Slashdot Mirror


350 Megapixel Camera

Remy Hathaway writes "Ars Technica just posted an article on the "MegaCam", a 350 megapixel camera. The original story is from the Honolulu Advertiser."

34 comments

  1. oops by Kevin+Stevens · · Score: 1

    well actually... its called 'megaprime' according to the article...

    1. Re:oops by cyb97 · · Score: 1

      i guess somebody didn't read the article...

    2. Re:oops by Kevin+Stevens · · Score: 1

      oops, well I only read the ARS article (seems like a better source than the 'Honolulu Times' ) and it only mentions 'megaprime', not 'megacam'... I guess its really not a big deal either way.

    3. Re:oops by cyb97 · · Score: 1

      It's called mageprime in both articles... I guess somebody (read: slashdot-editor) didn't read the article before posting...

    4. Re:oops by Anonymous Coward · · Score: 1, Funny

      "It's called mageprime in both articles"

      its name changes by the minute... what next? "mepagrime"? "grimey pap"? "papery gimp"? "preppy game"?

    5. Re:oops by Anonymous Coward · · Score: 1, Funny

      "Megaprime"? Wasn't he one of the Transformers?

    6. Re:oops by mdielmann · · Score: 2, Insightful

      Actally, if you follow the link to megaprime from ARS, you'll find that the camera is called MegaCam, from the MegaPrime project...
      http://www-dapnia.cea.fr/Phys/Sap/Acti vites/Projet s/Megacam/index.html

      --
      Sure I'm paranoid, but am I paranoid enough?
    7. Re:oops by matrix29 · · Score: 1

      Actually, if you follow the link to megaprime from ARS, you'll find that the camera is called MegaCam, from the MegaPrime project...
      http://www-dapnia.cea.fr/ ... /index.html (Link works, but Slashdot pushes spaces in words over 40
      chars

      Hmm... what I thought of instantly was the fellow that turned a flatbed scanner into a wide-field still image camera.
      Building a megapixel digital camera from a flatbed scanner

      And the later revisions in the concept by interested folks with science as their tool of genius.
      Industry always begins with hacking existing toys Thus the ultimate reason the DMCA is a BAD BAD BAD law.
      Improved Scanning Digital Camera

      I see the article is about an array of the imaging sensors with one bigass lens. I guess they'll get a quicker image this way.

      --
      "Face it, a nation that maintains a 72% approval rating on George W. Bush is a nation with a very loose grip on reality.
  2. Filesize of the pictures by neur0maniak · · Score: 5, Interesting

    Just imagine how big they'll be.. 350Mb(yes, megabits!) x 24bpp = 8.4Gb (Gigabits) = 1.05GB (Gigabytes). And that's without compression. How long would that take to compress, too?

    1. Re:Filesize of the pictures by p4ul13 · · Score: 2, Insightful

      I'm not sure about compression, but there is an X10 joke in here somewhere...

      Looking through the links I saw pleanty of images of the camera itself, but no actual images from the camera. Yes, yes, I know they're not going to put the +-1GB image on a site, but a largely scaled down and compressed sample for the gee-whiz factor would be nice.

      --
      Paul Lenhart writes words!
    2. Re:Filesize of the pictures by lfourrier · · Score: 5, Funny

      I think we need some breakthrough in technology. I've heard about some technic using a silver based substrat. The file is written in parallel, using some light encoding. Total writing of the file can take as little as 1/1000 second. Calculate the througput, it is very impressing. Compression is lossy, and it is more similar to cd-r than to cd-rw, and long term stability is not guaranted. But give it a few years, and I'm sure we will have a solution derived from this principle.

    3. Re:Filesize of the pictures by alonsoac · · Score: 1

      Even though the images are so big there is probably no problem in handling them since each science team will take home only a few images at most. I understood from the article that the images are wide views so unless they conduct a full sky survey you only need one image to see what you want. They probably just split it into several smaller images and then compress those.

    4. Re:Filesize of the pictures by Idarubicin · · Score: 2, Funny

      Not only that, but it could readily be made compatible with a large installed base of legacy technology.

      --
      ~Idarubicin
    5. Re:Filesize of the pictures by faxafloi · · Score: 2, Interesting
      From the project description at CEA:
      Each exposure will produce about 770 MB of data; the mosaic will be read out in about 20 seconds which means that Megacam will produce approximately 100 images (science fields and calibration) per night, ie 77 GB of data each night or about 1 TB of data for an average observing run.

      So it will likely be 16bpp, not 24. Astronomical images are usually FITS, not JPEG.

      Large images like this are becoming the norm in astronomy. Double the dimensions of a CCD and you quadruple the file size. With mosaiced chips like this one, you can easily get monster images. Then there's the processing, where you're usually juggling several similar-sized images. Looks like CEA is addressing this.

      Incidentally, if they did want to compress these, some lossy algorithms (wavelets, Starck) do well on astronomical images. Most of what you lose in those cases is the sky noise, as long as you don't select too high a compression factor. The DSS did very well with 10x wavelet compression.
      --
      Exit, pursued by a bear.
    6. Re:Filesize of the pictures by faxafloi · · Score: 1

      Even though the images are so big there is probably no problem in handling them since each science team will take home only a few images at most.

      But presumably they'll be archived (at CADC?). So someone will have to handle them.

      --
      Exit, pursued by a bear.
    7. Re:Filesize of the pictures by rabidcow · · Score: 1

      350Mb(yes, megabits!) x 24bpp = 8.4Gb (Gigabits) = 1.05GB (Gigabytes)

      No, megapixels.

      Think of it this way:

      bpp = bits/pixel
      pixels * bits/pixel = bits
      bits * bits/pixel = bits^2/pixel

  3. MegaCam???? by Red+Rocket · · Score: 1


    ...it's a .35 GigaCam !

    --
    - Hail to our fearless misleader! Fool speed ahead!
  4. oh my... by dauvis · · Score: 1

    Lord help us if the paparazzi gets hold of this one.

    1. Re:oh my... by jzs9783 · · Score: 3, Funny

      but then those pictures of Britney Spears having sex with a donkey while spanking Natalie Portman's ass will actually be clear!!!

    2. Re:oh my... by jo42 · · Score: 1

      Zoomable Porn!!!

  5. Additional info on their website... by stienman · · Score: 5, Informative

    For those too busy to go to their website:

    The imager is made up of 40 silicon imagers, each of which has two imaging surfaces on it, for a total of 80 channels of about 4.3 megapixels each. There are gaps between the silicon imagers. Images they make have black borders around them so that the image spatially is like looking through a window with bars. The imagers are rectangular, and are set up in a pattern somewhat like this

    .HHHHHHHHH
    HHHHHHHHHHH
    HHHHHHHHHHH
    .HHHHHHHHH
    The periods are due to slashdot's inability to do & nbsp;, & #160;, etc. The #'s I was going to use in place of H caused the lameness filter to spew "Too many junk characters." I guess I didn't realize I had a junk character account, nevermind that it is apparently overdrawn!

    It takes a full ten seconds to get the data from the sensors, and there is a rotating shutter above it. The time to take the image and then copy it off the array is long enough that they can only obtain about 1TB of data on a typical veiwing night.

    -Adam

    1. Re:Additional info on their website... by SeanAhern · · Score: 1
      The
       tag is your friend:

      ####
      ######
      ######
      ####
  6. god... by C21 · · Score: 1

    I can just imagine the way the noise reduction frames would look to clean those pics up...God, that camera had better be COOLED.

    --
    this is not a sig.
    1. Re:god... by _pruegel_ · · Score: 2, Informative
      Quote from their camera pictures page:
      Cryogenic The cooling is done using a "pulse tube", a closed-cycle cooler with no moving part, therefore no vibrations. The cold capacity is used to keep the CCD mosaic cold during the installation of the instrument at the prime focus.
      So yes, it is cooled.
  7. Megaprime by Strange+Ranger · · Score: 3, Funny


    So Optimus Prime can take pictures during battle?

    350 megapixels really is More Than Meets The Eye.

    --

    Operator, give me the number for 911!
  8. CCD mosaic by Anonymous Coward · · Score: 2, Informative
    they actually use CCD mosaic. There are 38 CCD chips of which 36 are usable, each 9 Mega pixel.
    Total resolution is about 18000x18000. At 0.2" visibility (the best you get anywhere on earth), this gives 1 degree FOV. It is large in total pixels, but each chip is not. Largest single CCD ever made is about 63 Mega pixel. Largest in 35 mm digital cameras are 16 mega pixels. Mosaics are suitable in astronomy where they don't mind a line here and there not exposed.



    The most important characteristic of astronomy CCD chip is back illumination which gives ultra high efficiency and extremely low noise.



    I believe, there is another project where they plan more than giga pixel mosaic. But forgot the project name or link.

  9. no big deal by g4dget · · Score: 1
    It's an array of CCDs, so it goes well with an array of machines. In different words, there is no need to push all those pixels through a single machine. The most natural thing to do would be to have a common trigger and timestamp the images, but to capture, compress, and store them on separate machines.

    You could get in the ballpark of this by buying 40 off-the-shelf 5 Mpixel consumer cameras and 40 laptops to hook them up to. Voila--instant "200 Mpixel camera", plus the bandwidth to process and store all those images.

  10. Damn review sites by JUSTONEMORELATTE · · Score: 1

    How come these digi-cam review sites never tell you how many pictures you can take with one charge of the NiMh cells?!?
    Pixel resolution is one thing, but it's worthless if the batteries die in the middle of my family's vacation!

  11. how many LOCs? by snot+whistle · · Score: 2, Funny

    how many LOCs per cubic hectare per fortnight is that? imagine a beowulf cluster of THESE babys! what a great day for pr0n! in soviet russia, they take pictures of you!! all of your pixel are belong to us! (this post is just a lameness filter test)

    --
    Where's Robin Hood? We could kinda really use him now.
  12. Skip the beowulf jokes but by phorm · · Score: 1

    The purpose of this camera seems to be taking pictures of large areas of sky, as opposed to getting superfine resolution on a particular area. Wouldn't an array of properly configured and aligned cameras serve much the same purpose? It seems that really all we're doing is compacting the components of many cameras and cramming them into one anyways?

    Still, I don't see this camera's technology coming out for home use - there's just no point. Could you imagine emailing home photos to relatives (photoshopping would also take forever and huge amounts of RAM). Rather, cameras with better colours sensors (as mentioned in previous /. articles) should be the future for home users.

    Now emailing dayatthebeach.jpg, estimated time 13d-4h-3m...