Photographers and cinematographers love lenses. We romanticize this part of our kit more than any other, going on about “the Cooke look” or the unique character of a vintage Canon with spherical elements. But the time may come when our big, heavy chunks of polished glass are replaced with arrays of small, unremarkable image gathering sensors that not only boast improved image quality over their traditional ancestors, but can be tuned to emulate any of their qualities.
What if you could choose any lens characteristic you desired, including focal length and even depth of field — without ever changing lenses? What if you could make these decisions after the photo was shot?
What if you could not only specify how much (or little) depth of field your photos had during post processing, but you could also change which part of the image was in focus?
What if you could simulate any lens, from an NASA Zeiss F0.7 50mm that Stanley Kubrick used on Barry Lyndon, to a tilt-shift, to a razor-sharp telephoto at ƒ/22? All with a single camera the thickness of a point-and-shoot?
This potential future of photography has a new player with a very fancy web site and a pre-order form. Meet the (website for the) Light L16 Camera.
When the Photo is Built, Not Taken
The Light L16 is based on a concept known as computational photography. What that means is that the final photo that it generates is not a 1:1 record of light intensities captured on a photosensitive surface, but rather a reconstruction, based on multiple imaging sources.
Remember the Lytro camera? It was the first commercially-available computational camera. Known as a "plenoptic," or "light field" camera, the Lytro features an array of microlenses directly on its sensor. A traditional lens focuses the scene onto this micro-lens array, and the sensor records many small images of the scene — one per micro-lens.
They key is that each micro-lens sees a slightly different perspective on the scene, just as a grid of cameras aimed at the same subject would. Software can use these multiple perspectives to build a depth map of the scene, and reconstruct a photo at a higher resolution than any one of the capture sites. This reconstructed image can have any depth-of field you like (up to the limits related to the full sensor size) and can even be re-imaged from the perspective of any of the micro-lenses, giving you the ability to create a paralax effect in post-processing.
It’s a bit like a police sketch artist creating a crisp likeness based on multiple blurry photographs. With enough of them to work with, the drawing can look nearly perfect.
You can see live examples of this madness at Lytro’s gallery. Hover your mouse over the images to see the parallax effect, and click to refocus the photo wherever in the image you like.
How Many Cameras Does it Take
Last year the HTC released the One M8, a smart phone with a second rear camera devoted entirely to recording depth information. Its built-in software allows adjusting the appearance of shallow DOF. Some examples I found were convincing, some not so much.
This two-camera approach was an improvement over previous attempts at this kind of thing in mobile phone photography, where multiple photos taken at slightly difffernt times could be combined to offer post-processing focus options.
If the multiple-imagers to make one picture thing is not exactly new, what’s special about what the Light L16 promises?
Choose Your Fetish
The Light L16 has 16 separate cameras at various focal lengths. Many are quite a bit larger than what you'd find in a smartphone, packed in sideways and firing through mirrors. The resultant arrangement of apertures looks almost insect-like. Depending on the zoom level you choose at capture time, ten of the 16 cameras are fired to gather material to merge into a final image.
What does this buy you?
I recently wrote about 3,000 words about how you might not want the overhead of the Sony a7R II’s 42.4 megapixels. The Light L16 promises stills of up to 52 megapixels. This is presumably possible because of a technique known as superresolution, where multiple samples are combined to create an image of higher resolution than any one of its sources. If you want to play with this technique today, try the Cortex Camera app for iPhone.
This isn’t truly an “optical” zoom, because the images are synthesized, but since some of the cameras on the L16 are genuinely longer in focal length, you should expect high (super)resolution results at any virtual focal length in that range.
A long zoom is hard thing to find in a pocket camera, and when you do see it, it’s always at the expense of my fetish:
Shallow Depth of Field
The L16 promises to be able to emulate a ƒ/1.2 aperture in its post processing. But don’t worry, you can decide just how shallow your DOF should be in post — as you see happening in Light’s lavishly-produced demo video:
Of course, shallow focus is just one option. MIT Technology Review’s Rachel Metz’s wrote about her visit to Light’s offices:
Light didn’t show me any working camera arrays, though I did see an image of one of the company’s engineers that was shot with a test array of four eight-megapixel sensors and combined with software. In a close-up of her face, her hair and the background were quite sharp.
That sounds terrible to me — but here’s one pocketable camera that promises Rachel her infinite DOF and me my crazy shallow focus. And I’m sure there are occasions where each of us want the opposite.
Because light field cameras offer the ability to select what’s in focus after the shot is captured, not only can you choose focus manually, but the camera’s decisions about what should be in focus could be made after the shutter is depressed rather than before, eliminating some delay.
And then in post processing, you grab a slider and make that final tweak to sharp-up the eye rather than the eyelash.
The L16 video shows how multiple exposures are combined to create low-noise images even in low-light situations. I’m curious as to why Light doesn’t play up the HDR possibilities here more, but low light performance is definitely nice, and also something that traditionally would come at the expense of the other features listed here.
Perfect focus every time. A 35–150mm equivalent zoom at ƒ/1.2. Great low-light performance and more resolution than a $27,000 medium-format Hasselblad. It’s easy to see why folks are excited — this camera seems to have it all.
Of course, it doesn’t exist yet.
The Light Plan
Light plans to ship the L16 in “Late Summer” of 2016 at a cost of $1699. You can pre-order now, and your $199 (refundable) down payment earns you a $400 discount on the list price.
Should You Order One?
A lot can happen between now and Spring. Light seems to have moved from a business plan of building camera modules for mobile phone manufacturers to include in handsets, to a selling directly to consumers. But their Plan A is, very likely, everyone’s Plan A. My estimate is that within two to four years, technology like this will be commonplace in mobile phones.
I’ve been hoping that would be true for a while now, so I might be overly optimistic. Apple can barely squeeze their single tiny camera into their current iPhone. Digital imaging maestro and former Apple employee Ron Brinkmann wrote about the possibilities of a bug-eyed camera array in an iPhone back in 2011 when the Lytro was announced.
It seems inevitable that the future of mobile phone photography is multi-capture computational, but it’s not clear exactly when that future will be. Light is promising to put it in your hands early next year. But I’d guess they’re also working on making their technology available to others at the same time.
What I Like
Lytro seemed positively obsessed with the idea that their cameras made a new thing called “Live Photos,” meaning the focus and parallax hijinks should be exposed to the viewer, rather than provided to the photographer as editing tools with the goal of producing a superior traditional photographic result. It’s fun to play with in their gallery, but it’s not photography. When I make a photo, I frame it. I decide what you see, and that includes focus. Focus is framing in depth. When you ask the viewer to chose it for you, it’s like beginning to tell them a joke, and then asking them to provide the punchline. It’s dissatisfying to them, and an abdication of your responsibilities as a photographer.
I like that Light seems to be dispensing with this technological showcasing and focusing on giving photographers tools to make the best pictures possible.
I also really like that whole thing about perfect focus every time, a 35-150mm zoom at ƒ/1.2, great low-light performance and more resolution than a Hasselblad.
It simply seems miraculous, and I want it all.
These Are Our Concerns
I am always suspicious of miracles in photography. At best, they may result images that appear to have had miracles worked on them.
We romanticize lenses in part because lenses are weird. Strange things happen with them. Have you ever noticed a background distorting along the edge of an out-of-focus foreground? Or how heat ripples refocus light? How spectacular (and hard to fake) anamorphic lens flares are? Or how sharp silhouettes appear in big circles of boke? Look at this shot from Nocturne:
Notice how clearly your eye can resolve the silhouette of the approaching skateboard wheels, sharply outlined within the boke of the distant headlights. Those of us who fake reality for a living know how hard it is to generate this kind of organic, optical reality. This shot is a light sculpture, a freak of nature. It’s weird, and wonderful, and it’s actually communicating a lot of information in those blobs of light. I’m not sure you could make anything like it by reconstructing multiple images.
Not that the L16 would claim to — it’s important to note that the L16 is promising 4K video from a single-sensor only, so none of the fancy DOF or refocusing will be available in motion. Maybe light-field video will be next year’s pre-order.
One big issue I have looking at Light’s sample gallery is that there are very few examples of shallow focus. The few comparisons they present are very conservative — more so than what they show in the video. If the L16 can really emulate ƒ/1.2, we've yet to see it.
The sample gallery seems tuned to Rachel Metz's tastes, not mine. There's a lot of pop and contrast, and little indication of broad dynamic range, or sultry shallow focus. Light would do well to get some moodier, sexier samples into that gallery.
It’s worth thinking about about how a camera like this would integrate with your daily photography habit, which, for me, revolves around Lightroom. These photos obviously require Light’s own proprietary software for all the fun tricks, but I want it all — Lightroom’s color and organizational control, without giving up my ever-re-editable DOF and focus sliders.
I have one last concern, and it’s a big one. Next summer is soon. This camera doesn’t exist yet. Let’s have a little Shyamalan flashback to that MIT Technology Review quote: “Light didn’t show me any working camera arrays.” That was in April of this year.
An anonymous friend (and Lytro owner) shared his skepticism with me:
Acording to that same article, Light has partenered with Foxconn to finance and manufacture the L16. Still, my sense is that we'd all be wise to treat this pre-order situation like any crowdfunding campaign from Kickstarter or the like — with the proviso that Light promises to refund your down payment if you ask in time.
Pre-Ordering the Future
Today we get to decide if we want to pay now to potentially play with the far future in the near future, or just wait for the future to arrive on its own. Sometimes the future is amazing. Sometimes, when it arrives too early, it winds up sitting on a shelf — like my friend’s Lytro.
Even with the $400 discount, $1,300 is a lot to spend on an experimantal camera that doesn't integrate into your existing workflow. On the other hand, what the L16 promises cannot be had at any price today. What would you expect to pay for a F1.2 35–150 zoom lens?
For now, Light has $200 of my money. And I have an invitation to visit them and learn more in person. So this won’t be the last you read about the L16 here.