The Lytro camera: Too little for too much

My editor will be glad to know that this week, in contrast to the last few Gearhead columns I will not be discussing AT&T, ADSL (+ or otherwise), or Motorola DSL modems.

Nope, this week, something completely different: The Lytro camera, which is based on a really geeky, way cool, leading edge piece of research; light field photography.

Actually, it's only leading edge because the technology that makes it possible has only been around for a couple of decades. The original theoretical foundations were laid out way back at the turn of the last century.

What's so interesting about this technology is it not only has a place in the consumer world, but it will also be a boon in the business world simply by making it incredibly easy to take good photographs that will always be in focus.

The underlying principle, the light field, is the concept that a scene is comprised of all the rays of light, no matter what direction they are heading. A conventional camera captures only a restricted subset of these rays that are focused by the lens onto the imaging device in the camera, and while the color and intensity of rays are recorded, all of directional data about where the rays came from is ignored.

Conventional camera systems have a focal "plane", the point in front of the lens at which objects are perfectly in focus. The amount of light and other factors determine the "depth of field," the zone that extends in front of and behind the focal plane where, as far as the human eye is concerned, objects are apparently in focus. Anything outside of this zone will appear blurry.

A light field camera uses a very different imaging system that consists of, as Wikipedia explains: "an array of microlenses [placed] at the focal plane of the camera main lens. The image sensor is positioned slightly behind the microlenses. Using such images the displacement of image parts that are not in focus can be analyzed and depth information can be extracted."

So, by analyzing many tiny images and then using software to cleverly manipulate and combine them into a single image, a light field camera can capture, in effect, an enormous depth of field. This means an image captured by a light field camera can be used to create other versions of the image with, in theory, any focal plane and depth of field. So you could use a light field image to create versions where the foreground, mid-ground, or background objects are in focus. Even more intriguingly, at least in theory, a light field camera could even produce an image where everything in the scene is in focus!

This ability of light field imaging technology to produce "refocusable" images means that the ultimate "point and shoot" camera could be built ... no focusing, no aperture adjustment, just, er, point and shoot.

But, there's a potential problem. As the Wikipedia entry explains "The drawback of such a system is the low resolution that the final images have. As one microlens samples the light directions at one spatial point an increase in the number of image pixels can only be done by increasing the number of microlenses by the same amount."

Thus, the depth of field is dependent on the number of microlenses and the resolution is dependent on the number of pixels per microlens. So, if you increase the number of microlenses for a given sensor, the resolution drops but the depth of field increases. In other words, the resolution and depth of field are inversely proportional for given sensor geometry.

It's worth noting that the Lytro camera is specified as capturing 11 "Megarays" ... a wholly meaningless value as there's no definition anywhere as to what a megaray might actually mean.

Enough theory, let's get to the Lytro camera. First of all, the design is sensational. A small rectangular box (1.6 in x 1.6 in x 4.4 in) with a single large lens (f/2) at one end (with a magnetic lens cap -- very slick) and a touch sensitive screen at the other.

The screen end of the box is a textured, rubber-like surface while the lens end of the is a smooth metallic surface. The overall weight and balance of the device is perfect.

On the textured part of the top surface is a depression that you press to take a photo, and sliding your finger left and right above the screen optically zooms in and out. On the bottom surface at the screen end is the power switch and a pop-off cover for the USB 2.0 connector. Physically, that's it; a very clean design and very elegant.

The touch sensitive screen can be swiped left to review and manage photos (you can zoom in and out and refocus on any point) and swiped up to reveal the menu system. When you are taking photos In the "Creative" mode you can tap on the screen to choose the focus point.

I won't bother explaining how the menus work -- that's all covered quite well by Lytro's Web site.

So, the design is great but what of the results? In a word, disappointing.

The pictures are generally not very high resolution and display obvious "jaggies" and aliasing of dark objects in front of bright backgrounds as well as "fringing" under anything other than good lighting conditions.

Moreover, photos are not always in focus at all and refocusing on a specific area doesn't always produce significantly better results. In pretty much anything less than almost full daylight the noise level (visible as colored "graininess") is very noticeable. The company claims "By using all of the available light in a scene, the Lytro performs well in lowlight environments without the use of a flash." This is simply not true.

The software (it currently only runs on OS X) allows you to download and manage images from the Lytro camera (these are in a proprietary format) over USB, select what part of the photo is to be in focus, and upload images to the Lytro site where they can be refocussed dynamically by people viewing your work. You can also select the focal point in a Lytro image and export it to JPEG format though this feature is fairly clumsy.

Add to that the software is slow for almost every task (when it comes to exporting links to Facebook, painfully slow).

In fact, when it comes to pretty much everything a user really cares about in photography, the Lytro camera underwhelms.

One thing that really bothers me is that when you look at the images on the Lytro Web site you'd think that those are typical of what you're going to get -- clear, sharp, and "alive". The reality is far from what those images promise to the extent that I consider them to be false advertising.

So, bottom line; technically, conceptually, and physical design-wise, the Lytro camera is terrific, but the results -- the stuff users will really care about -- are, given the pricing ($399 for the 8GB, 350 picture model and $499 for the 16GB, 750 picture version), simply not worth it when you consider what you can get from a conventional camera for a lot less money. I'm giving the Lytro camera a rating of 2 out of 5 with a "nice try."

Light field technology will become a huge and redefining force in camera technology over the next few years. The Lytro camera is perhaps ahead of its time and delivers too little for too much. I'm sending mine back for a refund.

Gibbs is underwhelmed in Ventura, Calif. Your snaps to gearhead@gibbs.com and follow him on Twitter (@quistuipater) and on Facebook (quistuipater).

Read more about pc in Network World's PC section.