BREAKING NEWS: The end of JPEG is in sight

Discussion in 'Photoshop Tutorials' started by +/-, Sep 30, 2005.

  1. +/-

    +/- Guest

    Finally, JPEG is doomed, algorithm geeks unite! This is the quantum leap, no
    compromise technology. Young genius about to rock the DI world.

    WHAT'S NEXT
    Honey, I Shrunk the JPEG
    By Shailaja Neelakantan, September 21, 2005
    BUSINESS 2.0

    If downloading digital photos stalls your PC, spare a thought for the
    data networks in hospitals. A midsize hospital typically gets 60 requests
    every hour for MRIs and echocardiograms. At 10 megabytes apiece, the
    enormous images can quickly cripple a network.

    Enter 25-year-old Arvind Thiagarajan, co-founder of Singapore-based
    startup MatrixView, who wants to revolutionize digital imaging. The
    data-compression algorithm he invented shrinks images into a format called a
    MatrixView Universal, or MVU, which is 15 to 300 percent smaller than a
    JPEG. But unlike a JPEG, which omits details, an MVU is as precise as the
    original. "Data loss is unacceptable in medical diagnosis," Thiagarajan
    says. That's why the startup is focusing on health care first. A well-known
    hospital in Bangalore is using the technology, and MatrixView plans to ink
    deals in the coming year with several Fortune 100 health-care companies in
    the United States. MatrixView is also targeting other subsets of the $9
    billion U.S. digital-imaging market. Right now it's negotiating with
    chipmakers to embed the technology in cameras and fit more files on storage
    cards. MRIs today, vacation snaps tomorrow.

    http://www.business2.com/b2/web/articles/0,17863,1106847,00.html

    http://www.matrixview.com/

    Download white paper:
    http://matrixview.com/files/ABO white paper.pdf
     
    +/-, Sep 30, 2005
    #1
    1. Advertisements

  2. +/-

    Matt Ion Guest

    Unless it's freely available to all developers, and doesn't include some
    cockeyed protection scheme that makes it difficult for one to
    backup/edit/copy one's own pictures, it'll never fly for the mass market.


    ---
    avast! Antivirus: Outbound message clean.
    Virus Database (VPS): 0539-2, 09/29/2005
    Tested on: 9/29/2005 11:16:05 PM
    avast! - copyright (c) 1988-2005 ALWIL Software.
    http://www.avast.com
     
    Matt Ion, Sep 30, 2005
    #2
    1. Advertisements

  3. : Unless it's freely available to all developers, and doesn't include some
    : cockeyed protection scheme that makes it difficult for one to
    : backup/edit/copy one's own pictures, it'll never fly for the mass market.

    Also nothing is said about processing time. If a sufficiently complex
    program is run, it is very possible to compress any photo much further
    than any currently used photo image. For the compression of an x-ray or
    MRI is "fast" if the result shows up in a few min while the dr and patient
    are walking back to the dr's office. But how many of us are going to be
    happy waiting 2 min between shots on our digital camera, just to save 1/3
    to 1/2 the memory space. Personally I think that I would rather purchase
    more memory than to have to wait several min (or even 10's of seconds)
    between normal shots. JMHO

    Now, if such a program were developed for archiving photos in more compact
    but lossless forms, It could have a big impact.

    Randy

    ==========
    Randy Berbaum
    Champaign, IL
     
    Randy Berbaum, Sep 30, 2005
    #3
  4. +/-

    Mike Henley Guest

    Indeed, Jpeg is an industrywide standard by a joint ISO/IEC and ITU-T
    committee.

    AMD is now making 5Ghz processors. Broadband is now being offered at
    24Mb. So hardware-bandwidth is no issue. Last thing the industry will
    do is entrust its data formats to a proprietary one from an obscure
    upstart.
     
    Mike Henley, Sep 30, 2005
    #4
  5. +/-

    Trevor Guest

    data formats are like standards - there are so many to choose from. I woudl
    rather put my trust in JPEG2000, but that's taking its time getting to the
    masses - anyone up to date on the Lizardtech claims?
     
    Trevor, Sep 30, 2005
    #5
  6. 10 megabytes only takes about 1 second on 100 Mbps ethernet. Is that a big
    deal?

    Anyhow, a 10 megabyte jpeg is probably more than 50 Mpixels. I don't know what
    kind of viewing devices they have in hospitals, but starting out with
    lower resolution images and then getting high res crops from the real
    image (cropping at block boundaries is cheap in jpeg) strikes me as a good
    solution to reduce bandwidth.
    The usual snake-oil. Lossless compression doesn't work all that well on
    images that contain noise. Any algorithm that deletes noise is also going
    to delete some image detail (unless the algorithm has some much domain
    specific knowledge that you can save just the 'contents' of the image
    and not the pixels.)
     
    Philip Homburg, Sep 30, 2005
    #6
  7. +/-

    Ron Hunter Guest

    Well, let me know with Irfanview has it, and PhotoShop adopts it, THEN I
    will be impressed.
    BTW, NO MENTION was made of color!
     
    Ron Hunter, Sep 30, 2005
    #7
  8. My bullshit senses are tingling.


    Cheers,
    Nicholas Sherlock
     
    Nicholas Sherlock, Sep 30, 2005
    #8
  9. +/-

    Ron Hunter Guest

    Moreover, the math in the quote leaves much to be desired. A 10
    megabyte image across a gigabit ethernet connection takes less than 1
    second to transmit. 60 of those an hour is hardly a significant network
    load.
    Then there is the aspect that MRI's are NOT COLOR. I am sure that going
    back to B&W is not an option for most of us.
     
    Ron Hunter, Sep 30, 2005
    #9
  10. +/-

    Kingdom Guest

    Doubt well ever even see this format nevermind use it, if they want cash
    from hospitals they real are greedy bastards and it's about 2 years too
    late we now have high speed everything!
     
    Kingdom, Sep 30, 2005
    #10
  11. +/-

    *-- Jinn --* Guest

    ABO's feautre (according to the page) is speed. It involves nothing but
    integer manipulations.
    Well, FWIW, that's what ABO seems to offer.
     
    *-- Jinn --*, Sep 30, 2005
    #11
  12. +/-

    *-- Jinn --* Guest

    Why?

    Perhaps the forecast of doom is premature, but better algorithms for compression
    aren't technically impossible.

    Not like MP3 hasn't been bested from numerous angles.

    Perhaps you simply like to express negativity to new ideas /just because/?
     
    *-- Jinn --*, Sep 30, 2005
    #12
  13. Remember how Foveon was supposed to revolutionize digital photography
    and what a zero it turned out to be... All of us poor sods using
    conventional CCD's were supposed to have hopelessly obsolete equipment
    by now.
     
    Bruce Coryell, Sep 30, 2005
    #13
  14. +/-

    Chris Brown Guest

    Sssh! You'll wake *him* up...
     
    Chris Brown, Sep 30, 2005
    #14
  15. +/-

    Charlie Self Guest

    Yup. Let's stick it in the catapult and see if it flies when it reaches
    the end of the shot.
     
    Charlie Self, Sep 30, 2005
    #15
  16. +/-

    Stewy Guest

    Depends. Exactly how fast is your broadband? I'm on a LAN rated at
    10mbps but do I ever get that? Well, if everyone else on the LAN logged
    off, then maybe yes. As it is, I'm lucky to get 100kbps for either music
    downloads or binaries.
    So data compression without loss (I'm assuming these are B&W/false color
    - ie 32 or 256 colors) IS very useful. How it deals with full color
    JPEGs is another matter.
     
    Stewy, Sep 30, 2005
    #16
  17. Adobe DNG does a pretty good job with RAW file storage. It shrinks my
    NEF files from my D70 by about 25% [rough estimate].
     
    Thomas T. Veldhouse, Sep 30, 2005
    #17
  18. +/-

    Mark Roberts Guest

    The "15 to 300 percent smaller than a JPEG" certainly sets off alarm
    bells. 15 to 300 percent smaller than what JPEG quality level? And with
    what *kind* of image (in terms of content): This has an influence an how
    effective JPEG compression is.

    Perhaps, a "naive, uninformed reporter" detector or "overhyping press
    release" detector might be a better term than "bullshit senses", but
    suspicion is certainly merited.
     
    Mark Roberts, Sep 30, 2005
    #18
  19. +/-

    none Guest

    Yes, they are. Huffman coding
    (http://en.wikipedia.org/wiki/Huffman_coding) provably gives the most
    efficient result for lossless compression. This is the algorithm that is
    used for TIFF.

    If you managed to find a general method for loslessly compressing
    bitmapped images to 30% the size of a JPEG, you'd get more attention
    than a 150-word press release on some no-name website.

    -Mike
     
    none, Sep 30, 2005
    #19
  20. +/-

    mark Guest

    did it?
     
    mark, Sep 30, 2005
    #20
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.