The Lytro Camera: Impressive Technology and Some Big Drawbacks 220
waderoush writes "The venture backers behind Lytro, the Silicon Valley startup that just released its new light field camera, say the device will upend consumer photography the way the iPhone upended the mobile business. This review takes that assertion at face value, enumerating the features that made the iPhone an overnight success and asking whether the Lytro camera and its refocusable 'living pictures' offer consumers an equivalent set of advantages. The verdict: not yet. But while the first Lytro model may not an overnight success, light field cameras and refocusable images are just the first taste of a revolution in computational photography that's going to change the way consumers think about pictures."
DPReview has a review (Score:5, Informative)
DP Review [dpreview.com] has a review of this camera. It sounds like it has a long way to go. Due to the way lightfield works, the final resolution is fairly low, in this case only 1024x1024. I don't know if there's really a way around it, since they're substituting resolution for the depth of field focus feature.
Re:Of two minds? (Score:5, Informative)
Re:A pity... (Score:5, Informative)
Unfortunately, though, the move to release it at a (barely) 'consumer toy' price point really led to a product slightly too compromised to be useful: The optics you need for the light field capture eat so much of the sensor's available resolution that the resolution of the images you can get out of the thing is hovering slightly below 1 megapixel.
I'd love to see the same technology applied at a price point and form factor where the sheer sacrifice of available pixels wouldn't be so keenly felt.
The reason the camera is only 1 megapixel has nothing to do with the optics. The technology requires many pixels in the imager for each pixel in the resulting image. So, the CCD (or CMOS imager, I don't which it uses) probably has at least 10MP, despite the output of only 1MP.
It's a fundamental limit of the technology, and it'll be a while until we see more than 2 MP using it.
Re:New medium awaiting new aesthetics and explorat (Score:0, Informative)
WTF is a prosumer? Fucking marketers are destroying the language.
P.S. You suck.
Re:New medium awaiting new aesthetics and explorat (Score:5, Informative)
Why do we need "focus" at all? Why not have photographs where everything is in focus? Depth of field is an artifact of lenses, whether they're in your eye or in your camera.
Focus can be used in composition to guide the viewer to the important elements in the story. Just as "left", "up","down", etc. define the field of view, so does focus.
Simular effect with simple filter (Score:5, Informative)
The Lytro camera has special optics that basically separates the light entering the lens from different angles. Knowing the rough angle of the light rays allows you to combine them in different ways to change the focal length of the image, as opposed to a traditional camera, in which they are permanently combined as the CCD captures the light at a set focal length. This comes with a trade-offs as light from each set of angles is essentially captured as a separate image, giving you say 12x12 sub images on the CCD, so the resolution of each sub-image is much lower than you would get using the full CCD for an image.
Since Ren Ng published his seminal paper making the connection between refocusing a light-feild and Fourier Slice theory, there has been additional work which shows that you can achieve the same thing using a simple filter, rather than a whole new set of optics. The benefit of this is that it is cheaper to manufacture, and you can easily switch out the filter to adjust the trade-off between image resolution and depth of field, but come with an additional cost of a slight loss of total light (due to the filter). Here is one of those papers [umd.edu].
There are two basic approaches. The first heterodynes the light (a filter acts as multiplication) such that light that enters at different angles is shifted to different frequencies. So with this approach you get "subimages" in the frequency domain rather than the spacial domain, which can be seperated and recombined in software. The result and trade-offs are essentially the same but with simpler hardware.
The other is based on refocusing as a deconvolution operation, but the filter modifies the point-spread-function of the camera, such that it's frequency response doesn't have any zeros, so you don't loose data at those frequencies like you would with a simple rectangular aperture.
Re:New medium awaiting new aesthetics and explorat (Score:4, Informative)
Depth of field effects are considered part of the art of photography, much like amplifier distortion is part of the art of playing electric guitar. People pay a great deal for the capacity to get *narrower* depth of field: compare the price of Canon's 85mm f/1.8 and f/1.2 lenses. People most often buy the f/1.2 as a very very narrow depth of field portrait lens, rather than a very very low-light lens. Other lenses are known for the particular way that they throw backgrounds out of focus -- Nikon will even sell you one where you can choose exactly how the background is defocuses.
I think this trend in photography is overblown (I don't see the appeal of portraits where half of one eye is out of focus), but there's no doubt that artistic manipulation of depth of field is a big part of the art.
Re:A pity... (Score:5, Informative)
No you can't. You just think that because you don't understand how an aperture works.
Camera lenses focus by directing light through a small hole. At the point of focus, any light which bounces off an object then hits the lens will be directed in such a way that it hits the sensor in exactly the same place as it would have if it had bounced exactly at the center of that hole to begin with, meaning all light from that position hits the same place, giving a sharp image. Away from the point of focus, light bounces off the object, then when it hits the lens, it bends either too far, or too little, giving a soft edge. Thus when an image is out of focus, then the light projecting onto the sensor is actually wrong, no amount of sensitivity will fix that. This is why optics and focus have always been the most important part of getting a nice image out of any digital camera.
A light field camera fixes this by capturing the direction of the light and reconstructing an image of where the light actually came from, not just where it hits the sensor. Thus it can calculate a 100% in focus image covering the entire depth range without having to focus. Previously, only a relatively small range of distances could be kept in focus, and for that it was required to have a small apature and either a long exposure or a grainy image (cellphone style). Now you can have a sharp image with a wide range of focus without motion blur or grain and that's fantastic.
Resizing a 12 megapixel image into 1 megapixel will give you the same image, with less grain, exactly the same image as if you had stuck a 1 megapixel sensor in to begin with (lower resolution sensors of the same size format give less grain because of larger size per pixel and lower photosensitivity). It will never be any better than the image projected on the sensor to begin with, so it doesn't get you anywhere.
Re:New medium awaiting new aesthetics and explorat (Score:5, Informative)
IIRC, it's an 11 megapixel sensor, to get a 1 megapixel image.
So, not TOO far off from 4k video, to get a low HD quality Lytro video.