Question about "raw" and in-camera image processing

Discussion in 'Digital SLR' started by RichA, Jun 23, 2005.

  1. RichA

    RichA Guest

    I've noticed different "looks" to pictures in
    tests done on various DSLRs. Not just the result
    of pixel count differences (there may be none, say,
    between the Olympus E-300 and Canon Rebel XT)
    but because of what is happening to the image
    once it's aquired. What I'd like to know is,
    which cameras seem to have the most agressive
    image processing firmware and is this applied
    to raw files? I know that some cameras have a selectable
    noise reduction feature, I wasn't really interested
    in it's effect, so the question pertains to cameras
    with that feature turned off.
     
    RichA, Jun 23, 2005
    #1
    1. Advertisements

  2. RichA

    Ed Ruf Guest

    No. That's what RAW means.
     
    Ed Ruf, Jun 23, 2005
    #2
    1. Advertisements

  3. RichA

    RichA Guest

    So the raw file's composition is determined solely
    by the sensor itself?
     
    RichA, Jun 23, 2005
    #3
  4. RichA

    eawckyegcy Guest

    Even a brief perusal of a grep at google will tell you as much.

    You just babbling because you are lonely or something?
     
    eawckyegcy, Jun 23, 2005
    #4
  5. RichA

    Stacey Guest

    I've read that the cmos sensors are doing noise reduction at the sensor
    level so there is at least some processing going on with a cmos camera even
    if you believe there is -zero- processing on the RAW file. Of course I
    believe there is no contrast or sharpening, WB etc going on but I wouldn't
    be shocked to find that some image processing is being done even to a "RAW"
    file before the RAW data is written to the card. SInce almost every camera
    has it's own RAW file format, there must be some sort of processing done to
    even generate the file?
     
    Stacey, Jun 24, 2005
    #5
  6. There is processing, of course, but the question is whether it changes
    the Raw sensor data. But ....

    First, of course, there has to be processing even to obtain what we
    think of as the Raw sensor data. Sensors deliver electrons, and cameras
    do not just store a count of electrons into the Raw file. So they must
    have an algorithm for converting the count of electrons, or amount of
    charge, into the numbers held in the Raw file or used for further
    processing.

    Second, I have seen a suggestion that the Nikon D70 does further
    processing in order to compress the sensor data further. I think this
    is the only case of lossy compression of sensor data I've read about,
    and I have no independent evidence for this statement:

    "The Compressed NEF format is the only one that comes close to
    retaining the full data set the D70 is capable of acquiring. I say
    "close to retaining" because the compression scheme Nikon uses is not
    lossless. Basically, the camera takes the highlight data and places
    them into groups (essentially a rounding of many of the data points),
    producing the equivalent to somewhere between 9 and 10 bits of data.
    When this is returned to 12-bit form, there's a bit of posterization in
    the highlight data. The reason this works as a visually lossless scheme
    is that our eyes really can't resolve more than about an 8-bit value
    can produce (and our eyes aren't linear in response to light, either).
    For the most part it isn't a big thing that the compression loses data,
    though there may be some post-processing manipulations that will render
    slightly differently because of the data rounding."

    http://www.bythom.com/D70REVIEW.HTM

    Third, generating the file itself is not really an issue. A Raw file is
    a container, and generating the tags or whatever of the container
    doesn't itself mean that the sensor data is being changed.

    None of the above is about any type of processing that uses data from
    more than one pixel (or "sensel"?) at a time. (Sharpening, noise
    reduction, etc). I think that is what people are normally talking about
    when they distinguish between Raw and other formats.
     
    Barry Pearson, Jun 24, 2005
    #6
  7. RichA

    Stacey Guest

    Barry Pearson wrote:


    But how would we ever really know if this is the case? They could easily
    "slip" some processing in that you'd never know they had done. Not that we
    could ever do anything about it or that it even really matters but just
    because it's "RAW" doesn't mean it's just a sensor dump straight to the
    card either.
     
    Stacey, Jun 24, 2005
    #7
  8. RichA

    JPS Guest

    In message <>,
    There's no such thing as a straight dump of analog data into binary
    data.
    --
     
    JPS, Jun 25, 2005
    #8
  9. RichA

    Stacey Guest

    Yet people "believe" that's what RAW data is and there is no posibility
    that it has been processed by the camera. Seems obvious to me Ed Ruf
    believes this from his coment "No. That's what RAW means." when the
    question of extra processing was brought up.
     
    Stacey, Jun 25, 2005
    #9
  10. RichA

    Ed Ruf Guest

    You can believe what you like.
    IF you reread the OP
    You may see that one can interpret in another way. Given the emphasis by
    the OP on "aggressive image processing firmware" and "tests done on various
    DSLRs", I took this to be inquiring about the many comparison images in
    sites such as dpreview which are comparisons of JPGs produced by the
    camera. These are affected by the default in camera processing in such
    areas as sharpening, contrast and saturation. My reply was stating THESE
    in camera adjustments have no effect on the raw images.
     
    Ed Ruf, Jun 25, 2005
    #10
  11. RichA

    l e o Guest


    So what? Each cell is doing the subtraction from it's own ambient noise
    level to get the signal value. How do you see this as an image
    enhancement? [I assume your refer processing as noise
    reduction/enhancement.] And do you know how much is the noise floor
    level of your CCD? How about the noise generated by the amplifier used
    to capture the signal from the CCD? If you don't like CMOS, then don't
    dream of D2X.
     
    l e o, Jul 7, 2005
    #11
    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.