Colour blindness, photography and colour management

Discussion in 'Digital Photography' started by Tor Lillqvist, May 18, 2004.

  1. I am slightly red-green colour blind. I haven't had my colour perception
    amomalies exactly measured or anything, though. It's just something that
    was noticed at school as a teenager, using those Ishihara plates. As is
    common, it was quite a surprise to me, I hadn't noticed anything wrong
    earlier... (There went my childhood dream of becoming a train driver ;-)

    I don't know whether I have protanomaly or deuteranomaly. Presumably I
    am not blatantly dichromatic, though, as I do see the difference between
    red and green in general. It's just some hues that are hard to
    distinguish for me, but apparently easy for others, etc.

    Now, I have been thinking about how colour blindness relates to digital
    photography and colour management in particular. Is a colour managed
    workflow worth it at all for me? Colour management (and the whole CIE
    colour model) is based on the "standard observer"...

    Even if I would have a calibrated and profiled monitor, and use a
    profiled printer (either an own or use a on-line service bureau), when I
    tweak an image on my machine before sending it to the printer, I tweak
    it to look correct for me. But presumably the colours might well be
    quite off to a person with normal colour vision?

    And even if I would consider only myself, and ignore showing my pics to
    other people: Doesn't colour management strive to make images look the
    same when produced by different devices (modulo the rendering intent)
    for people with normal vision? Isn't it entirely possible, that as my
    colour receptors have different spectral response than those of the
    standard observer, an image on the monitor and when printed won't look
    at all the same to me, even if it would to a person with normal colour
    vision?

    Basically, I guess what I am asking is whether it makes any sense for me
    to invest some hundreds of euros in colour management hardware and
    software, if it by design won't work for me anyway...

    Thanks in advance for any insights people can share.

    --tml
     
    Tor Lillqvist, May 18, 2004
    #1
    1. Advertisements

  2. Tor Lillqvist

    Beck Guest


    That is not meant to sound flippant or offensive, but have you considered
    specialising in monochrome pictures only?
    I have an acquaintance who is colour blind, but he excels in black and white
    images.
     
    Beck, May 18, 2004
    #2
    1. Advertisements

  3. Tor Lillqvist

    Mark Weaver Guest

    As I understand it, the point of color management is to get the same results
    across devices (or as nearly as possible given the differences). So the
    point of color management for a color-blind person would be to have the
    image look 'right' to you both on your monitor and after you print it (or
    have it printed). As to whether or not it looks right to others -- have you
    run into an actual image where you thought the colors looked right but
    others thought they were off?

    Mark
     
    Mark Weaver, May 18, 2004
    #3

  4. I would not consider this a significant issue, especially if your
    difficulties lie in such a narrow area of the color spectrum.

    Unless you're shooting artwork, tile patterns, or something that
    basically can be limited to those wavelengths you have difficulties with,
    there will almost always be enough of the rest of the spectrum in the image
    to judge accurately enough. Most things that might be green, for instance,
    are going to be obviously green (grass, foliage, etc), or will transition
    across the spectrum to include those areas you can see well. A balloon shot
    in high contrast conditions may run the gamut from almost pure white where
    it reflects the light, to black in the shadows. By concentrating on those
    areas when doing color correction, you shouldn't be off the mark by much,
    if any.

    And that's part of the secret to color correction anyway - you watch
    the highlights, and/or you pick something white that transitions into
    shadow. If a white shirt shows a color cast, other than grey, as it goes
    into shadow, chances are you need to tweak the color (provided the lighting
    spectrum supports this - don't do it for sunset pics, for example, because
    they're supposed to be yellow ;-))

    But even if this presents a problem, there are other methods. If you
    tend to shoot controlled or studio work, you can waste a frame on a color
    chart, and as long as the rest of the frames are done in the same lighting
    conditions, you can correct the entire session based on the readings from
    the chart. A grey card can work too, but tends to be less accurate.

    And photo editing programs can help as well. You can snatch a color
    sample from some area of the image and look at the actual RGB (or CMYK)
    values. Tricky, but with a little practice (and perhaps a guideline chart
    or image) you can tweak the colors as needed. Again, this works better in
    neutral areas - white, grey, and black. But it goes hand-in-hand with the
    chart idea above. This will also work if you ever run into something that
    isn't obvious and falls into those problem areas you have - a quick sample
    will tell you red or green easily enough.

    But to be sure your vision isn't shifted overall (for instance, what
    seems pure white to you is actually slightly reddish), correct a few images
    and have someone review them. Preferably someone who's done it many times
    before - color correction is very subjective until you're used to doing it.
    At the same time, use your color sampler to ensure yourself. BUT, only
    after your monitor has been reasonably calibrated. Doesn't have to be
    fancy, just use a three-color gamma program (like Adobe Gamma) first.

    And be wary. The color sampler in Photoshop tends to pick individual
    pixels, which in a neutral grey area can still run the gamut of colors. Use
    a command, such as the Noise/Median filter or any blur command, to get the
    pixels evened out within an area before sampling.

    Good luck!


    - Al.
     
    Al Denelsbeck, May 18, 2004
    #4
  5. SNIP
    You can set the sampler to a 1, 3 or 5 pixel pixel square to sample from.

    Bart
     
    Bart van der Wolf, May 18, 2004
    #5
  6. Tor Lillqvist

    GraemeGill Guest

    I'm no expert on color blindness, but the impression I've got from several
    books/articles that lightly cover this topic, is that the majority of
    colorblindness is due to weakness in one or more receptors, not a shift
    in their spectral sensitivities. If this is indeed the case, then normal
    tri-stimulus color matching should be of benefit. For someone who is weak
    in distinguishing certain ranges of colors, then "doing it by the numbers"
    with a CMS, may be the only way of producing work that can be confidently
    viewed by non color deficient viewers.
    As with investing in any expensive set of tools, the best thing is to
    insist on trying them out, before committing to buy ...

    Graeme Gill.
     
    GraemeGill, May 18, 2004
    #6
  7. It will still work. Your colour blindness is only part of the
    calibration issue, and in fact the questions of black point, gamma and
    brightness of images are critical for good prints - a calibration device
    sorts these out as well as trying to make all colours almost match.
    Colour calibration is, in any case, not all that successful even with
    'eye' type sensors; it's often battling against extreme nonlinearity by
    imposing extreme correction curves in very critical areas like the
    transitions from deep shadows to visible detail, and can result in odd
    effects like greenish hues appearing on these boundaries in prints when
    the monitor looks perfect. You might miss seeing such effects and that
    would be one risk of using calibration!

    After years of using such devices, I tend to calibrate contrast only
    now. I accept the different colour 'flavours' of my various monitors and
    don't mess with their gain curves, and nor do I adjust my images using
    different curves on RGB etc. I just make simple adjustments to
    brightness, daylight/white balance, contrast and so on. This works well
    enough. I have not used a digital camera yet which had crossed curves!

    David
     
    David Kilpatrick, May 19, 2004
    #7
  8. If you take one of those Ishihara plates, convert to 'cone space' (using
    Bradford) drop the appropriate channel (red, green, or blue) en convert back
    to a normal color space, you can simulate the effects of color blindness
    quite well.
     
    Philip Homburg, May 19, 2004
    #8
  9. Tor Lillqvist

    Roger Breton Guest

    Is that all the 'Bradford' transform does to a set of XYZ values, convert to
    'cone space'?

    Is it the same thing as converting from XYZ to LMS then?

    Roger Breton
     
    Roger Breton, May 20, 2004
    #9
  10. Bart van der Wolf, May 20, 2004
    #10
  11. Or even several different types of cone spaced, depending on the used
    CAT (Bradford, CAT02, Sharp, ...).

    Gerhard
     
    Gerhard Fuernkranz, May 20, 2004
    #11
  12. It is designed for white-point transformations. And it happens to work on
    the Ishihara plates. I don't know how big the difference is with what is
    really going on.
    I guess that LMS is more than a linear transformation of XYZ (but I don't
    know).
     
    Philip Homburg, May 20, 2004
    #12
  13. Tor Lillqvist

    Waldo Guest

    If you take one of those Ishihara plates, convert to 'cone space' (using
    Or try this website:

    http://www.vischeck.com/

    Includes a plug-in for Photoshop...

    Waldo
     
    Waldo, May 24, 2004
    #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.