6 apr 2007 kl. 17.07 skrev ADavidhazy:
Hi, I am not sure we discussed this on the list sometime in the
past so if we did forgive me. Someone posed the following question
to me:
"What is the RGB value for 18% gray (or 18% grey for that matter)
in Photoshop?"
This simple question led to three somewhat different answers from
three instructors in the vicinity.
One simply said: 177
Another said:
RGB is linear scale. It all depends on how your camera is set up. If
it is linear from 0-256 - some cameras are set up this way - then
its
18% of 256 or 46!
And yet another:
The only thing I have to add is that the sensors are linear
(mostly) and
so is the associated RAW data. If you has access to such RAW data,
then
Don's calculations are valid.
However, the data the user generally has for use in programs like
Photoshop (is invariably GAMMA CORRECTED for perceptual uniformity
during image editing... The gamma correction used will vary by the
type
of color space used (2.2 for sRGB/A-RGB but 1.8 for Profoto RGB).
(Incidentally, the moment a RAW file is opened in PS, it is GAMMA
corrected).
So basically, one has to undo this GAMMA correction before assuming
linear data. After that, you take 18% of whatever your scale is
(0-255
for 8 bit, 0-65535 for 16 bit etc)
P.S.: Incidentally, meters don't use 18% gray for calibration...I
think
ANSI standards are more like 12%. But then, does anyone use hand-held
meters anymore in this digital world :-)
...... so, what to make of all this? Anyone on the list have
something to contribute???
;)
andy
Wow; don´t know if I have even 18% of gray matter available between
my ears right now....
Anyway, my limited readings on the subject leads me to believe the
third alternative is correct: it depends on the colour space used
(mainly the gamma value).
I did a quick experiment with Photoshop; opened a new document in my
default working colour space (which is ProPhoto, since that´s what
Lightroom uses). Then I drew rectangles with the marquee tool and
filled them up with the bucket. First one I set R,G,B all to 46.
Way too dark! (no surprise...) K value was 75% in this space.
Next one got 177 on all three: a bit too light; K value26%.
Third one got 128 on all three; looked about right, like it should.
K value was 43%.
Last, I created a new layer in Overlay mode, and checked the "fill
with overlay neutral (50% gray)" option. Then changed mode to Normal
and dragged the layer a bi to the side, so I could see the other
squares. A little darker than the last square, but still looking
about right (no use putting a Kodak card beside a LCD screen, so I
have to guess from memory...).
The K value of the new layer was 50%, just as it should, but the RGB
values were all 110.
So, yes, it depends on the space. I might try the same with
different spaces (even use e.g. a printer profile as working space,
which would probably make R, G and B unequal for a visibly neutral tone.
Phew... Now I´ve spent most of that 18% gray matter for tonight...
Per Öfverbeck
http://ofverbeck.se
"In a world without walls or fences, who needs Windows or Gates?"
