Lytro

Remember that camera that supposedly captures enough quantum hoo-ha that you can re-focus your pictures after the fact? Apparently it's actually shipping:

Since you'll capture the color, intensity, and direction of all the light, you can experience the first major light field capability - focusing after the fact. Focus and re-focus, anywhere in the picture. You can refocus your pictures at anytime, after the fact.

What's missing? Nothing you'll miss. No auto-focus, no shutter lag, no unnecessary modes, dials, or settings. And no flash, because Lytro can handle many low light settings. So, no obstacles to the perfect shot.

I'm curious how it works in the dark.

Tags: , , ,

27 Responses:

  1. Not that Jamie says:

    I'm intensely curious about this thing. As someone with way too much camera gear and nowhere nearly enough free lunches under my belt, I think there has to be something fatal, somewhere.

  2. Vincent Janelle says:

    Only ships in the states, and not until 2012.

  3. jwm says:

    Curious that the lack of mention of exposure and white balance control, lack of quantification of it's light sensitivity and the phrase “And no flash, because Lytro can handle many low light settings.” all say that this is a happy snapper camera, yet they bother to mention that the lens is f/2 along the entire zoom range. Surely in a camera that's designed to allow you to set an arbitrarily shallow depth of field anywhere you like in post, this is entirely beside the point?

    Still, interesting to see if gets exposure right and is fast enough for kids and sports.

    • gryazi says:

      There'd be less of a point in having the magically rigged sensor without a shallow DoF (now someone who actually understands optical physics can cough up the term I'm looking for here). So opening it up gives you both more light-gathering as a benefit of the magic trick and more opportunity to actually see the effect of the magic trick.

      I'm thinking the sensitivity can't be much worse than other cameras, since the subpixels should be the same size as the pixels on an equivalent model - you're just sacrificing resolution in exchange for magical-postfocus, and the pixels you're actually using should be receiving basically the same amount of light as they would when a conventional lens is properly focused on the same sensor without the microarray, right?

      • Without the f/2 lense, you won't have a shallow depth of field. If you have a lense that stops down as the optical length increases, you lose the ability to have areas out-of-focus.

        On the other hand, with the light field capture (which I personally think is a crock of shit, and not real photography) you would easily be able to emulate depth of field by combining multiple focii (that is, get images with different points in focus and composite them); perhaps this is even allowed for in the camera's software.

        • gryazi says:

          Find the last post about this stuff and you'll find someone linked to a PDF from one of the academic groups that came up with the whole idea.

          Real photography involves capturing an interesting image. HTML isn't a real manuscript painstakingly handlettered by monks but I don't hear much complaining (and anything that reduces the urge to run to Photoshop to completely fabricate pixels in search of a pleasing result probably isn't a net loss).

          • jwz says:

            HTML isn't painstakingly handlettered by monks

            Excuse me, speak for yourself!

          • That's exactly my point. Real photography is about capturing an interesting image. Not capturing "something" (a light field) and then making it interesting when you go back to your Macintosh, for shitty redistribution on facebook at a resolution that frankly makes my old Apple QuickTake look high-tech.

            • Ben Brockert says:

              In any online photo there is a process of going from photons in real life to photons out of a display. That process can include any set of lenses, chemistry, types of light sensors, computers, algorithms, math, and prayer.

              Arbitrarily deciding that all processing of the image be done on a computer inside the camera is just that: arbitrary. Is cropping a photo more authentic if you do it in the camera's processor using the back display than doing it on another computer afterward?

              • I don't know about you, but I don't pray that my HC-110 works the same way it always has when I visit my darkroom.

                While it's true there's always a process, I dislike the trend towards massively reworking large swathes of image on the computer. I prefer minimal intervention between camera and end result - at most, generally some contrast/levels adjustment and (if necessary) denoise. Or, if I'm using film, I generally don't do any adjustments.

                Good photos will stand out even without being molested in the darkroom or Photoshop. Stunning ones sometimes need to be massaged to get from "good" to "stunning" - but let's face it, the target audience of the Lytro is likely neither capable of nor aspiring to stunning photos. It's a happy snap camera. It's not designed for Real Photography, and I refuse to acknowledge it as a tool for such.

                • Ben Brockert says:

                  If there was a One True Film that perfectly represented the light that fell upon it, and a process to display that exposure that was similarly perfect, not doing adjustments to it would be a meaningful thing to say.

                  But there isn't a Film, there are a whole lot of different films, and after that there are a lot of different papers, projectors, and scanners. Each one of those is essentially running a filter on the exposure, massively reworking the image to give a (hopefully) specific response to different intensities and wavelengths of light, as well as adding granularity.

                  I really like the look of gelatin prints from large format negatives, but it's no less 'molested' than a shopped shot out of a digital rebel.

  4. gryazi says:

    Huh, this explains the otherwise-insane form-factor (and don't ask me why it's https):
    https://www.lytro.com/science_inside

    ...and otherwise admits to what other people in the last thread found - the 'quantum hoo-ha' boils down to slapping an extra set of lenses across, say, a 14MP sensor, to get your 1MP image by choosing which pixels (now subpixels) to use. Which is a perfectly reasonable (completely awesome) application of Moore's Law.

  5. Vivek says:

    Thom Hogan has the most informative analysis: http://www.bythom.com/

    Basically, you have a something like a 720px X 720px camera that doesn't need focusing. AND, and, the EULA appears to put some limits on what you can do with the photos!

    • nix says:

      I think Thom Hogan is reading the terms of service for photos published on lytro.com and concluding that Lytro gets rights to your private photos too. Seems unlikely.

      • sherm says:

        Unless the camera is storing data in some kind of proprietary, patented format and it requires their (surely not shitty!) app or Flash monstrosity "player" to actually get a picture out of it. The marketing-speak seems to suggest that's what they're thinking (LET YOUR FRIENDS REFOCUS YOUR PHOTOS!!!).

        • gryazi says:

          Even if they fuck that up it's not like you aren't going to be able to retrieve a still by screenshot.

          At least until they scale past screen resolutions.

    • tegeran says:

      What in the name of...?! That site is an utterly incomprehensible mess, and once I finally realized where the "informative analysis" was, I found it completely uninformative. And nix is right, he's interpreting terms that are specific to their web service, and are hardly abnormal for such a service.

      Read the actual terms and it's quite clear. It even says that the license granted to Lytro is "with respect to content that you submit to Lytro.com".

      • tegeran says:

        By the way, the thing about requiring lytro's explicit concent is *explicitly in reference to Lytro's own content*,

        Lytro.com contains both Lytro content and user content. The content on Lytro.com and any authorized copies are licensed, not sold. Such content is protected by copyright laws and, if applicable, by international treaties.

        Lytro Content

        Lytro content is the content posted by us on Lytro.com. It may include, for example, living pictures, photos, videos, text, graphics, and other materials we post from time to time. Subject to your compliance with these Terms of Use, we authorize you to:

      • Vivek says:

        You'll have to read into this part a bit:

        What's missing in the details are a couple of things: angle of view of the lens, the number of focus zones, and the final pixel count of the images. What we get instead is a new marketing term: 11 Megarays. What that means in real pixels is currently unknown, but my best guess so far is something like 8 focus zones at a resolution of maybe 720x720 pixels. Wait a second, what's a focus zone? Well, the way a Light Field camera works is by collecting information about how the light gets to the final sensor. A microlens set far forward of the sensor is the primary manner in which that information is obtained. But it means that you have multiple photosites on the sensor that are getting information from a microlens. The number of photosites that collect a microlenses information determines how many light angles you capture, and that number will determine how many discrete focus positions you can mimic, what I'm calling focus zones. My 720x720 guess comes from looking at sample images and what was said at the announcement ("equivalent to HD"). The math says an 11mp sensor would have 16-20 photosites per microlens, and a 16-photosite Light Field array makes sense to me.

        This makes it clear that the final resolution of the image in the Lytro approach will be locked to about 1/16th of what the general state of the art of the day for the given sensor size is. Focus is a non-issue for any application which doesn't need more than 720px X 720px today. For small images, whether from point-and-shoots or phone cameras , focus is a solved/non-existent problem.

        It's interesting that Lytro has solved this problem, but no one was looking for the solution.

        • gryazi says:

          re: focus, have you ever actually looked at a television ever?

          720x720 will indeed be a little sad but as soon as this scales to actual HD ... well, 1920x1080 is a standard that's going to be with us for a decent while, and looks pretty good-enough at poster size from a reasonable viewing distance, let alone ~ 8"x10".

        • nix says:

          Their first product doesn't look very compelling because they are making a big deal about focusing in post. Whatever, most people can't even be bothered to recrop.
          The exciting thing here is the absence of moving parts. Sensors are getting better fast, while shutters and mechanical focusing systems will improve relatively slowly.

  6. Sam Carter says:

    By "actually shipping", I think you mean that they are "actually taking money, but not shipping until 2012." This is still vaporware (try to click on a Buy link).

  7. chao says:

    May future films be shot with this? And it would be more fun than just 3D.

  8. trollop says:

    They are rather vague about the output. I had signed up to get the first shot at buying one a while back. And I was toying with that notion. Their examples aren't very big, plus the "focusing" isn't in clear sharp focus. So I asked them what the output format is, and if the examples on the site are as good as it gets with regards to sharpening. And also is it true that images will live on their servers.
    Here's their reply:
    "Thanks for your email. This version of the camera is designed for people who enjoy sharing pictures on the internet via Twitter, Facebook, and embedded images in blogs and email. Yes, the pictures are hosted on our site. The pictures you see in the gallery is the quality of pictures you will be capable of making. This camera is not designed to replace professional cameras or to take professional quality shots. It's geared towards casual use. This is something you should consider."

    I am surprised that you don't keep your own images. Are they going to be filtering them for content? Plus one assumes that if they don't succeed, then you loose all your images.
    So I guess I'm not the target market. I hate facebook. I am perfectly happy with my iphone camera for this my internet photo purposes-plus you can print a good image from those! And I don't really want to carry around an extra gadget to take pictures. $400 seems like a lot for a toy camera that doesn't even let you keep your own images.

    • trollop says:

      I got a follow up email about the output... Turns out it does jpeg at least 1080x1080. Also it only lives on their servers if you want to share them. So that seems a lot more reasonable.