Which DSLR has worst JPEGs compared to RAW images?

Discussion in 'Digital Cameras' started by Rich, Mar 8, 2006.

  1. Rich

    Rich Guest

    I've heard JPEGs vary quite a bit between cameras.
    -Rich
     
    Rich, Mar 8, 2006
    #1
    1. Advertisements

  2. It's all to do with compression.
    The smaller the file, the more artifacts. Even these can be dealt with,
    incidentally. What critera do you seek to pass judgment on?
     
    wraped in canvas, Mar 8, 2006
    #2
    1. Advertisements

  3. Rich

    Pete D Guest

    All you have to do is have a look at the review photos side by side and then
    you realise that the reviewer was using a very large magnifying glass and
    the differences are actually very small and even if printing at above 8x10
    you will not be able to see a difference in the jpeg quality unless you use
    a very large magnifying glass as well. Reality is that the Pentaxes actually
    produce a very good result and handle particularly well and have a better
    feature set than most of the competition.
     
    Pete D, Mar 8, 2006
    #3
  4. Rich

    Martin Brown Guest

    Testing only the highest quality JPEG setting would be sufficient. None
    of the others can possibly compete with RAW. MOst people would be unable
    to tell the differnce between a top quality JPEG and RAW in a blind test.

    Some high end cameras have a highest quality that sets every entry in
    the the quantisation matrix to 1. There is no theoretically better JPEG
    encoding possible from the original data (although exact choices of in
    camera processing, unsharp mask, colour space, white balance, contrast
    etc can compromise what is saved). RAW gives you a lot more latitude.

    The JPEG codec is pretty standard now but the parameters used
    (especially for quantisation) do vary significantly. Canon is very good,
    Nikon and Pentax arguably are OTT.
    I am surprised by this since the Pentax istDS JPEGs use the very highest
    quality settings - and I have never seen any JPEG issues in my images.
    It will, but the Pentax JPEG is nowhere near as bad as that review seems
    to suggest. It would be interesting to look at the original images to
    see what other in camera settings have compromised image quality.
    That seems likely. No point blaming the JPEG for some incorrect settings
    used on the camera for preprocessing. Obviously if you save RAW you can
    do anything afterwards but if you save a defective in camera processed
    image as a JPEG you no longer have the raw sensor data to work from.

    Regards,
    Martin Brown
     
    Martin Brown, Mar 8, 2006
    #4
  5. There is a very curious asumption being made in this thread. That the RAW
    files from a camera are actually images. They are not. The file is raw
    (processed to be sure) data from the camera's processor before converting to
    an image. All a JPEG camera image is is that data converted to an image
    using the Camera makers idea of what an image which will satisfy most of the
    people most of the time should look like.

    Some DSLRs allow you to alter the makers decissions and produce a JPEG file
    with wider contrast range but a flatter looking image. Sharper of not, etc.
    The amount of compression of a JPEG is usually predetermined. Olympus for
    example create a huge RAW file but very clean JPEGs. Oly 'E' series owners
    won't see a lot of difference between a RAW file and a JPEG altered with
    custom functions to suit a users idea of "right". They make a good camera
    for Newspaper work.

    Canon DSLRs will benefit from the use of RAW data processed after the shoot
    because of the degree of clipping and compression built into Canon cameras.
    Nikon's are heading the way of Canon with their recent models.
     
    wraped in canvas, Mar 8, 2006
    #5
  6. Rich

    Martin Brown Guest

    Over The Top. The incremental increase in filesize becomes excessively
    large for an almost negligible improvement in image quality.

    Regards,
    Martin Brown
     
    Martin Brown, Mar 8, 2006
    #6
  7. Rich

    Martin Brown Guest

    It is usually the RAW Bayer mask sensor data stored at 11-12bits per
    element, possibly with preliminary calibration and perhaps stored with
    lossless compression. It is the original RAW sensor data needed to
    compute a 24bit colour image. Where did I say anything else?
    And the same is true of in camera conversion to 8-bit TIFF.
    It varies with the models and the only way to be sure is to test it to
    see if it will do what you want. If there is any doubt or the subject is
    high contrast with shadow detail and important highlights then raw
    provides a lot more latitude for post shoot processing.

    Regards,
    Martin Brown
     
    Martin Brown, Mar 8, 2006
    #7
  8. Rich

    Matt Ion Guest

    Most people would be unable to tell the difference between black and
    white in a 'blind' test ;)


    ---
    avast! Antivirus: Outbound message clean.
    Virus Database (VPS): 0610-0, 03/06/2006
    Tested on: 3/8/2006 7:54:21 AM
    avast! - copyright (c) 1988-2005 ALWIL Software.
    http://www.avast.com
     
    Matt Ion, Mar 8, 2006
    #8
  9. Aoooouch!

    Yet, as in doing a good deed that no one sees, the photogapher will
    know.... could be critical in some cases.
     
    John McWilliams, Mar 8, 2006
    #9
  10. Rich

    Martin Brown Guest

    Only a partial study of the formats I have encountered.

    Pentax and some Nikons are using 100 on the IJG scale (all quantisation
    coefficients are identically 1).

    Several other makers have maximum qualities that are more highly
    compressed (IJG ~95 ish) but with custom quantisation tables - not a
    scaled version of the canonical "example" ones from the standard that
    almost everyone uses.
    I haven't seen that one. 444 chroma subsampling certainly is very
    optimistic if the file came off a Bayer mask sensor. It might make sense
    for a Foveon sensor or a high end scanner.

    Regards,
    Martin Brown
     
    Martin Brown, Mar 8, 2006
    #10
  11. There is a firmware upgrade for the E300 that fixes the white balance
    problem. You have to first upgrade the software you use to access the camera
    (maybe twice) then all the upgrades for your lenses, flash and camera are
    available to you. The firmware fix was so radical, I thought I had gotten a
    later model camera! You are right too, about the almost no difference
    between raw and JPEG on this camera.
    --
    Douglas MacDonald
    www.photosbydouglas.com
    www.weprint2canvas.com
    If you really can't resist writing to me personally,
    use my first name at either of the domains.
     
    wraped in canvas, Mar 8, 2006
    #11
  12. Rich

    Matt Ion Guest

    The one advantage of shooting RAW is the ability to adjust the white
    balance later, as the camera isn't doing it for you.

    One of the other benefits is the extended dynamic range of the RAW
    format, which is useful if you have extremely high contrast situations,
    such as very low light or very bright scenes. For your shooting, that's
    probably not so much of an issue.



    ---
    avast! Antivirus: Outbound message clean.
    Virus Database (VPS): 0610-1, 03/08/2006
    Tested on: 3/9/2006 4:00:36 PM
    avast! - copyright (c) 1988-2005 ALWIL Software.
    http://www.avast.com
     
    Matt Ion, Mar 10, 2006
    #12
  13. Rich

    Matt Ion Guest

    Thank you, thank you, I'll be here all week...


    ---
    avast! Antivirus: Outbound message clean.
    Virus Database (VPS): 0610-1, 03/08/2006
    Tested on: 3/9/2006 4:01:03 PM
    avast! - copyright (c) 1988-2005 ALWIL Software.
    http://www.avast.com
     
    Matt Ion, Mar 10, 2006
    #13
    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.