How an open source camera will change photography

How an open source camera will change photography

Summary: It seems amazing that no camera company has yet sought to build an ecosystem based on software, but this is an area where open source really can innovate, since every application will be a new one.

SHARE:

What Stanford calls the Frankencamera does not look like much.

It's big and clumsy and you don't get much better performance as a result of that.

(This is a close-up from our Crave blog. The hands are those of graduate student Andrew Adams. No word on whether he's related to Ansel.)

But because the new camera is based on a Nokia N95 smartphone, whose software is licensed by the open source Symbian Foundation, it can become a lot more.

Professor Marc Levoy plans to release a complete implementation for the camera in a year, a platform on which apps can be built.

Already he has created software for the camera that does things no commercial camera can do, like extend its "dynamic range" so all distances are optimally lit, and enhance the resolution of videos with still images.

The applications are endless, going well beyond hobbies.

Cameras that take pictures of speeders could have programs that enhance and re-take those images on cars that try to gray-out their license tags to avoid detection. Any attempt to evade photo detection might be automatically countered with the right combination of hardware and software.

It seems amazing that no camera company has yet sought to build an ecosystem based on software, but this is an area where open source really can innovate, since every application will be a new one.

Open source smart phone groups like Android, LiMo and Moblin should all be anxious to replicate what Levoy is doing in their phones.

One point Levoy did not make is that the Frankencamera software could be integrated with existing open source imaging software, like The Gimp, so developers of those programs can jump-start the ecosystem.

Levoy's idea is also so obvious it's impossible to believe a proprietary company can't adapt it quickly enough.

How soon before we see camera apps at the Apple app store?

Topics: Open Source, Hardware, Software

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

17 comments
Log in or register to join the discussion
  • Hobbist gimmic

    It's funny how open source always comes down to that annoying little detail called hardware.

    How nice when someone invests a couple million to start making a new toy for him to play with.

    Honestly, this is a non-starter. 99% of photographic needs are met with a commercial camera in conjuction with gimp or photoshop.

    There is nothing about photography to be changed here.
    croberts
  • It's all about the glass

    And just like the Gimp, this idea will give us software that's
    five years behind the curve. Commercial camera companies
    (the good ones anyway) are built on quality glass and quality
    sensors. A cell phone camera is a cell phone camera.
    MC_z
    • RE: How an open source camera will change photography

      One point Levoy did not make is that the Frankencamera software could be integrated with existing open source imaging software, like The Gimp, so developers of those programs can jump-start the ecosystem.<a href="http://ipadbagblog.com/"><font color="LightGrey"> k</font></a>
      zakkiromi
  • How does this help the camera company?

    I hear what it will do for "the community" but how does this help the camera company?
    GuidingLight
    • it trains engineers

      this is from an educational institution and it is aimed a researchers and
      educators.
      shis-ka-bob
    • Light dawns

      That's why it hasn't happened yet. But you can build a lot of software companies on an open source camera.
      DanaBlankenhorn
  • Optics make the camera

    While there is SOME innovation which can be done in the camera software, most of the same tasks can already be accomplished on the desktop after the image was captured. Either way, good cameras come down to great optics and sensors. After that, it doesn't matter much where the image processing happens. In fact, the same techniques he's putting inside the camera could (and probably should) also be developed as mini-applications for desktops.

    Then again, there is some appeal to a camera that hobbyists could "hack."
    BillDem
    • that is why this might matter

      There are all sorts of niches that could be filled in camera design. For
      example, I have a fabulous Zeiss 50mm f/1.4 lens that is unused
      because it uses the old 35mm film. I would love to see somebody
      combine a new digital camera with one of the 'classic' Contax, Leica,
      Canon or Nikon lens mounts. I like lenses that are glass and brass and
      that I have to focus by hand. If the software for a camera becomes open
      source, we are one step closer to a wider range of niche cameras. Canon,
      Nikon, Zeiss et al. are happy that I have to buy a new generation of glass,
      but I would love to keep using some of their old glass.
      shis-ka-bob
    • it can matter where the image processing happens

      Read this: http://graphics.stanford.edu/papers/adaptive-flash/ziegler-
      adaptive-flash-TR08.pdf

      This a a completely different idea than 'photoshopping' to correct for
      flash imbalances. This could be heaven sent for group photographers.
      And it only works if the electronics are built into the camera. The paper
      is from the research group that designed the Frankencamera.
      shis-ka-bob
  • RE: How an open source camera will change photography

    I agree that the frankencamera, based on a cell phone, is
    not likely going to cause any worry to Canon and Nikon,
    etc. But it does open an interesting possibility. What
    would happen if one of the big camera makers took a
    model that they were about to retire, and re-released it
    with its software open-sourced? Imagine the
    possibilities...

    I know that this would be a niche market, aimed at
    hobbyists and artsy types, but I would bet that there would
    be a strong market for it. Look at the "toy camera" craze
    that continues. If the camera was already developed, and
    had already paid for its development through years of
    sales, then the only additional cost would be the software.

    I can see the camera maker wanting to go through the
    existing proprietary software and removing the bits they
    don't want released to the public domain.... but it came
    with just the basic software to control it open sourced the
    rest would be added in. Just a thought.
    snberk341
    • Exactly

      Even if the initial market is small it's worth going for.

      The problem here is Moore's Law. Camera costs are falling exponentially -- it's basically two chips now.

      How do you maintain margins? I suspect this is an answer.
      DanaBlankenhorn
      • No Moore's Law for Cameras

        The control electronics in a camera may follow Moore's Law, but as the
        commentors note, image quality depends critically on the glass. Fixed
        focal length lenses have 'only' 6 elements (Canon EF 50mm f/1.8) and
        complex zooms can have up to 18 element (e.g. Canon f/2.8 IS 70-
        200mm L). For the lenses, servo motors, shutters, prisms and controls
        and you have nothing resembling Moore's Law.
        shis-ka-bob
        • Likewise, the sensors do not follow Moore's Hypothesis.

          In that the size remains almost constant. In fact, the bigger the sensor
          the better the image quality. Moore's Hypothesis only applies when you
          can shrink the topography to improve devices per unit area.
          Bruizer
    • Like the original Rebel.

      That camera was quickly hacked and functions added and changed to it.
      Bruizer
  • Uh...

    [i]How an open source camera will change photography[/i]

    Or not.
    Qbt
  • RE: How an open source camera will change photography

    I'd love to see a camera in which only the software necessary to run the camera, the other stuff left to post exposure processing on another computer .... to the degree that this is possible. This would simplify the actual picture taking I'm sure.
    timeven
  • a camera as a computer peripheral

    gosh! Just imagine, taking a picture with a camera, downloading it to your computer, optimizing it with image manipulation software then uploading to the camera the information to take the same picture just like you optimized it!

    This isn't the kind of thing that would be of much use to the point and shoot crowd.

    You would essentially be using the Gimp to program your own pre-sets. It doesn't get much cooler than that. Gimp even has it's own scripting language which would make it ideal for approximating objects in a depth of field. For security purposes or other applications where the camera is used in a fixed setting, a passerby or other variably present object could be made the focus of data collection saving storage space.

    And don't forget video. You could use a preset you made from a still and shoot an entire video segment with it. That would beat the heck out of manipulating the image frame by frame.
    valvestate@...