Thursday, October 08, 2015

Light Co. Announces its First Product

Light Co. introduces its first multi-aperture camera product. The L16 camera will retail for $1,699 and ship in late summer 2016. A limited quantity will be available for pre-order through November 6, 2015 at a special price of $1,299. The camera offers:
  • Adjust focus and depth of field even after a photo is taken, all the way to f/1.2
  • Built-in 35-150mm true optical zoom
  • Exceptional low light performance
  • Low image noise
  • 3 fast prime lenses
  • DSLR quality images up to 52MP
The company's Vimeo video talks about the camera design:




Update: Imaging Resource adds few more details about the L16 camera:

"L16 packs in sixteen 13-megapixel camera modules at three different focal lengths – 35mm (5), 70mm (5) and 150mm (6) – for a total of 52-megapixels. Each of these modules captures an individual image, which is then stitched together via software to yield a photo that Light claims is equal to the quality of a DSLR with a large lens attached."

12 comments:

  1. So basically it all depends on how precise is the depth information, If it's grainy and loaded with imperfections (like any other product out there.. nobody does it perfectt) nobody is going to use it because the post-processing like adding artificial lens blur will always come up with some nasty artifacts.

    ReplyDelete
  2. We will have to see what the camera reviews say about this camera performance, but my intuition tells me the price point is set way too high, like 2x too high, maybe 3x. At $300-$400 I would take the risk and order one. If it performs like a 52Mpixel DSLR with a high performance lens, then maybe the price point is ok. Do they do all the processing in the camera? What is the battery life in that case?? I have to assume it basically works as advertised or they will be dead after the first professional review. Anyway, good luck to them. Always good to try something new.

    ReplyDelete
    Replies
    1. They just want to tell Unicorn they are able to merge 16 cameras data to one, then they could sell themselves a fat check.

      Delete
    2. Agree 100%, would have loved to try it out was it priced more like an ordinary camera, rather than a DSLR. Also, as good as this camera might be, it will get measured relatively with respect to the best smartphone cameras, latest iPhone, Galaxy S, Xperia Z and Lumias. Considering how they are constantly improving, this has its task cut out. Mass market adoption will be hard for this given the price and that not many are willing to lug an extra device around (camera)...can they convince serious amateurs who use DSLRs to switch? I also hope they do well...its time someone saw an actual product based on computational imaging that appeals to a consumer :)

      Delete
  3. Very impressive feat, would love to see more pictures and video's.

    What ISP is being used for the processing is it custom (or multiple ISP's) or off the shelf.

    ReplyDelete
    Replies
    1. Would be very interesting to know which ISP, it must be custom. What off the shelf ISP (SoC) can process 16 cameras?

      Delete
  4. Nice idea. but... the light gathering is limited by lens diameter and angle of view. And this does not look good compared to a 'dslr'. This has 17 lenses; a DSLR has one lens.. A cheap DSLR has f/4@35mm and f/6@150mm. That is a lens entry pupil diameter of 9 to 25 mm. So this would need 17 lenses with an entry pupil diameter of 2.1 mm for wide angle and 6 mm for tele - for all cameras @correct focal length.

    So it may not be totally there yet with regard to signal/noise ratio. That said it may deliver nice out-of-focus blur and the center of wide angle pictures could be rather sharp.

    ReplyDelete
  5. ??? Completely way off what they claimed to promote:
    http://image-sensors-world.blogspot.de/2015/07/light-closes-25m-series-b-financing.html
    => a 60$ 52MP hardware for smartphones?!

    ReplyDelete
  6. At the end of the day you can do only so much with the processing, the math. certainly intriguing to have multiple focal points/exposures. But like many things, it is quality in, quality out. Can this approach rival a full frame sensor with quality glass? How would you get past the noise performance/issues of small pixels compared to high end sensors with huge comparative full wells. You still have the limitations of the smallest resolvable spot size vs tiny pixels. Is there a real break through in physics?

    ReplyDelete
  7. I'm getting so tired of this. So many start-ups think up some weird new camera/lens system, and try to sell it to mobile phone and DSLR makers. Both are too risk-averse to adopt the technology, leaving the start-up without any significant customers. Rather than rethinking their market focus, the start-up management decides to try and sell their own consumer product. It NEVER works. They always discover that making a product is a lot harder than they realized. This blog is littered with companies that have tried this, and they've all ended up nowhere.

    ReplyDelete
  8. Agree with the light gathering ability commented above. I believe the light coupling efficiency of multiple lens/sensors vs a single lens/sensor will be inferior. Coupling efficiency drops with the square of the F# and the square of the demagnification (smaller sensor -> higher demagnification).

    ReplyDelete
  9. The first part of this video is interesting, although it seems they are still showing mock ups and not an operating camera.
    http://petapixel.com/2015/10/16/light-ceo-dave-grannan-on-the-new-l16-camera/#disqus_thread

    Also, at the end of the article, they claim a lot of presales. Hard to say if this is BS or not but perhaps there are a lot of early adopters out there with money to spend. Shipping August 2016, they say. Sounds like they still have some work to do.

    ReplyDelete

All comments are moderated to avoid spam.