How to remove black from RGB files
Tony
GSp
Tony <ton...@telus.net> wrote in message news:3B8C9ED8...@telus.net...
Mr. Jim
>In RGB files, how can you remove black from colored areas?
Convert the image to CMYK, fill the black channel with white,
and then convert back to RGB.
--
Jim Hall
Tony
Thanks a lot, Jim and Gsp! Those were fast responses!
JIm: I was told by two graphics professionals that it's not good to
switch back and forth between RGB and CYMK because every time you do
that, there's a risk of losing some color data that is irretrievable.
I'm not sure why or how this would occur. Maybe somebody could clarify
this.
Tony
Mr. Jim
>JIm: I was told by two graphics professionals that it's not good to
>switch back and forth between RGB and CYMK because every time you do
>that, there's a risk of losing some color data that is irretrievable.
That's mostly true ... CMYK is a colorspace. RGB is a different colorspace.
Converting between the two will tend to clip/cut/lose colors.
The spaces overlap, but any portion that does not overlap will be lost!
There are times when you will want to swap back and forth, for color
correction.
--
Jim Hall
Tony
> Selection by color, then change or delete
Thanks, G.S.P.
Do you mean select by color range? I don't see how that would work.
If if select by color range in my sky, I select the area of sky that contains
the small percentage of black (as well as large percentages of cyan and
magenta) but then how do I delete the black?
Tony
Tony
Thanks, Jim.
I also thought of selecting my RGB sky, copying it, pasting it into a new
file, converting to CMYK, deleting the black using Levels, converting back to
RGB, pasting the sky into the original file and merging the two layers, but
I'm concerned that by doing that I may lose some of the cyan or magenta in the
sky.
Tacit
>the small percentage of black (as well as large percentages of cyan and
>magenta) but then how do I delete the black?
In an RGB file? You don't.
If you wish to color-correct in CMYK, *work in CMYK.* Don't work in RGB.
Separate the file to CMYK, using a separation setup that's appropriate for the
job. Then leave it in CMYK! Continue to work with it that way. Don't try to
color correct in RGB if your ultimate goal is CMYK color correction.
Also, when you look at the CMYK stats for an RGB file, be aware that the
settings in the CMYK Setup dialog WILL affect the numbers you see! Too much
black overall? Then maybe you should be using light black generation, not
medium black generation.
Incidentally, if you do convert to CMYK, don't convert back.
There are colors in RGB that are not in CMYK, and there are colors in CMYK that
are not in RGB. It should never be necessary to convert more than once between
color spaces. If you do, the resulting image will contain only those colors
that exist in both CMYK *and* RGB. This is A Bad Thing.
If for some reason you need to separate the file again, convert from CMYK to
Lab, not RGB; then back to CMYK. By going to Lab, you don't clip colors.
And finally: If you convert an RGB image to CMYK, remove all the black from an
area, then go back to RGB, when you go to CMYK again you may find black in
those areas again! The black plate is generated anew each time you separate
from RGB to CMYK. The black plate is created based on the RGB values and on the
parameters in the Separation Setup dialog. It won't "remember" that you removed
black before.
--
Literary forums; Onyx, a game of sexual exploration; Xero, the industrial
magazine of art, fiction and photography; fine-art photo gallery--all at
http://www.xeromag.com/franklin.html
Tony
> Separate the file to CMYK, using a separation setup that's appropriate for the
> job.
Thanks, Tacit.
I'm preparing files to print on an HP-5000. How do I set up separation?
Tony
Tacit
Most inkjet printers expect RGB data, not CMYK data.
The numbers you see in Photoshop will not represent the numbers that your
inkjet printer actually prints, because the printer performs its own
separation, using separation parameters different from Photoshop.
If you print CMYK data to an inkjet printer, it typically converts back to RGB,
then back again to CMYK.
So I guess the question is: What is your goal? What are you trying to do?
Directly controlling CMYK values printed to an inkjet printer is not always
easy.
Mike Russell
that CMYK is a major color correcting resource that many people ignore because of blanket statements
about data damage converting between color spaces.
Those who spend some time experimenting will discover a new "lost world" of color correcting power
waiting for them in CMYK as well as Lab. The black channel in particular provides leverage for both
color correction and control of shadow detail. Check out any book by Dan Margulis for more info.
I disagree with Tacit and others who are excellent contributors to this news group on this issue of
whether to routinely use CMYK for color correction. I routinely convert a variety of normal
photographs back and forth between CMYK and RGB with minimal changes to the image data.
Before going on, I'm going to propose that we begin, to illustrate our points with images, rather
than with words. I will follow up this article a web page that demonstrates conversion of an image
between RGB and CMYK dozens of times, and quantifies the changes that occur. I suggest that Tacit
and anyone else who wants to chime in do the same when illustrating the supposed degradation of
photographic images converting from RGB to and from CMYK images.
Yes, there is a gamut clip that may or may not occur on your first conversion to CMYK. This clip
will primarily affect bright, saturated blues. Once that clip has occurred subsequent conversions
have almost no effect on a normal photographic image. Note: in color settings, set your intent to
relative colorimetric, not perceptual, to minimize the loss of general color saturation.
Keeping in mind I'm talking about photographic images, not computer generated or graphic ones, I'll
address some of Tacit's points re use of CMYK:
--
"Tacit" <tac...@aol.com> wrote in message news:20010829133902...@mb-fi.aol.com...
> >If if select by color range in my sky, I select the area of sky that contains
> >the small percentage of black (as well as large percentages of cyan and
> >magenta) but then how do I delete the black?
>
> In an RGB file? You don't. If you wish to color-correct in CMYK, *work in CMYK.* Don't work in
RGB.
> Separate the file to CMYK, using a separation setup that's appropriate for the
> job. Then leave it in CMYK! Continue to work with it that way. Don't try to
> color correct in RGB if your ultimate goal is CMYK color correction.
It is in fact quite practical to go back and forth between the three major color spaces supported by
PhotoShop. This includes RGB, Lab, and CMYK.
> Also, when you look at the CMYK stats for an RGB file, be aware that the
> settings in the CMYK Setup dialog WILL affect the numbers you see! Too much
> black overall? Then maybe you should be using light black generation, not
> medium black generation.
Tacit and I agree here. The heaviness of the black plate is important variable, and it is equally
important that the intent setting be relative colrimetric (relcol for short) and not perceptual.
Note: How to change the black plate? in PS6, you may set the heaviness of the black plate using
"Edit>Color Settings...". Click on your CMYK settings, and select "Custom" from the drop down list.
Then make sure the GCR (stands for gray component removal) radio button has a dot in it, and select
anything from "None" to "Maximum" for the amount of black plate you want. I generally leave this
set to "Light", to allow more leeway for color correction.
> Incidentally, if you do convert to CMYK, don't convert back.
I disagree with this recommendation, of course. I do this all the time, both for images destined
for web use, and for those that I intend to print. It's necessary to be careful of pure blue
colors, but this is no reason to avoid CMYK ,even if your eventual destination is RGB.
> There are colors in RGB that are not in CMYK, and there are colors in CMYK that
> are not in RGB. It should never be necessary to convert more than once between
> color spaces. If you do, the resulting image will contain only those colors
> that exist in both CMYK *and* RGB. This is A Bad Thing.
If your eventual intent is to print the image, then you are going to be confined to CMYK at anyway.
Printed images, generally, are done at a higher quality level, both in terms of tonality and pixels
per inch. This sort of caution when converting color spaces is appropriate when converting to and
from indexed color modes, but inappropriate going to and from CMYK and Lab.
> If for some reason you need to separate the file again, convert from CMYK to
> Lab, not RGB; then back to CMYK. By going to Lab, you don't clip colors.
More specifcally, you may lose very bright yellows going from RGB to CMYK. Besides the fact that
these yellows are almost un-seeable by the human eye, they are impossible to capture in a
photograph - though very rarely we may decide to use these yellows in a CMYK version of the image.
An example of this would be a full page add featuring a photograph of a lemon that we want to jump
off the page. Clipping these colors in RGB is not a big deal, as long as we get a good yellow.
I would further add that there is an insignificant, but measurable data change doing an
RGB->Lab->RGB conversion. Try the following experiment: Dup your RGB image, convert the copy to
and from Lab. Use Image>Adjust>Calculations to difference the red channels, and store the result in
a new document.
Look at the histogram of the difference and you will see that a goodly number of pixels have changed
value. Normally we ignore this small variation, and I am arguing that RGB->CMYK conversions is in
the same ballpark, particularly after the first RGB->CMYK conversion.
> And finally: If you convert an RGB image to CMYK, remove all the black from an
> area, then go back to RGB, when you go to CMYK again you may find black in
> those areas again! The black plate is generated anew each time you separate
> from RGB to CMYK. The black plate is created based on the RGB values and on the
> parameters in the Separation Setup dialog. It won't "remember" that you removed
> black before.
Quite true. For a variety of reasons, printed images look better if you mix a bit of color into
your grays, so deleting the K channel will only delete some of the gray color. You may minimize
this effect by setting your GCR to "Maximum". This will remove all but a tiny amount of gray from
the image, and selecting the amount of GCR will control the black-removal effect that the subject of
this thread refers to.
My final point: CMYK is a major resource, as are Lab and RGB. You paid for them, so use them.
Experiment and I think you'll find you can make your images look better, whether they are destined
for the web or printing on an inkjet.
Now I've promised a web page of example images, and will post to this thread when it is ready. I
will provide web space for others joining this discussion who wish to provide images.
Mike Russell
> I'm preparing files to print on an HP-5000. How do I set up separation?
Consumer inkjets will not allow you to directly control the CMYK ink that reaches the page. You may
print directly to CMYK using a PostScript printer, or a software RIP such as Adobe PressReady (now
discontinued, unfortunately).
Start by printing the CMYK image from PhotoShop - some black will leak through but you will get a
close approximation to what you need.
There is a more refined method if you want to experiment further, you can print the image using
several passes, just the way a real printing press does. First print your yellow image, then
magenta, then cyan, feeding the paper through separately each time.
As Tacit mentioned, you will not get 100% control this way, for example your cyan may have bits of
the other colors in it. But you will elimiate most if not all of the black, and get somewhat closer
than you would printing just the RGB image.
There are additional things you can do to purify the colors, using a microscope and a little more
work in PhotoShop setting up the printer curves. My guess is you do not need to go this far for
what you're doing.
Tony
> Yes, there is a gamut clip that may or may not occur on your first conversion to CMYK. This clip
> will primarily affect bright, saturated blues. Once that clip has occurred subsequent conversions
> have almost no effect on a normal photographic image. Note: in color settings, set your intent to
> relative colorimetric, not perceptual, to minimize the loss of general color saturation.
How do you do that? It's not in File > Color Settings in Photoshop 5.
> Keeping in mind I'm talking about photographic images, not computer generated or graphic ones,
I'm referring to my scanned, fine art images, which I assume you would categorize as graphic, so I'm
not sure that that advice applies in my case.
> Tacit and I agree here. The heaviness of the black plate is important variable, and it is equally
> important that the intent setting be relative colrimetric (relcol for short) and not perceptual.
> Note: How to change the black plate? in PS6, you may set the heaviness of the black plate using
> "Edit>Color Settings...". Click on your CMYK settings, and select "Custom" from the drop down list.
> Then make sure the GCR (stands for gray component removal) radio button has a dot in it, and select
> anything from "None" to "Maximum" for the amount of black plate you want. I generally leave this
> set to "Light", to allow more leeway for color correction.
I did that and it worked! All the black is gone from my sky!
> > Incidentally, if you do convert to CMYK, don't convert back.
>
> I disagree with this recommendation, of course. I do this all the time, both for images destined
> for web use, and for those that I intend to print. It's necessary to be careful of pure blue
> colors, but this is no reason to avoid CMYK ,even if your eventual destination is RGB.
>
> > There are colors in RGB that are not in CMYK, and there are colors in CMYK that
> > are not in RGB. It should never be necessary to convert more than once between
> > color spaces. If you do, the resulting image will contain only those colors
> > that exist in both CMYK *and* RGB. This is A Bad Thing.
>
> If your eventual intent is to print the image, then you are going to be confined to CMYK at anyway.
But I like working in RGB because the colors are nicely bright and they look closer to the proofs that
have been printed for me so far on this job.
> More specifcally, you may lose very bright yellows going from RGB to CMYK. Besides the fact that
> these yellows are almost un-seeable by the human eye, they are impossible to capture in a
> photograph - though very rarely we may decide to use these yellows in a CMYK version of the image.
> An example of this would be a full page add featuring a photograph of a lemon that we want to jump
> off the page. Clipping these colors in RGB is not a big deal, as long as we get a good yellow.
But sometimes I want very bright yellows -- and blues. I'm talking here about printing limited edition
art prints , and consequently the colors must be of the finest calibre.
Thanks for your comprehensive response, MIke.
Tony
Tacit
>in this group. I believe
>that CMYK is a major color correcting resource that many people ignore because
>of blanket statements
>about data damage converting between color spaces.
I agree completely. 90% of the work I do in Photoshop is in CMYK. I rarely use
RGB; when I do, it is for Web work. I never work on any file intended for print
in RGB; it's scanned in CMYK, retouched in CMYK, color-corrected in CMYK,
proofed in CMYK, and printed in CMYK.
>I disagree with Tacit and others who are excellent contributors to this
>news group on this issue of
>whether to routinely use CMYK for color correction. I routinely convert
>a variety of normal
>photographs back and forth between CMYK and RGB with minimal changes to
>the image data.
>
>Before going on, I'm going to propose that we begin, to illustrate our points
>with images, rather
>than with words. I will follow up this article a web page that demonstrates
>conversion of an image
>between RGB and CMYK dozens of times, and quantifies the changes that occur.
I would suggest that such a comparison may not be valid, for two reasons:
1. A JPEG image on the Web will always have data loss, and this degredation may
mask degredation in the uncompressed image.
2. My argument is specifically against converting a CMYK image to RGB. If you
do this, you clip the CMYK colors that are outside the RGB gamut. Since all
monitors and all Web browsers display only RGB, you cannot see this loss in an
RGB image. It can be seen only in a printed proof of a CMYK image. Ergo, a Web
page will never reveal this particular form of image degredation.
>Yes, there is a gamut clip that may or may not occur on your first conversion
>to CMYK. This clip
>will primarily affect bright, saturated blues. Once that clip has occurred
>subsequent conversions
>have almost no effect on a normal photographic image.
This is correct with respect to an RGB->CMYK separation, but incorrect with
respect to a CMYK->RGB separation.
When you go from CMYK to RGB, the image loses detail in the shadows. This is
because CMYK can represent a much greater range of shadows than RGB can.
To see what I mean:
1. Create a new CMYK image.
2. Switch to the Cyan channel. Type the word "Shadow Detail" in 100% Cyan.
3. Switch tot eh Black channel. Fill with 100% black.
When you proof this image, you will be able to read the words "Shadow Detail,"
even though you cannot see them on your RGB monitor.
Now. convert to RGB and back to CMYK. The image becomes filled entirely with
Photoshop's default black. The words are gone.
In a real-life, less artifical setting, the same thing can still be seen.
Scan a transparency containing lots of deep shadow detail, ona drum scanner, in
CMYK.
Convert to RGB.
Convert back to CMYK. On your RGB monitor, it will look the same.
Now, print a proof of the original scan, and the CMYK->RGB->CMYK image.
The one that was never converted to RGB will have more visible shadow detail.
On your RGB monitor, they will look identical On the proof, they will not.
>It's necessary to be careful of pure blue
>colors, but this is no reason to avoid CMYK ,even if your eventual destination
>is RGB.
If your final destination is RGB, thee's no compelling reason to avoid CMYK,
except that CMYK has a restricted gamut; however, if your destination is CMYK,
there are compelling reasons (such as the one I outlined above) to avoid RGB.
>If your eventual intent is to print the image, then you are going to be
>confined to CMYK at anyway.
Yes. But by going to RGB, you sacrifice shadow detail and all colors that exist
in CMYK but not RGB.
>More specifcally, you may lose very bright yellows going from RGB to CMYK.
And you will lose vivid yellows from CMYK to RGB. You will also lose rich
blacks, dark browns, deep navy colors, and other rich, dark colors when going
from CMYK to RGB.
Tacit
>look closer to the proofs that
>have been printed for me so far on this job.
Then your monitor is not set correctly.
If your job is to be printed on press, you gain NOTHING by continuing to edit
in RGB. The image will reproduce CMYK; any rich RGB colors that exist in the
RGB image but can't be reproduced in CMYK will not be reproduced on the final,
printed piece.
Are you planning to print the final piece on a press, or on an inkjet printer?
If the answer is "on a press," then I submit you gain nothing by continuing to
work in RGB.
>But sometimes I want very bright yellows -- and blues. I'm talking here
>about printing limited edition
>art prints , and consequently the colors must be of the finest calibre.
If you are on a press, you don't get that option. CMYK can not reproduce vivid,
bright blues. You can't do it, no matter what color model you start out in. If
you wish to do this on a printing press, you must add a "touch plate"--an area
of blue ink. Otherwise, what you want is impossible; there are simply colors
that can't be reproduced in CMYK, no matter what.
Mike Russell
> I'm referring to my scanned, fine art images, which I assume you would categorize as graphic, so
I'm not sure that that advice applies in my case. <
Tony,
Fine art images on paper scanned in RGB are probably not a problem, but it does depend on the type
of work you are doing and how saturated your colors are.
> But sometimes I want very bright yellows -- and blues. I'm talking here about printing limited
edition art prints , and consequently the colors must be of the finest calibre.
I spotted an error in my previous post re which direction yellow clipping occurs. I should have
said saturated yellows suffer going from CMYK to RGB. Summarizing, RGB is great for pure blues,
CMYK is great for pure yellows.
Your final printed images can't really have RGB-style saturated blues, since you are using an
inkjet, but your blues will still look good, just a little dark. Change to special inks, or use
spot colors on a press, or for that matter output to slide film and you can get around this.
You may find the 3 pass method I mentioned before gives you the most control, especially over the
use of black. Most printers will inject substantial amounts black even for overlapping magenta and
cyan.
Fred Athearn
wrote:
>Convert the image to CMYK, fill the black channel with white,
>and then convert back to RGB.
Or just right click black channel and select delete which gives a
slightly different effect.
Fred Athearn
Fred Athearn
> If you convert an RGB image to CMYK, remove all the black from an
>area, then go back to RGB, when you go to CMYK again you may find black in
>those areas again! The black plate is generated anew each time you separate
>from RGB to CMYK
Wow the conversion really does make changes. If you convert back and
forth deleting the black channel over and over you very quickly see
some rather alarming changes in the image.
Fred Athearn
Tony
> >I'm preparing files to print on an HP-5000. How do I set up separation?
>
> Most inkjet printers expect RGB data, not CMYK data.
One of the employees at the service beareau said the HP-5000 -- a high-resolution
inkjet printer -- seems to print better colors in RGB, although one of the
partners of the company told me it makes no difference if I give them the files in
RGB or CMYK.
> The numbers you see in Photoshop will not represent the numbers that your
> inkjet printer actually prints, because the printer performs its own
> separation, using separation parameters different from Photoshop.
>
> If you print CMYK data to an inkjet printer, it typically converts back to RGB,
> then back again to CMYK.
>
> So I guess the question is: What is your goal? What are you trying to do?
I'm trying to get the art prints printed on the HP-5000 to look like good on
paper. The images were scanned onto my computer from my paintings.
Tony
Tony
> From: "Tony" <ton...@telus.net>
>
> > I'm preparing files to print on an HP-5000. How do I set up separation?
>
> Consumer inkjets will not allow you to directly control the CMYK ink that reaches the page.
The HP-5000 is a high-end inkjet printer that costs, I think, almost $10,000.
The gamut of the inkset is a bit narrower than normal because the inket is a UV inkset.
> You may
> print directly to CMYK using a PostScript printer, or a software RIP such as Adobe PressReady (now
> discontinued, unfortunately).
>
> Start by printing the CMYK image from PhotoShop - some black will leak through but you will get a
> close approximation to what you need.
>
> There is a more refined method if you want to experiment further, you can print the image using
> several passes, just the way a real printing press does. First print your yellow image, then
> magenta, then cyan, feeding the paper through separately each time.
>
> As Tacit mentioned, you will not get 100% control this way, for example your cyan may have bits of
> the other colors in it. But you will elimiate most if not all of the black, and get somewhat closer
> than you would printing just the RGB image.
I don't think the service beaureau prints that way. The sheet size is 48 by 36 inches, which makes it
awkward to feed the sheets through the printer four times for each sheet.
Tony
Tony
> "Tony" <ton...@telus.net> wrote in message news:3B8D6F21...@telus.net...
>
> > I'm referring to my scanned, fine art images, which I assume you would categorize as graphic, so
> I'm not sure that that advice applies in my case. <
>
> Tony,
>
> Fine art images on paper scanned in RGB are probably not a problem, but it does depend on the type
> of work you are doing and how saturated your colors are.
>
> > But sometimes I want very bright yellows -- and blues. I'm talking here about printing limited
> edition art prints , and consequently the colors must be of the finest calibre.
>
> I spotted an error in my previous post re which direction yellow clipping occurs. I should have
> said saturated yellows suffer going from CMYK to RGB. Summarizing, RGB is great for pure blues,
> CMYK is great for pure yellows.
>
> Your final printed images can't really have RGB-style saturated blues, since you are using an
> inkjet, but your blues will still look good, just a little dark. Change to special inks, or use
> spot colors on a press, or for that matter output to slide film and you can get around this.
I've chosen fade-resistant ultraviolet inks for printing on the HP-5000, which narrows the gamut, but
the printer uses six colors, which compensates somewhat for the narrow gamut.
I don't think I can run spot colors because it's not a litho press.
Tony
> >
> > There is a more refined method if you want to experiment further, you can print the image using
> > several passes, just the way a real printing press does. First print your yellow image, then
> > magenta, then cyan, feeding the paper through separately each time.
> >
> > As Tacit mentioned, you will not get 100% control this way, for example your cyan may have bits of
> > the other colors in it. But you will elimiate most if not all of the black, and get somewhat closer
> > than you would printing just the RGB image.
>
> I don't think the service beaureau prints that way. The sheet size is 48 by 36 inches, which makes it
> awkward to feed the sheets through the printer four times for each sheet.
>
> Tony
Actually, the HP-5000 is going to print my job on watercolor paper using a roll of paper three feet wide,
so the idea of using four passes (or six for this six-colour printer) isn't feasible.
Tony
Tony
> Note: in color settings, set your intent to
> relative colorimetric, not perceptual, to minimize the loss of general color saturation.
Sometimes when I open Photoshop (I don't know why it's only sometimes) a color conversion window
appears.
It says:
The embedded profile does not match the current RGB setup. Specify desired input conversion:
Input Conversion:
From: Custom RGB
To: RGB
Engine: Built-in
Intent: Perceptual (Images)
Black Point Compensation is checked.
I have been told to always choose "Don't Convert", and that's what I always do.
I was told by the service bureau that every combination of inks and paper have different needs for
color correcting, and because of that I'm not choosing ICC profiles.
Instead, i'm getting proofs done on the HP-5000.
Then I make an adjustment layer (or layers) in Photoshop to make the monitor limages ook like the
proofs.
Then I make an adjustment layer (or layers) in Photoshop to make the monitor images look the way I want
them to.
Then I hide the first set of adjustment layers and choose "Save a Copy", which makes an EPS file of the
layered .PSD file.
Then I copy those EPS files onto c.d.s and give them to the service bureau to make another set of
proofs.
I'm going to continue this process until the proofs look the way I want them to.
I hope I'm doing everything correctly.
Tacit
high-resolution
He's correct.
>although one of the
>partners of the company told me it makes no difference if I give them the
>files in RGB or CMYK.
Quite likely, if you give them CMYK files, they just convert to RGB.
Here's another question, though. Since you're making fine-art reproductions,
have you considered going with a giclee (pronounced GEE-klay) process, using an
Iris printer, rather than reproducing your work on an HP printer?
The HP printer was really designed more for banners and trade-show displays
than for fine-art reproduction. Many companies create giclee fine-art
reproductions; these giclees, which use a similar process, are made on an Iris
printer--it's kind of like an inkjet printer in some ways, but it was designed
and built specifically for high-quality fine-art reproduction.
A giclee typically costs around $90 to make, provided you supply the scan.
Giclees can be printed on watercolor paper, canvas, and other materials. A
giclee offers a much wider color range than an HP inkjet print, at much higher
resolution, with superior fade resistance, UV resistance, and permanence.
One company I've dealt with that does excellent fine-art giclees is called
Gamma One Conversions. Their Web site is at
As an artist, you will probably be much happier with a giclee than with an HP
inkjet print. Giclee reproductions are suitable for presentation in galleries;
they're that good.
Tacit
>color conversion window
>appears.
>
>It says:
>
>The embedded profile does not match the current RGB setup.
This happens because:
1. You are instructing Photoshop to save a copy of your color settings inside
your files whenever you save a file; and
2. You have instructed Photoshop to ask you what to do whenever the color
setting you are using does not match the color profile saved in your files.
You can change these settings. In Photoshop 5, use the File->Color
Settings->Profile Setup command. Turn off all embedded profiles, and tell
Photoshop to ignore all profile mismatches.
In Photoshop 6, use the Edit->Color Settings command, choose Advanced, and turn
the Color Management Policies to "Preserve Embedded Profiles" or "Off."
Mike Russell
CMYK: http://geigy.2y.net/DigPhoto/CMYK-RGB.htm
I would appreciate your comments.
Matti Vuori
>
> I have put up a page describing the results of converting a photographic image to and from RGB to
> CMYK: http://geigy.2y.net/DigPhoto/CMYK-RGB.htm
>
> I would appreciate your comments.
I would not draw much conclusions from one file, which as you say does't
contain any saturated blue (or much saturated anything). Why not try a
file of which people would say "now that file's going to suffer!".
Also, doing the conversion 10 times should not prove anything, since if
there will be any gamut clipping, it will happen during the first
conversion.
--
Matti Vuori, <http://sivut.koti.soon.fi/mvuori>
Vapaana työmarkkinoilla!
<http://sivut.koti.soon.fi/mvuori/ammattilainen/vapaana.htm>
Kuvataide-uutta! Kuvasarja "Zen. Kuvia ei-mistään"
<http://sivut.koti.soon.fi/mvuori/grafiikka/zen/>
Mr. Jim
>http://geigy.2y.net/DigPhoto/CMYK-RGB.htm
The graphics need work, but that is possibly due to bandwidth
considerations, I know.
--
Jim Hall
Mike Russell
> Mike Russell wrote:
> >
> > I have put up a page describing the results of converting a photographic image to and from RGB
to
> > CMYK: http://geigy.2y.net/DigPhoto/CMYK-RGB.htm
> >
> > I would appreciate your comments.
>
> I would not draw much conclusions from one file, which as you say does't
> contain any saturated blue (or much saturated anything). Why not try a
> file of which people would say "now that file's going to suffer!".
Matti,
You'd be surprised - this file was chosen because someone else said "sure it'll work with bright
saturated colors, but try one with shadows and you'll see an immediate deterioration." Still, your
point is well taken, and perhaps someone will provide me with an image, or try the experiment
themselves on an image of their own choosing.
> Also, doing the conversion 10 times should not prove anything, since if
> there will be any gamut clipping, it will happen during the first
> conversion.
I agree with this completely, and in fact this is one of the main conclusions I came to: that there
is no large cumulative degradation with each RGB->CMYK->RGB conversion.
Tacit
>it'll work with bright
>saturated colors, but try one with shadows and you'll see an immediate
deterioration."
Actually, I think you're referring to my message.
What I said was: If you convert an image that has been scanned in CMYK (and has
never been in RGB at all, ever), and you convert that image to RGB, you will
see a deteroration in the shadows--BUT NOT ON YOUR RGB MONITOR.
You will only be able to see the loss of shadow detail if you print a CONTRACT
PROOF, like a press operator would use to print the job on a printing press.
CMYK allows for greater shadow detail than RGB. Once you go from CMYK to RGB,
it's gone.
You'll never see this shadow detail on a computer screen, because a computer
screen is RGB. You can only see it on a proof, or on a printed piece.
Mike Russell
> >You'd be surprised - this file was chosen because someone else said "sure
> >it'll work with bright
> >saturated colors, but try one with shadows and you'll see an immediate
> deterioration."
>
> Actually, I think you're referring to my message.
>
> What I said was: If you convert an image that has been scanned in CMYK (and has
> never been in RGB at all, ever), and you convert that image to RGB, you will
> see a deteroration in the shadows--BUT NOT ON YOUR RGB MONITOR.
>
> You will only be able to see the loss of shadow detail if you print a CONTRACT
> PROOF, like a press operator would use to print the job on a printing press.
>
> CMYK allows for greater shadow detail than RGB. Once you go from CMYK to RGB,
> it's gone.
>
> You'll never see this shadow detail on a computer screen, because a computer
> screen is RGB. You can only see it on a proof, or on a printed piece.
Even a drum scanner scans in RGB, then uses it's internal tables or a profile to convert to CMYK,
just as PhotoShop does. Why should that conversion produce an image that's different from a
conversion that PhotoShop does?
I'm not asking this rhetorically, but I'm skeptical of an image whose shadow problems show up only
on print and not on the screen, and again I await an example.
BTW - I think the previous points about my example image are well taken, and will work on an example
with more saturated colors.
Mike Russell
RGB->CMYK conversion:
http://geigy.2y.net/DigPhoto/CMYK-RGB/CMYK-RGB.htm
Tacit
>to convert to CMYK,
>just as PhotoShop does.
Actually, that depends on the drum scanner. Sophisticated drum scanners don't
create an RGB file, then convert that file to CMYK; in fact, old-fashioned drum
scanners are not even digital devices at all, aren't connected to computers,
and don't create "files" per se.
>I'm not asking this rhetorically, but I'm skeptical of an image whose shadow
>problems show up only
>on print and not on the screen, and again I await an example.
You can create one. Make a new CMYK image. Fill the Cyan channel with a
gradient, or text, or whatever you like. Fill teh Black channel with solid
black.
Print a contract proof. Convert the CMYK image to RGB. Convert back to CMYK.
Look at the Cyan channel; it will be a solid tint of color, no gradient or text
or whatever. Prnt a contract proof and look at the difference.
On your RGB monitor, both files will look identical. On the proof, they will be
*worlds* apart.
Mike Russell
>
> You can create [an image whose shadow detail is different on the screen versus in print].
> Make a new CMYK image. Fill the Cyan channel with a
> gradient, or text, or whatever you like. Fill teh Black channel with solid
> black.
Yes, that's an excellent example of an artificially generated image that behaves as you say. But
I'm confining my "CMYK is OK" assertion to photographs.
> Print a contract proof. Convert the CMYK image to RGB. Convert back to CMYK.
> Look at the Cyan channel; it will be a solid tint of color, no gradient or text
> or whatever. Prnt a contract proof and look at the difference.
>
> On your RGB monitor, both files will look identical. On the proof, they will be
> *worlds* apart.
Yes, it's the difference between a pure black and a rich black, which is darker than 100% black
because it contains cyan. I think we'd both agree that there has never been a photograph that
duplicates this scenario.
And if the answer to the previous question is yes, does this re-enforce my statement that for the
great majority of photographs, the gamut clip going from RGB to CMYK and back again has no
appreciable effect?
> >Even a drum scanner scans in RGB, then uses it's internal tables or a profile
> >to convert to CMYK,
> >just as PhotoShop does.
>
> Actually, that depends on the drum scanner. Sophisticated drum scanners don't
> create an RGB file, then convert that file to CMYK; in fact, old-fashioned drum
> scanners are not even digital devices at all, aren't connected to computers,
> and don't create "files" per se.
Even in scanners with analog color correction circuitry, the RGB exists internally as a signal, and
that signal is processed in a way that PhotoShop can duplicate digitally. PhotoShop's
Image>Adjust>Selective color is a direct descendant of the 24 analog knobs you see in pictures of
the older scanners.
Verbal discussion is fine, and you bring up some interesting points, but I'm still waiting for a
photograph that has real problems with being converted to CMYK mode. Time will tell, but I'm
beginning to suspect that such images are few and far between, and that they may not exist at all.
Tacit
>darker than 100% black
>because it contains cyan. I think we'd both agree that there has never
>been a photograph that
>duplicates this scenario.
Download the following image:
http://www.xeromag.com/test/35539.TIF
(CAUTION: This is a 3.4MB CMYK TIFF image. You won't be able to view it online,
because it's CMYK.)
Proof it. Then, convert it to RGB and back to CMYK, and proof it again, or
examine the channels closely. Pay particular attention to the area surrounding
the panther's right eye, especially in the cyan and magenta channels of the
converted image.
The difference is quite subtle; nevertheless, a sufficiently critical observer
can spot the difference on the proofs. The black is slightly flatter and
slightly cooler in the converted image.
You can argue, of course, that the difference is so subtle that it makes no
difference; and for many people, I would agree. Only a very critical viewer
will notice.
Nevertheless, the difference DOES exist, and it IS noticeable.
Mike Russell
photographic elements, and CMYK colors in the yellows and dark shadows that would not easily be
generated from an RGB photograph.
Your image points out an interesting issue that could be an item for a future web page, that there
are benefits to being able to output directly to CMYK that are not there for RGB-based consumer
inkjet printers. Yes it would be nice if this were otherwise.
But that potential benefit does not detract CMYK as a "color safe" environment for photographers.
There have, IMHO, been too many warnings about staying in RGB, and I hope that encouraging people to
try out CMYK will open the door for more people to try color correcting an RGB photograph in CMYK.
My thanks to Tacit for contributing to the discussion - I hope he will say more. I have added some
discussion to my web page that covers some of the points he and his image have brought up.
The page is located at: http://geigy.2y.net/DigPhoto/CMYK-RGB/CMYK-RGB.htm . I hope others will
become curious and check it out.
--
http://www.zocalo.net/~mgr
"Tacit" <tac...@aol.com> wrote in message news:20010902123705...@mb-fo.aol.com...
Tacit
>it as a montage with
>photographic elements, and CMYK colors in the yellows and dark shadows that
>would not easily be
>generated from an RGB photograph.
The image is actually a CMYK drum scan of a painting. The scan was created in
CMYK on a Hell Chromagraph CP341--an analog drum scanner.
>But that potential benefit does not detract CMYK as a "color safe" environment
>for photographers.
>There have, IMHO, been too many warnings about staying in RGB, and I hope
>that encouraging people to
>try out CMYK will open the door for more people to try color correcting
>an RGB photograph in CMYK.
I agree. And, as I said before, I work almost entirely in CMYK.
But I think that choosing the correct color model for color correction requires
making an informed decision based on a strong knowledge of the limitations of
the various color models. Many people are aware that there are RGB colors which
don't exist in CMYK; fewer people are aware that tehre are CMYK colors which
don't exist in RGB.
Ont he whole, you can color correct a photographic RGB image in CMYK--and, as
you've pointed out, sometimes this gives better control over the shadow end
than correcting in RGB does.
But going the other way--starting with a CMYK drum scan that's destined for
CMYK output, and color correcting it in RGB--is something I try to avoid.
So, in short: Color correcting an RGB image in CMYK offers a powerful way to
control color very precisely. Color correcting a CMYK image in RGB, on the
other hand, offers no significant benefits and a few liabilities.
Mike Russell
> >Tacit, this is an interesting and well done picture, but I would describe
> >it as a montage with
> >photographic elements, and CMYK colors in the yellows and dark shadows that
> >would not easily be
> >generated from an RGB photograph.
>
> The image is actually a CMYK drum scan of a painting. The scan was created in
> CMYK on a Hell Chromagraph CP341--an analog drum scanner.
LOL. You got me then. It's an interesting image, and with your permission I may include in a set
of web pages I'm thinking about called "CMYK curiosities".
--
http://www.zocalo.net/~mgr
Tony W.
> Here's another question, though. Since you're making fine-art reproductions,
> have you considered going with a giclee (pronounced GEE-klay) process, using an
> Iris printer, rather than reproducing your work on an HP printer?
>
> The HP printer was really designed more for banners and trade-show displays
> than for fine-art reproduction. Many companies create giclee fine-art
> reproductions; these giclees, which use a similar process, are made on an Iris
> printer--it's kind of like an inkjet printer in some ways, but it was designed
> and built specifically for high-quality fine-art reproduction.
>
> A giclee typically costs around $90 to make, provided you supply the scan.
> Giclees can be printed on watercolor paper, canvas, and other materials. A
> giclee offers a much wider color range than an HP inkjet print, at much higher
> resolution, with superior fade resistance, UV resistance, and permanence.
>
> One company I've dealt with that does excellent fine-art giclees is called
> Gamma One Conversions. Their Web site is at
>
> http://www.gammaone1.com/
>
> As an artist, you will probably be much happier with a giclee than with an HP
> inkjet print. Giclee reproductions are suitable for presentation in galleries;
> they're that good.
Thanks, Tacit.
The HP-5000 can be used as a giclée printer.
Giclée is a French term meaning "a spray of ink". Giclée printers are any
high-resolution inkjet printers that provide significant lightfastness of the
output.
My own Epson Stylus Photo 1270 is a giclée printer. It has a high resolution that
makes prints look like photos, and when using Epson Matte Heavyweight paper, no
significant fading occurs for 24 to 26 years, under museum lighting conditions.
See this giclée Web site:
http://www.fineartGicleePrinters.org/pdfs_fine_art_giclee_printers/free_pdf_reports_wide_forma.htm
This site says of the HP Design Jet 5000, "The color gamut was impressive. Lots of
reds, blues, bright yellows, everything you need. Now (June 2001) the word from
the grapevine is that the UV inks have stability and longevity ratings over 150
years when displayed under museum-like conditions." That's similar lightfastness
to what the Irises can produce.
I got a proof printed on the HP 5000 with a paper distributed by Azon. I was told
that the Wilhelm Imaging Institute rated that paper (in combination with the HP
Design Jet 5000 UV inkset) to have a 100-year lightfastenss rating. (I didn't go
with that paper, though, because it's too thin.)
The article continues: "Since we are photographers and artists, we are very picky
about color gamut. When you hear the trials and tribulations of people getting
dull gray (instead of black) or lack of reds or blues on their Epson piezo systems
we know why we have the HP 5000 as our flagship printer."
I doubt your statement that Irises provide a much wider gamut and finer resolution
than the Hewlett-Packard Design Jet 5000.
There is company in my city (Vancouver) that specializes in giclée printing, but
the cost is more than double of cost at the company here that has the HP-5000.
I've tried a few other giclée printing companies over the past several years, but
their prices are all way too high for me to have much chance of selling the prints
and making a reasonable profit.