8-Bits vs 12 or 16 bits/pixel; When does more than 8 bits count ?

Discussion in 'Digital Photography' started by Al Dykes, Dec 29, 2003.

  1. Al Dykes

    Al Dykes Guest

    Cameras and film scanners generate 12, 16, or more bits per pixel.
    This data is retained as long as the image is in raw or tiff format,
    or when manipulated in a professional software tool like
    Photoshop. jpeg format is 8 bits, right ?

    Assuming you do all your image manipulation at maximum color depth and
    generate your results at a color depth that's no better than it has to
    be, when does 8 or 16 bits become perceptably degraded when compared to
    the original ?

    Does manipulation in a software tool that uses 8-bits, internally
    produce a poorer image than one one done in a larger color depth and
    reduced to 8 bits for final output ?

    What color depth can a good home color printer or a high quality CRT
    or LCD PC screen render ?

    Comments ?
    Al Dykes, Dec 29, 2003
    1. Advertisements

  2. Al Dykes

    Rafe B. Guest

    Aggressive tonal manipulation benefits from the
    extra bit depth.

    Nowadays most film scanners operate internally
    (in the firmware and driver) with high bit depths.

    So, if you do your most aggressive tonal moves
    in the scanner driver, there's not much benefit to
    importing 16-bit (per color) image files into

    OTOH, if you prefer to scan "raw" and defer all
    color correction to Photoshop, you will need to
    import 16-bit (per color) files into Photoshop.

    If the output of the scan (or digicam) has a limited
    tonal range, you better hope you've got lots of bit
    depth to fix it up.

    If memory is no issue, if you don't mind image
    files double the size, and if your image editing
    tools can handle it, there's no reason not to use
    16-bit (per color) image files, aside from the
    extra processing time within the image editor.

    This is a very controversial topic, and has been
    hotly debated on this forum and others. Do a
    web search on Dan Margulis -- there's a transcript
    on his site of an excellent in-depth discussion of
    this very topic.

    rafe b.
    Rafe B., Dec 29, 2003
    1. Advertisements

  3. Many questions. I know that for raw format, the 12-bits of data
    means that you can darken the picture, and get details from the
    highlights that was burned out in 8-bit. The same goes for shadow,
    where you can extract more detail. Like doing +/- 2 EV after the
    picture is taken.

    Same for color correction/manipulation. It will eliminate some
    of the banding that might otherwise crop up.

    Well manipulated RAW pictures can be made very impressive.
    Povl H. Pedersen, Dec 29, 2003
  4. Al Dykes

    Tony Spadaro Guest

    When you are doing colour correction. The bit depth refers to the number of
    shades of colour available 256 per colour for 8 bit, 4096 per colour for 12
    bit. If you multiply 256x256x256 and then multiply 4096x4096x4096 you will
    see that there are a whole lot more shades at 12 bits per colour. Once you
    have done colour corrections everythign above 8 bits per colour is just
    taking up space, and will have to be converted to print.

    home of The Camera-ist's Manifesto
    The Improved Links Pages are at
    A sample chapter from my novel "Haight-Ashbury" is at
    Tony Spadaro, Dec 29, 2003
    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.