Increase in saturation can DECREASE contrast (gurus refuted)

Discussion in 'Digital SLR' started by David Virgil Hobbs, Dec 29, 2006.

  1. It is mind-boggling, judging from the opinions presented on the
    internet, the lack of understanding regarding color saturation in
    photography. Most of the so-called experts seem to be mindlessly
    reciting mantras they have heard somewhere without understanding the
    topic. They all self-confidently proclaim that an increase in
    saturation INCREASES contrast.

    In reality, when two colors are similar in that in both colors the same
    element of the RGB (red-green-blue) components dominates, an increase
    in saturation DECREASES contrast between the two colors; but when two
    colors are dissimilar in that in one color one element of the RGB
    dominates while in the other color another element of the RGB
    dominates, an increase in saturation INCREASES contrast between the two
    colors. This is important because so many photos of persons are wrecked
    by an excess of contrast between two similar colors such as on a
    person's face.

    In digital photography colors are composed of a mix of R for red, G for
    green, and B for blue values. I here prove to you that when two colors
    both feature the R value being dominant, or when two colors both
    feature the G value being dominant, or when two colors both feature the
    B value being dominant, an increase in saturation results in a
    DECREASE, not an INCREASE in contrast between the two colors; and when
    two colors feature in one color one of the rgb values being dominant
    and in the other color another of the rgb values being dominant then
    saturation INCREASES contrast between the two colors.

    All except one of the definitions of saturation I found on the internet
    I found to be incomprehensible; they all sounded as if someone was
    reciting a mantra he had memorized that he did not understand. The
    definition I found to be comprehensible was:

    "What is saturation, and why it is so important?
    The saturation defines the level of pureness or a color. All the colors
    derives by a mix of the tree primary colors - red, yellow and blue (or
    Red, Green and Blue in the common RGB color space). The more a color
    is saturate, the more it is close the one of the primary colors :
    theoretically, if you raise the saturation to its upper limit the photo
    would be composed by just red, green and blue."
    makes a color in which the R value is dominant more red. From working
    with HTML I have seen what the photography gurus seem to fail to
    understand, which is that in shades of gray, the R G and B values are
    equal. The photography gurus all agree that as saturation is decreased
    the colors draw closer to being shades of gray as you can see working
    with saturation in a photo editor.

    Thus when there is zero saturation, or shades of gray, that means that
    in each color, the difference between r and g and b, is zero, ie
    rgb(192, 192, 192), rgb(127, 127, 127), rgb(67, 67, 67). This means
    (my personal inference) that as saturation is decreased, the
    differences in the r and the g and the b in the color decrease as the r
    g and b values all draw closer to an average of the r and the g and the
    b values in the color.

    As saturation is increased, the difference between the r g or b element
    that is strongest in the color and the other colors in the RGB trinity
    is increased, as I can tell from my own personal experience and from
    the quotation from included above.

    Thus, an increase in saturation will DECREASE the contrast between two
    colors both of which have say the R component amongst r g and b
    dominating the g and b components; and an increase in saturation will
    increase the contrast between two colors when one color has say the r
    component dominating the g and b components, and the other color has
    say the b component dominating the r and g components.

    Take for example two colors found in a man's face. One is
    rgb(210,131,118); the other is rgb(233,164,135). In both colors the r
    dominates. The difference between the r's is 23, between the g's is 33,
    between the b's is 17 for a total differential of 73. If, increasing
    saturation, you push the r's half way to the maximum 255, and the g's
    and the b's half way to the minimum of 0, you end up with the first
    color being 232,65,59 and the second color being 244, 82, 67 so that
    the difference between the colors has declined from 73 to 37, a
    decrease in contrast. Check out what these colors look like in even a
    simple program the internet gurus are too good for such as Microsoft
    Paint and you can see the decline in contrast with your own eyes.

    Again, take for example two numbers 249 and 55; the difference between
    the two is 194. Push 249 half way to 255 and you get 252, push 55 half
    way to 255 and you get 155; the difference between the two numbers
    declines from 194 to only 97.

    Yet again, take for example two numbers 150 and 100; the difference
    between the two is 50. Push 150 half way to 255 and you get 202, push
    100 half way to 255 and you get 177; the difference between the two
    numbers declines from 50 to only 25.

    Now take two colors, with one featuring the R or red component
    dominating the G and B components, and the other featuring the B or
    blue component dominating the R and G components, say rgb(210,131,118)
    and rgb(55,85,121). The difference between the r's is 155, between the
    g's is 46, and between the b's is 3, for a total differential of 204.
    If you increase saturation pushing the first color's dominating r value
    half way to 255 and the first color's g and b values half way to zero
    you change the first color to 232,65,59. If you increase saturation
    pushing the second color's dominating b or blue value half way to 255
    and the second color's r and g values half way to zero, the second
    color changes to 27,42,188. The overall differential between the two
    colors becomes 205 in the r's, 23 in the g's, and 129 in the b's for a
    total of 357, whereas previously the total differential was only 204,
    and the contrast INCREASES instead of decreases because in this case in
    one color the R value dominates and in the other color the B value

    I find it significant and incredible that the digital photography world
    has apparently failed to understand this key point, that increasing
    saturation decreases contrast between two similar colors while
    increasing contrast between two dissimilar colors.

    Many of my photos of humans were unacceptable using a digital camera
    that a short while ago sold for $600 but now sells for only $100 (I
    wonder why?)--until I cured the hyper-contrast in the faces of the
    subjects by increasing saturation and then brought the tint back to
    normal. Using said camera, my photos of a person would make the person
    look like a different person from photo to photo. Photos of humans can
    be wrecked by hyper-contrast between two similar colors in a person's
    face resulting from a lack of saturation, but all you ever hear from
    the internet photography gurus is that increasing saturation INCREASES

    I confess to being proud that I, an amateur dabbler, have been able to
    figure out how a lack of saturation can result in hyper-contrast,
    without any help from anyone, so as to put down the photography gurus
    who proclaim that increasing saturation increases contrast. Too bad
    employers are so mindlessly concerned with credentials, experience, and
    coddling those whom the rude label as 'dorks', that they de-emphasize
    all kinds of important qualities and--since hyper-emphasis of one thing
    leads to de-emphasis of other things--under-value persons such as

    @2006 David Virgil Hobbs
    David Virgil Hobbs, Dec 29, 2006
    1. Advertisements

  2. David Virgil Hobbs

    Pat Guest

    Did someone get up on the wrong side of the bed this morning. You are
    ranting against a non-problem. To start with, I don't remember "all"
    of those experts going on as you said. If I did see it, it went in one
    eye and out the other because I don't necessarily believe everything I
    read. Contracts, saturation, everything is sort of picture-specific.
    That's why you have the ability to adjust it by hand. Sometimes I up
    it. Sometimes I lower it. It all depends. So go have another cup of
    coffee and calm down.

    If many of the pictures were unacceptable, it shouldn't have taken you
    long to figure out how to adjust you camera to make them acceptable.
    Most cameras allow you to adjust contract and saturation inside the
    camera. They do that for a reason, ya know. And maybe you shouldn't
    be using a $100 camera

    Again, who are these "internet photography gurus".

    And just because their theory doesn't work for your particular
    circumstance doesn't mean it doesn't work for most people most of the
    I'm glad you got the whole saturation/contract thing figured out. Now
    if you could just figure out the whole URL thing on your link so that
    it works, you'd be doing good.
    Pat, Dec 29, 2006
    1. Advertisements

  3. David Virgil Hobbs wrote:

    <snip unprovoked rant>

    What brought that on?

    Richard Polhill, Dec 29, 2006
  4. David Virgil Hobbs

    Lionel Guest

    A poorly medicated case of paranoid-schizophrenia, at a guess, seeing
    as I've never even heard of any 'gurus' making claims anything like
    the ones he thinks he's refuted.
    (I wonder if he's the same loon who keeps on spewing all that demented
    spam about claims of imaginary MI5 harrassment?)
    Lionel, Dec 29, 2006
  5. Sorry for snipping most of that enormously long, mindless drivel, but..
    it was enormously long mindless drivel...
    Strong words. Name the 'gurus' you refute.
    Well done, Einstein! The post could have ended there.
    "so many" - f'rinstance?

    Maybe that was because they were right. The one you quote is horribly
    No. It is the intensity of the colour. 'Pureness' is a very loose
    term and could simply refer to the colour's accuracy or closeness to
    the original.
    Only when using the RGB model. Colors are actually defined by
    different wavelengths and mixtures of wavelengths. The RGB, CMYK, LAB,
    etc models are simply ways to try to represent colours using available
    technology or to simulate the operation of the eye.
    I disagree strongly with this interpretation. For example, if the
    colour had equal blue and green components, this definition suggests it
    would have to become 'more close' to *one* primary - so it would have
    to become *either* more blue or more green, in order to become more
    saturated. Stupid.
    'Apparently' is the key word here. It seems so to you, yet you haven't
    posted a single example. And the example you post as gospel is flawed.
    Maybe when the rest of the world thinks one thing, and you think
    another, there could be an alternative to the thought that *you* are
    the only correct one..
    Gee, really?
    Which you most likely introduced, either by excessive post-processing
    or bad camera settings.
    Gee, again. Maybe you should ask why you *had* to bring them back to
    Aha, there we have it - is this an attempt at an ego trip? Sorry. You

    Maybe you might get a bit less 'under-valued' if you stopped slanging
    off at imaginary others, waffling on forever, and making out there is a
    problem when none exists. Yes, some amateurs (and pseudo pro's) aren't
    very good at post processing and may screw up colours, contrast, and
    skin-tones with that. Indeed if you want to see saturation and
    contrast gone wrong (not many flesh tones though), try ... (O;

    But no 'experts' I have seen have made any misleading claims. We
    await, with breath baited, all your examples to the contrary.


    PS - maybe you should fix your homepage link...
    mark.thomas.7, Dec 29, 2006
  6. David Virgil Hobbs

    twilson Guest

    ....couldn't we all just get along?

    Come on everyone, there is nothing quite as entertaining as a good old
    fashioned rant in a public place. It's kind of like watching a husband
    and wife going three rounds in the grocery store over which brand of
    jam to buy, break out the snacks, sit back and enjoy... and when said
    rant gets copyrighted... whoa Daddy!

    I can just see all those gurus gnashing their teeth now that they have
    been so ultimately and finally refuted on that oh so topical saturation
    debate! The fate of the universe may very well have been settled right
    here in this very forum... take that you so called saturation experts
    (shake clenched fist at retreating lab coated, bespecked experts
    scrambling back under the rocks from whence they came)! The saturation
    gurus are likely all down at the local bar muttering under their breath
    trying to figure out how to get around that blasted copyright
    statement... personally I can hardly wait for the fireworks to begin!
    twilson, Dec 30, 2006
    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.