Grayscale with RGB values?

  • Thread starter Thread starter Frank Kolwicz
  • Start date Start date
F

Frank Kolwicz

All I want to do is check my scanners handling of gray tones in color.
I've bought and tried Stouffer and Kodak grayscales, both opaque and
transparent, and I have no way of telling just how neutral they are
(the opaque ones are obviously brown). This seems to be the case of
the IT-8 color charts, too: the ones I've seen have brown "gray"
scales.

A grayscale with calibrated colorimeter or spectrophotomer RGB values
would be perfect, but I haven't been able to find one. I know
individuals are making them in testing inksets for dedicated B&W
printers, but I can't find any for sale.

Suggestions? Reply here or to (e-mail address removed), my Google address is
dead and I can't access my settings to change it at the moment.

Thanks,

Frank
 
All I want to do is check my scanners handling of gray tones in color.
Click to expand...

Well, with a RGB scan of your grayscale references, Adobe Photoshop (or
Elements) has Info palette, shows exact blend of RGB anywhere you run the
mouse over, even pixel by pixel, if want to zoom in that far. If not
equal values of each, then will quickly show what the bias is. I suppose
some other image editors may offer the same?

Or have I misinterpreted your quest?
 
Perhaps he means that a scanned image of a gray scale does not show equal
RGB values. Once upon a time I decided to scan the Kodak 18% gray card, and
I found that even though the image looked gray the RGB values were not
equal.

To me, it just showed that you are not supposed to scan those things because
they look different to a scanner than they do to film.

Jim
 
Perhaps he means that a scanned image of a gray scale does not show equal
RGB values. Once upon a time I decided to scan the Kodak 18% gray card, and
I found that even though the image looked gray the RGB values were not
equal.

To me, it just showed that you are not supposed to scan those things because
they look different to a scanner than they do to film.

Jim
Click to expand...


That was what I though he meant.
And the Info tool would point that out.

Mac
======
 
Jim said:
Perhaps he means that a scanned image of a gray scale does not show equal
RGB values. Once upon a time I decided to scan the Kodak 18% gray card, and
I found that even though the image looked gray the RGB values were not
equal.
Click to expand...

Those Neutral gray references only reflect/transmit equal amounts of the
incident light (reduced silver, or carbon, usually offers such broad
spectral uniform absorption). If the light isn't seen as neutral to the CCD,
then the values will differ. However, in linear gamma, the relative
difference must be (close to) equal for all luminances/brightnesses covered.

IT8 targets don't really have to be exactly neutral in their GS steps,
because the deviation is given in the reference data file. However, they
should preferrably be close to the standard values, because that will give a
good spread of hues and luminances and allow better interpolation.

Bart
 
Sorry I haven't answered sooner - other things, you know?

Bart seems to have gotten it: I don't want to scan any old thing
(Kodak gray cards, stepscales, IT8s, etc.) UNLESS (it's a BIG unless)
it comes with RGB or other color calibration information that I can
use for comparison to my B&W prints and test images. Those mentioned
items are NOT color calibrated, the "gray" scales that I have seen
(Kodak and Stouffer and somebody's IT8) are really brownscales and no
one seems to know or care how brown they are, they are for another
purpose entirely.

I want to have something I can hold up to the light or scan alongside
my prints to check the prints' neutrality or lack of it. Without
something to calibrate against directly, my eyes are not good enough,
nor is my scanner.

It's a simple idea: a known RGB object for standardization.

I have finally found something that seems to fill the bill (a $55
bill, in fact). Munsell Color, a division of GretagMacbeth, sells a
"Neutral Value Scale", in matte and glossy versions, which have color
values calibrated in Munsell notation. A free program is available
from them for conversion of those values to standard RGB, and other
color values as we know them.

So, it's about 5 times as expensive as I had hoped, but maybe it will
suit the task.

Thank you for you replies,

Frank
 
Bart seems to have gotten it: I don't want to scan any old thing
(Kodak gray cards, stepscales, IT8s, etc.) UNLESS (it's a BIG unless)
it comes with RGB or other color calibration information that I can
use for comparison to my B&W prints and test images. Those mentioned
items are NOT color calibrated, the "gray" scales that I have seen
(Kodak and Stouffer and somebody's IT8) are really brownscales and no
one seems to know or care how brown they are, they are for another
purpose entirely.
Click to expand...

Sorry, you will not find here what you want. Some things you should
think of:

1. What light source is used to compare the test chart and the print?
Of course you need a test chart that is neutral for that specific
light source in order to have a correct reference. Grayscale targets
like the Kodak Q10 or so do appear neutral under most light sources
and thus are a cheap visual reference. Otherwise you basicly have to
use a target that let's say is neutral under D50 standard light source
and use it as reference... even if it does not appear perfectly
neutral under your actual light source. I do not know what print inks
you use, but for instance dyes used in modern ink jet printers can
appear very different with changing light sources.

2. So even if you do adjust your prints to match the grayscale target
under the specific light source, it might not be the case under a
different light source.

3. You say you want RGB values? RGB is a device dependent color space.
If you want RGB values with your target, than you also have to specify
what type of RGB standard/device you mean (sRGB?). The choosen RGB
space should be big enough to include all colors in the test chart
(sRGB fails to do that already for some color patches of the IT 8/7/2
targets... and compared to slide films the color gamut of sRGB is a
nightmare).

4. Regarding your mentioned IT 8 targets:

Note that with IT 8 targets you do get a reference file that does
contain the measured color values. Once you define a wanted RGB space,
you can calculate the RGB values from the XYZ color values listed in
all IT 8 target reference files. Calculating color values for
different light sources from the IT 8 reference files usualy does
result in an added fault unless you do get the spectral data for the
target.

Printing a perfectly neutral grayscale is extremly hard to achive
especialy on photo paper used for IT 8 targets as there is no black
dye. While the IT 8 standard does define a certain fault tolerance
that must be meet, you will find that most IT 8 manufacturers have
big problems getting small fault values and remaining faults are well
visible. Actually, if you are looking for an IT 8 target with extremly
small faults, check out my targets on http://www.targets.coloraid.de
and download the currently shipping reference file for charge R031126.
Compare the color values provided in the reference files with other
manufacturers and you should be impressed ;-) I can provide spectral
data for the targets in case you want to calculate things for
different observers, light sources or even RGB values (very easy to
do with programs like X-Rites ColorShop). But I think IT 8 targets are
not really what you need...
I want to have something I can hold up to the light or scan alongside
my prints to check the prints' neutrality or lack of it. Without
something to calibrate against directly, my eyes are not good enough,
nor is my scanner.
Click to expand...

I guess, the cheapest way really is to use a Kodak Q10 target and do
it visualy under the final light source used and forget the RGB
numbers. That is, if you do not want to invest in
spectrometer/profiler.... But maybe someone else has a better idea
here that is also affordable for you....
 
Back
Top