting story, but i was hoping for much more. Seems the holodeck is still some way off.
I’m interested to see if you could have it focus everything in the picture. Imagine you taking a picture with your friends in the foreground and still being able to see the background in sharp focus.
I think the resolution would initially be a problem. But 14-16 megapixel cameras are now cheap. So when 25-50 megapixels come down in price, and then this technology is applied, you’d end up with a 2-4 megapixel result, which for most consumers, is good enough.
Saw it on the news this morning. Need to do some research into this tech. Might be a scam or iffy marketing ploy.
Certainly an interesting bit of technology. But not ready for Prime Time. Not that that stopped anything from becoming popular. Witness laptops with 1 hour battery life, VHS, Windows 3.1, cell phones*, AOL, etc.
But when push comes to shove, it is really not much different than taking several shots in quick succession, each at a different focus, and deciding later which one to keep. Film photographers did the same thing with f stops and shutter speeds for years. It’s called “bracketing”. The difference is that as the technology needed to capture high quality multiple images very quickly and to store them economically improves every day.
It does remind me of speech recognition technology. I remember being a test subject for some speech software 15 years ago. Not much came of it. I think the designers and dreamers had good ideas, but were woefully ignorant of the vast, vast amount of computing power needed to even come close to making it work reliably. We MAY be getting there now. Not sure.
*Cell phones. They STILL DON’T MAKE RELIABLE, INTELLIGIBLE CALLS. Who cares if they can scan take pictures, search the web, make bank deposits or be used as a marital device!
Let me guess. It will be few years before I can afford a good one.
You can already buy six and eight core processors if you have the cash. The only thing holding back much bigger monster chips is almost nobody needs them so I’m sure the processor power will become available.
Pretty exciting technology. Great for small point and shoots and especially good for candids.
The lenslet array converts the camera into a wavefront sensor that can correct the aberrations in the image – all of them. That means it can change focus, correct for inaccuracies and misalignment in the lens – everything but exposure time.
It does work, but the resolution is very low compared to what we’re used to with conventional digital cameras. What they don’t show is that you can also bring the entire image into focus, not just the near/far parts.
The technology is ahead of its time, but it will catch on in the future with better CCDs, more memory, and faster processors. Give it a few years. Right now the company probably wants to get some experience (and exposure – hahaha) in the market before it ramps up production for real.