If I shoot a grey card, should this end up as 127 grey?

Discussion in 'Digital Photography' started by Alan F Cross, Feb 25, 2004.

  1. Alan F Cross

    Alan F Cross Guest

    Talking theoretically ...

    If I take a shot of an 18% grey card, should I find this to be at level
    127 (on an 8-bit grey-scale)?

    Coming at the question from the other end, if I make a print of an
    electronically-generated 127 patch, should this look like my 18% grey
    card, or some other grey shade?

    I know in practice things won't work out perfectly, but I want to
    understand what I should be aiming for.

    TIA.
    --
    Alan F Cross
     
    Alan F Cross, Feb 25, 2004
    #1
    1. Advertising

  2. Alan F Cross

    Lionel Guest

    Kibo informs me that Alan F Cross <> stated
    that:

    >Talking theoretically ...
    >
    >If I take a shot of an 18% grey card, should I find this to be at level
    >127 (on an 8-bit grey-scale)?


    No. ;)

    In the unlikely event that there's any linear relationship, it'd be more
    likely to be about 46.


    >Coming at the question from the other end, if I make a print of an
    >electronically-generated 127 patch, should this look like my 18% grey
    >card, or some other grey shade?


    Some other shade.

    --
    W
    . | ,. w , "Some people are alive only because
    \|/ \|/ it is illegal to kill them." Perna condita delenda est
    ---^----^---------------------------------------------------------------
     
    Lionel, Feb 25, 2004
    #2
    1. Advertising

  3. "Alan F Cross" <> wrote in message
    news:...
    > Talking theoretically ...
    >
    > If I take a shot of an 18% grey card, should I find this to be at level
    > 127 (on an 8-bit grey-scale)?
    >
    > Coming at the question from the other end, if I make a print of an
    > electronically-generated 127 patch, should this look like my 18% grey
    > card, or some other grey shade?
    >
    > I know in practice things won't work out perfectly, but I want to
    > understand what I should be aiming for.
    >
    > TIA.
    > Alan F Cross


    -------

    Analysis of the image INFO in PS will tell you...Window>show>info

    Mid-Gray (the 18% gray card value) will produce the following color
    analysis:

    R = 133-135 G = 133-135 B = 133-135 >>Lum = 127<< Sat = 160 Hue = 0

    C = 0; M = 0; Y = 0; K = 54%

    Journalist
     
    Journalist-North, Feb 25, 2004
    #3
  4. Alan F Cross <> wrote in
    news::

    > If I take a shot of an 18% grey card, should I find this to be at level
    > 127 (on an 8-bit grey-scale)?
    >
    > Coming at the question from the other end, if I make a print of an
    > electronically-generated 127 patch, should this look like my 18% grey
    > card, or some other grey shade?


    No and no :)

    First - the digital 8 bit pictures uses a gamma to
    improve the useful gradiation.

    Second - the grey card can be any value - it depends
    on your exposure.


    /Roland
     
    Roland Karlsson, Feb 25, 2004
    #4
  5. Alan F Cross

    bob Guest

    "Journalist-North" <> wrote in news:4a2%
    b.1911$:

    >
    > "Alan F Cross" <> wrote in message
    > news:...
    >> Talking theoretically ...
    >>
    >> If I take a shot of an 18% grey card, should I find this to be at level
    >> 127 (on an 8-bit grey-scale)?
    >>
    >> Coming at the question from the other end, if I make a print of an
    >> electronically-generated 127 patch, should this look like my 18% grey
    >> card, or some other grey shade?
    >>
    >> I know in practice things won't work out perfectly, but I want to
    >> understand what I should be aiming for.
    >>
    >> TIA.
    >> Alan F Cross

    >
    > -------
    >
    > Analysis of the image INFO in PS will tell you...Window>show>info
    >
    > Mid-Gray (the 18% gray card value) will produce the following color
    > analysis:
    >
    > R = 133-135 G = 133-135 B = 133-135 >>Lum = 127<< Sat = 160 Hue = 0
    >
    > C = 0; M = 0; Y = 0; K = 54%
    >
    > Journalist
    >
    >


    In Photoshop 6.0, the values associated with the 15% and 20% grey (default)
    swatches are:
    R=G=B = 218 and 205

    CMYK values are 13%, 10%, 11%, 0%; and 19%, 14%, 15%, and 0%

    I think an 18% grey card, if exposed for 18% grey (auto meter) should come
    up with an 18% exposure, which will print at 18%, and everything will be as
    expected. If an 18% grey were recorded at 50%, and then printed at 50%, it
    would be way too dark, wouldn't it?

    Bob

    --
    remove the backwards "SPAM" to reply.
     
    bob, Feb 25, 2004
    #5
  6. Alan F Cross

    Alan F Cross Guest

    In message <4a2%b.1911$>,
    Journalist-North <> writes
    >
    >"Alan F Cross" <> wrote in message
    >news:...
    >> Talking theoretically ...
    >>
    >> If I take a shot of an 18% grey card, should I find this to be at level
    >> 127 (on an 8-bit grey-scale)?
    >>
    >> Coming at the question from the other end, if I make a print of an
    >> electronically-generated 127 patch, should this look like my 18% grey
    >> card, or some other grey shade?
    >>
    >> I know in practice things won't work out perfectly, but I want to
    >> understand what I should be aiming for.
    >>
    >> TIA.
    >> Alan F Cross

    >
    >-------
    >
    >Analysis of the image INFO in PS will tell you...Window>show>info
    >
    >Mid-Gray (the 18% gray card value) will produce the following color
    >analysis:
    >
    >R = 133-135 G = 133-135 B = 133-135 >>Lum = 127<< Sat = 160 Hue = 0
    >
    >C = 0; M = 0; Y = 0; K = 54%
    >
    >Journalist
    >



    How do you produce an 18% grey in Photoshop, then? It can't be a scan,
    because that will be off by an unknown amount. Are you just holding up a
    grey card next to the screen, and tweaking till they look similar? The
    values will depend on gamma and to some extent colour space.

    Found this Q&A on Google:

    What is the calculation that yields 18% grey = 50% tonality, 128 on an
    8-bit grayscale, etc?

    That depends on the gamma of the color space used. ColorMatch, a 1.8
    gamma color space, will have 18% gray at around 103. sRGB and Adobe RGB
    (1998), both 2.2 gamma spaces, put it at around 125.

    So clearly - it depends!!
    --
    Alan F Cross
     
    Alan F Cross, Feb 25, 2004
    #6
  7. Alan F Cross

    Guest

    In message <>,
    Alan F Cross <> wrote:

    >Talking theoretically ...
    >
    >If I take a shot of an 18% grey card, should I find this to be at level
    >127 (on an 8-bit grey-scale)?


    Most cameras (JPEG and TIFF) and RAW converters aim for 127 or 128, some
    for 117 (I'v seen this value, but I don't remember where).

    >Coming at the question from the other end, if I make a print of an
    >electronically-generated 127 patch, should this look like my 18% grey
    >card, or some other grey shade?


    Make a computer-generated grid of 256 squares, with each value from 0 to
    255 (or skip 0 to 70, or so, to save ink). Print it, then see which one
    is the same as the grey card. Using the same printer and paper, and
    driver settings, print a whole page at that value, and you should get
    18% grey. Don't use glossy paper, and don't expect the print to stay
    the same over time.

    >I know in practice things won't work out perfectly, but I want to
    >understand what I should be aiming for.
    >
    >TIA.


    --

    <>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<>
    John P Sheehy <>
    ><<> <>>< <>>< ><<> <>>< ><<> ><<> <>><
     
    , Feb 25, 2004
    #7
  8. "Alan F Cross" <> wrote in message
    news:p...
    > In message <4a2%b.1911$>,
    > Journalist-North <> writes
    > >

    (snip)
    >
    > How do you produce an 18% grey in Photoshop, then? It can't be a scan,
    > because that will be off by an unknown amount. Are you just holding up a
    > grey card next to the screen, and tweaking till they look similar? The
    > values will depend on gamma and to some extent colour space.
    >
    > Found this Q&A on Google:
    >
    > What is the calculation that yields 18% grey = 50% tonality, 128 on an
    > 8-bit grayscale, etc?
    >
    > That depends on the gamma of the color space used. ColorMatch, a 1.8
    > gamma color space, will have 18% gray at around 103. sRGB and Adobe RGB
    > (1998), both 2.2 gamma spaces, put it at around 125.
    >
    > So clearly - it depends!!
    >
    > Alan F Cross

    --------

    Actually Alan the K value of CMYK is the target (see notes below), as well
    as the indicative Luminance value (127) as the RGB gray value has you
    working in RGB colour space rather than a black and white colour space (e.g
    Greyscale = Grey1.8G or Grey2.2G). As a colour photo will necessarily be
    working in a colour space then the following Adobe procedure can be used to
    correct the colour (using the mid tone grey as the reference)

    You can construct / simulate mid grey in Photoshop by creating a step wedge
    and measuring the mid grey colour values as generated on screen; as received
    from a scanner; and compared visually to the output from a printer, as well,
    against a standard 18% grey card in ANY uniform light condition.

    SEE PHOTOSHOP HELP: Eliminating unwanted colour casts

    --------

    This is quoted from the help screen:

    If your scanned image contains an unwanted colour cast, you can perform a
    simple test to determine whether the cast was introduced by your scanner. If
    it was, you can use the same test file to create a colour-cast correction
    for all images scanned with the scanner.

    To identify and correct a colour cast introduced by a scanner:

    1) Make sure that your monitor has been calibrated. (See Creating an ICC
    monitor profile.)
    2) Open a new Photoshop file, and use the linear gradient tool to create a
    blend from pure black to pure white.
    3) Choose Image > Adjust > Posterize, and posterize the blend using 11
    levels.
    4) Print the 11-step gray wedge on a black-and-white printer, and then scan
    it into Photoshop.

    Note: You can also perform this test using an 18-percent neutral gray card
    or an 11-step gray wedge from a photography store.

    5) Open the Info palette, and read the RGB values on-screen for each of the
    gray levels.
    >>> Uneven R, G, and B values indicate a colour cast. <<<

    6) Use Levels or Curves to correct the colour cast, and then save the dialog
    box settings. (See 4. Adjust the colour balance..)
    7) Open the scanned image you want to correct, reopen the dialog box you
    used to correct the cast in step 6, and load the saved settings.

    --------

    The values I quoted as

    R = 133-135 G = 133-135 B = 133-135 >>Lum = 127<< Sat = 160 Hue = 0

    C = 0; M = 0; Y = 0; K = 54%

    are nearly correct (in absolute terms) for mid grey, that is, the mid point
    in the grey scale or on the step wedge constructed above, between pure white
    and pure black, and working in an RGB colour space

    You will note that all the RGB values are the same (per Adobe they are
    perfectly balanced though slightly denser than 127 in all channels -
    possibly due to my calibration settings); they are not shifted by the hue
    control (value is zero) but the luminance value is 127 which is dead centre
    between the white (255) and black (0) COLOUR values. Thus, I do have a 3 way
    colour balance and the Luminance is the mid grey / 50% reflectance value.

    The K (black ink) value I quoted is "near" to 50% (50% black ink density
    printed on a pure white paper = mid grey) but the exact value I measured
    (54%) may be off a bit, again due to my calibration settings. The K value is
    a printer's reference value for mid grey.

    If you also look at the step wedge extremes you will see that white has
    values of R = 255; G = 255; B = 255; and Lum around that value (typically
    240+ depending on your calibration) = 100% reflectance; whereas the black
    end will have values of very near R = 0; G = 0; B = 0; and Lum = 0 / = 0%
    reflectance

    OTHER NOTES:

    Note 1) As a side note...for commercial printing to paper, or for that
    matter printing to your ink-jet which in fact actually prints in CMYK as
    well, the K value is NOT the ONLY reference point as "absolute" black can be
    improved (see note 2, below, on 4 colour B&W printing with 4 black inks) in
    how it appears in on the page in a publication by slightly shifting other
    colour values; but, this correction is almost NEVER made (in Photoshop or
    any other image editor) before the image is referenced to a layout
    application (such as QuarkXPress - where the colour values can be further
    adjusted in the pre-press workflow). These corrections will not be seen on
    screen, or even noticeable, but can be observed at the output of a printing
    press and printers describe this as the "blacker than black" effect. That
    is, the slight colour shift introduced (usually a bit of a CMY component
    shift towards magenta) actually produces, visually to the human eye, a
    blacker black on the page than just printing with "black" ink. So I am told,
    it is more akin to matching the appearance of the paper printed image to the
    non-linear gamma value of the human eye - more an arcane printer's "art"
    than science at work here - and, of course, it depends to some degree on the
    paper being used taking account of the whiteness and the non-white colour
    and reflectance values of that against the semi-transparent inks being laid
    down by the press - it also allows for "black" to be overprinted on, or
    joined to, other colours and still look "black."

    Note 2) 4 colour B&W ink-jet printing with all black ink cartridges
    installed. Some ink-jet printers can be fed 4 black toned inks for only
    printing photos in B&W. but allowing a wide gamut of black/grey toning; this
    allows for almost the same thing, in fact it is the exact same thing, as the
    printers are doing above by shifting from hard blacks to softer velvety
    tones or vice versa. They are all "black" inks in the cartridges, yes, but
    with slightly different colour shifts in their formulations. Ordinary paint
    works the same way as printer ink - go into any paint store and ask for the
    formula of a darkish shade of "charcoal grey" - more then likely you will
    find a formula that looks something like this: "mix black 90% with white
    9.9% and add yellow 0.1%." That ever so slight yellow shift does wonders
    when the paint is viewed in daylight conditions. You could never tell the
    difference between the above mix in the paint bucket and a mix of 90/10 B&W,
    only, in a second bucket...but you sure can see it once it is painted on a
    building.

    Journalist
     
    Journalist-North, Feb 26, 2004
    #8
  9. Alan F Cross wrote:

    > Talking theoretically ...
    >
    > If I take a shot of an 18% grey card, should I find this to be at level
    > 127 (on an 8-bit grey-scale)?
    >
    > Coming at the question from the other end, if I make a print of an
    > electronically-generated 127 patch, should this look like my 18% grey
    > card, or some other grey shade?
    >
    > I know in practice things won't work out perfectly, but I want to
    > understand what I should be aiming for.
    >
    > TIA.
    > --
    > Alan F Cross


    Remember that an 18% grey card isn't NECESSARILY neutral in hue.
    It is designed to have a standard REFLECTANCE equal to a middle grey
    tone.

    It was originally designed to help a photographer obtain proper exposure
    ,
    not proper white balance.
    I have two standard 18% grey cards, one by Kodak and one by MacBeth and
    they have noticeably different hues.
    However, they both give identical EXPOSURE readings on my camera because
    they both have the same reflectivity.

    Computer wise, we have 256 shades of grey, so I interpret "Middle Grey"
    as
    R=128, G=128, and B=128.
    I create a sheet in Photoshop and fill it with a swatch that is
    128,128,128, but I then convert it to Greyscale to remove ANY color.
    Then I print it on Matte paper with Black ink just to make sure that my
    printer does not introduce any color artifacts. This print CANNOT have a
    color bias because it has NO color at all. The resulting sheet gives the
    same reading as my Kodak Grey card and should also be perfect for White
    Balance, but I have never tried it for this purpose.
    Bob Williams
     
    Robert E. Williams, Feb 26, 2004
    #9
    1. Advertising

Want to reply to this thread or ask your own question?

It takes just 2 minutes to sign up (and it's free!). Just click the sign up button to choose a username and then you can ask your own questions on the forum.
Similar Threads
  1. Brian Bergin

    127.0.0.1 log entries

    Brian Bergin, Nov 21, 2003, in forum: Cisco
    Replies:
    7
    Views:
    993
  2. REMA

    ftp site 210.83.127.80

    REMA, Sep 3, 2003, in forum: MCSE
    Replies:
    3
    Views:
    945
  3. ha®®y

    127.0.0.1

    ha®®y, Aug 16, 2003, in forum: Computer Support
    Replies:
    7
    Views:
    633
    °Mike°
    Aug 16, 2003
  4. ROBERT ARMSTRONG

    The 127.0.0.1 mystery

    ROBERT ARMSTRONG, Nov 24, 2003, in forum: Computer Support
    Replies:
    1
    Views:
    503
    °Mike°
    Nov 24, 2003
  5. Alan F Cross
    Replies:
    40
    Views:
    1,849
    zbzbzb
    Mar 2, 2004
Loading...

Share This Page