Dan Margulis Applied Color Theory - GCR and Neutral Colors

Date: Sun, 21 Jul 2002 21:36:14 -0600
From: Chris Murphy
Subject: Re: neutral CMYK

faro mojahedi writes:

>How can I achive full color neutrality in a CMYK file.
>The file that I have is slightly off in some areas.
>Using curves and anchor points puts a cast on some
>other areas.
>I appreciate any suggestions.

You need to know what makes gray on the output device in question. It will be different for every kind of output. If you don't know this, or kind find someone who does, finding neutrality will be a guessing game.

Chris Murphy
Color Remedies (tm)
Boulder, CO
303-415-9932

 


 

Date: Mon, 22 Jul 2002 15:53:11 -0700 (PDT)
From: Faro Mojahedi
Subject: neutral CMYK

Chris,
The job that I am working on is going to several dofferent places overseas. There is no information available about any of their set ups.

There was a convesation going on about CMYK profiles and CMYK values(numbers) on this list just recenyly that is somewhat related to what I am doing. I figure if I produce a file where all the values read neutral in Photoshop, then I should be better off. I usually start with Adobe RGB color (vs neutral) images. Using various tools, I convert them to grayscale. Then I drag that into my CMYK file(US Web Coated(SWOP) V2 . The numbers in the info window are indicating neutrals in most but not all areas. I have tried resetting black and white points, anchoring curves, etc. The latter fixes some areas and creates shifts in others. I figured US Web Coated(SWOP) V2 should a graybalanced space and give me a full nuertal range but that is not the case. The only thing that has worked for me so far is that I reconverted the file to a custon CMYK with max balck. That gives me neutral colors all over but there is only black channel info. Printing black with black channel only will not come out as rich as black made of all four channels. Any suggestions?

Thanks,
Faro

 


 

Date: Tue, 23 Jul 2002 17:06:09 +1000
From: Stephen Marsh
Subject: Re: neutral CMYK

> The job that I am working on is going to several
> dofferent places overseas. There is no information
> available about any of their set ups.

Then any CMYK that you supply may not be gray balanced for those conditions, unless you can get some general aimpoints. Many printers can hit general aimpoints if the seps you give them are close to what is required. For most places supply neutrals with a SWOP aimpoint for gray balance, For Euro and Japan use the general aimpoints supplied by Photoshop or similar profiles for those regions.

> I usually start with Adobe RGB color (vs neutral)
> images. Using various tools, I convert them to
> grayscale. Then I drag that into my CMYK file(US Web
> Coated(SWOP) V2 . The numbers in the info window are
> indicating neutrals in most but not all areas. I have
> tried resetting black and white points,
> anchoring curves, etc. The latter fixes some areas and
> creates shifts in others.

What info numbers? CMYK or LAB or RGB or what? The separation should be totally dead neutral if the input was neutral...

Editing a CMYK grayscale while in CMYK can give you problems if your are not careful with the gray balance. It would be best to not edit the file after conversion or perhaps only to stick with K edits unless you are confident, let the sep algorithm do the magic for you.

> I figured US Web Coated(SWOP) V2 should a
> grayballanced space and give me a full nuertal range
> but that is not the case.

CMYK is a device space and not a safe neutral editing space like some RGB flavours are (A98, sRGB, ColorMatch RGB etc).

Ideally you target the CMYK for the proofing process in question, which is mimics the press process in question...If you do not have this luxury then a separation such as SWOP or another target might be supplied with the hope that the printer can make the tones neutral with the different inks and conditions thay they use.

> The only thing that has worked for me so far is that I
> reconverted the file to a custon CMYK with max balck.
> That gives me neutral colors all over but there is
> only black channel info. Printing black with black
> channel only will not come out as rich as black made
> of all four channels.

As suggested earlier, a medium GCR type sep will put more tones into the K plate while still giving you some density in the undercolours.

You can manually create your own four colour separations so that there is a K plate with some manual placement of undercolours in the threequarter tones and shadows to add density but not in the lighter tones which only have K. I usually just go with a heavy GCR sep in custom CMYK and do not bother with other methods. I also factor in some extra dot gain in the separation than with a regular colour image, even more so for newsprint or uncoated.

Regards,

Stephen Marsh.

 


 

Date: Tue, 23 Jul 2002 17:41:19 +1000
From: Stephen Marsh
Subject: Re: neutral CMYK

> > As suggested earlier, a medium GCR type sep will put more tones into the K
> > plate while still giving you some density in the undercolours.

Correction - although I wrote heavy GCR in my final paragraph -- the above snip slipped through...I meant HEAVY GCR and not medium, for a four colour grayscale. This reduces neutral tone issues (but does not reduce them, the lighter tone often have little K weight), but the dotgain will become more critical if priting overseas where platemaking, operator skill and climate may vary from what you are used to.

Stephen Marsh.

 


 

Date: Tue, 23 Jul 2002 08:04:29 -0400
From: Dan Remaley
Subject: RE: Re: neutral CMYK - excellent!

> From: Stephen Marsh

>You can manually create your own four colour separations so that there is a
>K plate with some manual placement of undercolours in the threequarter tones
>and shadows to add density but not in the lighter tones which only have K. I
>usually just go with a heavy GCR sep in custom CMYK and do not bother with
>other methods. I also factor in some extra dot gain in the separation than
>with a regular colour image, even more so for newsprint or uncoated.

Stephen, I agree 100% - this is an excellent way to produce color seps. The problem is 'teaching' the printer to control the black. They are so used to 'ghosted' blacks and 'pushing' the density for text pages.

I have a 4-color printed handout that shows the use of GCR vs conventional that's used in my Process Control Seminar, here at GATF. If anyone would like a copy please send me your mailing address.

Dan Remaley/GATF
412.741.6860x450

 


 

Date: Tue, 23 Jul 2002 08:12:14 -0700 (PDT)
From: Faro Mojahedi
Subject: neutral CMYK

Stephen Marsh wrote:

> Then any CMYK that you supply may not be gray
> balanced for those conditions,
> unless you can get some general aimpoints..
> Many printers can hit general
> aimpoints if the seps you give them are close to
> what is required. For most
> places supply neutrals with a SWOP aimpoint for gray
> balance,

Then I should be okay since I supply the printer(whoever that might be overseas) with a SWOP [US Web Coated(SWOP) v2]. It has good black and white points and with most of the tonal range being neutral(Info window CMYK numbers).

> What info numbers? CMYK or LAB or RGB or what?

I look at CMYK numbers in Info window.

>The separation should be
> totally dead neutral if the input was neutral...

The input is dead neutral since it is a grayscale image that I drag into the above CMYK file.

> Editing a CMYK grayscale while in CMYK can give you
> problems if your are not
> careful with the gray balance.

That is what I am trying to avoid.

> CMYK is a device space and not a safe neutral
> editing space like some RGB
> flavours are (A98, sRGB, ColorMatch RGB etc).

I picked US Web Coated(SWOP) V2 as my CMYK working
space to work in a standard CMYK space. Just about everybody is familiar with that space.

> Ideally you target the CMYK for the proofing process
> in question, which is
> mimics the press process in question...If you do not
> have this luxury then a
> separation such as SWOP or another target might be
> supplied with the hope
> that the printer can make the tones neutral with the
> different inks and
> conditions thay they use.

What happens if I use MatchPrint CMYK space as my working space since that is an industry standard and that is how we proof our jobs? The problem is that I do not get neutral CMYK numbers in Info window in MatchPrint space either.

> As suggested earlier, a medium GCR type sep will put
> more tones into the K
> plate while still giving you some density in the
> undercolours.

I will try that.

> You can manually create your own four colour
> separations so that there is a
> K plate with some manual placement of undercolours
> in the threequarter tones
> and shadows to add density but not in the lighter
> tones which only have K. I
> usually just go with a heavy GCR sep in custom CMYK
> and do not bother with
> other methods. I also factor in some extra dot gain

How much more?

> in the separation than
> with a regular colour image, even more so for
> newsprint or uncoated.

Thank you very much for all the info.
Faro

 


 

Date: Tue, 23 Jul 2002 11:17:03 EDT
From: Dan Margulis
Subject: Re: neutral CMYK

Faro writes,

>>I usually start with Adobe RGB color (vs neutral) images. Using various tools, I convert them to grayscale. Then I drag that into my CMYK file(US Web Coated(SWOP) V2 . The numbers in the info window are indicating neutrals in most but not all areas. I have tried resetting black and white points, anchoring curves, etc. The latter fixes some areas and creates shifts in others. I figured US Web Coated(SWOP) V2 should a graybalanced space and give me a full nuertal range but that is not the case.>>

You don't know this. It *is* the case, in the mind of the profile. Whether it will be neutral on press is another story.

>>The only thing that has worked for me so far is that I reconverted the file to a custon CMYK with max balck. That gives me neutral colors all over but there is only black channel info. Printing black with black channel only will not come out as rich as black made of all four channels.>>

You don't need Maximum GCR. Just use Heavy. That will give a 4/c separation with enough black so that even if the CMY isn't quite neutral nobody will know--the black will cover it up.

Yet another reason why Photoshop needs to be able to edit supplied profiles--the profile you were using generates a rather light black, which is usually good but happens to be unsuitable for this type of work.

Dan Margulis

 


 

Date: Tue, 23 Jul 2002 19:28:48 -0500
From: Chris Brown
Subject: GCR

Request for advise from the pundits,

Photoshop allows three different algorithms of GCR computation in its CMYK
setup options, No press profiles are allowed.

My question is:

When is it advisable to use Light GCR?
When is it advisable to use Medium GCR?
When is it advisable to use Heavy GCR?

Assume U.S. SWOP inkset, a commercial printer (not publication printer)
using direct-to-plate technology, Grade 1 or 2 gloss paperj, 150 (minimum) line screen, heat-set web press.

Image content is variable. It varies within three categories:
1) People
2) Gray & black plastic outlined against paper base white
3) Colorful flowers with stems and leafs.

Thanks!

Chris Brown

 


 

Date: Wed, 24 Jul 2002 17:30:30 +1000
From: Stephen Marsh
Subject: Re: neutral CMYK - excellent!

Dan Remaley replies:

> Stephen, I agree 100% - this is an excellent way to produce
> color seps.

Hi Dan, nice to see that you still read the list - it is nice to know that there is someone of your background willing to freely share some ideas about the technical issues of the process of putting ink on paper.

I have to admit that I usually do not use heavier K generation such as med GCR in custom CMYK for colour work, which I know is your 'thing'.<g>

But it does depend on the image content and the size and placement of the image, if known. If I am supplied some separation aimpoints and these include a separation type and perhaps a statement that GCR and a certain percentage of GCR be applied - I will usually try to factor in these specs most of the time, unless I know (my opinion) the image really needs less/more K than the spec indicates.

As for my custom CMYK with rich shadows - this is not for the faint of heart and you still should attempt to keep the undercolours in balance to create a neutral for the intended process, even though it is targeted to the darker areas only (since in this case you are so concerned about neutrality that the majority of the image is K). This is perhaps better suited to non CMYK work for those wishing to have richer shadows and are feeling adventurous.

This is basically mixing a good grayscale monotone and applying this to the K plate of a CMYK file. It is also applied and manipulated to the CMY plates to form a neutral undercolour that does not exceed the TIL when combined with the K. Often this will be quite low when compared to a regular shadow which may have higher undercolour strengths.

> The problem is 'teaching' the printer to control the black.
> They are so used to 'ghosted' blacks and 'pushing' the density for text
> pages.

Now we get to one of the real issues of successful print - K generation choice and the rest of the production chain.

It is my feeling that a commercial flatsheet printer is a lot more flexible than a publication, when it comes to custom seps tailored for the image content. For advertising image seps for publications, it can be hard to know how placement and other pages on the signature or other variables may affect the advertisement repro. Will K or another ink gain more in my image due to other page or sheet content and can I factor this in and should I attempt to do so...all these and more go through my head as I separate or make post sep edits. It can seem more of a crap shoot than most print jobs when it comes to advertising content, so in these conditions I would probably stick to specs more so than not. But as in this threads case, there are times to break a suggested GCR/UCR K generation amount from the spec. A flatsheet printer who is only doing my job can afford to give the image/s their best attention, while it seems that not all advertising content is treated equal - as you suggest the editorial content such as text may affect things. Apart from the gray balance issues - this is where the lighter 'UCR type' K plate is appealing (most often light GCR in custom CMYK). Easier for CMYK correction and easier for the press to push text and shadows - although some do not agree that satisfactory colour moves are possible on the press, this is where I see 'your local print shop' and the different work they do as being more accommodating to lighter K than say a publication that contains both editorial and advertising content. I don't think that I am doing a good job of explaining my position here, hope this makes sense. This is not a statement of fact - just things as I see them.<g>

Can you comment on my other belief/feeling--that a heavy GCR type CMYK grayscale repro could benefit from some small extra dot gain being factored in over the regular gain for these conditions as in say a regular colour image (say 5% or similar for all plates or perhaps just more in the K)?

Again, thanks for your time and input to the thread.

Sincerely,

Stephen Marsh.

 


 

Date: Wed, 24 Jul 2002 18:20:07 +1000
From: Stephen Marsh
Subject: Re: GCR

Chris writes:

> Photoshop allows three different algorithms of GCR computation in its CMYK
> setup options, No press profiles are allowed.

Using the legacy custom CMYK interface that gives similar results to the old Photoshop separation tables - there are the none options (CMY) and maximum GCR as well. There is also the custom K generation curve which seems like an afterthought. You cant edit a 'proper' ICC profile in the custom CMYK (which is not really a proper full fledged profile with all the extra options that a proper ICC profile has).

> When is it advisable to use Light GCR?

When post separation or press moves need to be made with more ease.

As a general separation rule unless you think that medium makes a better default for your workflow and press conditions. If you make post separation edits.

> When is it advisable to use Medium GCR?

When you are more concerned with midtone neutrals. When you don't make post sep edits and do not expect the press to need any flexibility in the K.

> When is it advisable to use Heavy GCR?

When you really need to hold neutrality and don't trust the press to hold gray balance or when you get a really tough image that may need more K than CMY to form neutrals and to perform the role of the opposing ink/s.

> Assume U.S. SWOP inkset, a commercial printer (not publication printer)
> using direct-to-plate technology, Grade 1 or 2 gloss paper, 150 (minimum)
> line screen, heat-set web press.
>
> Image content is variable. It varies within three categories:
> 1) People

Light GCR, but it can vary depending on the photography and other conditions - for caucasian skintones which are not shadowed it is common to have a UCR style separation.

> 2) Gray & black plastic outlined against paper base white

Med GCR for lighter grays and Heavy for darker ones. Things could get deeper, if you need to apply the Dan Margulis 'black cat' principle then it may be very image specific choice on the sep method.

> 3) Colorful flowers with stems and leafs.

UCR or light GCR may add some transition K tone density that UCR lacks once curved. Although in the greens the K generated by a medium GCR may add some more snap, so you might channel mix or selective colour some manual heavier GCR in the greens if using a lighter K.

There are no hard and fast rules - for any of the above situations images and conditions can be found which break the 'rules'.

I have a full colour image that could perhaps use heavy GCR in skintones, but this was a cyan colour cast situation and not a regular colour correction (art director wanted the cast preserved) - this is just an example that it's never a simple matter of just picking a GCR amount from a canned list to match a rough image category.

If I am ever in doubt - medium GCR seems a safe bet, when going lighter or darker may lead to just as many problems as using medium...I would take the middle road approach.

Sincerely,

Stephen Marsh.

 


 

Date: Wed, 24 Jul 2002 10:36:09 -0400
From: Floyd Rominski
Subject: neutral CMYK

>From: Faro Mojahedi

> reconverted the file to a custon CMYK with max
> balck.
> That gives me neutral colors all over but there is
> only black channel info. Printing black with black
> channel only will not come out as rich as black
> made of all four channels.

--------------------------------------------------

Faro

If i'm correctly understanding this thread you want a 4c b/w.
i posted a method for this May 8, as follows. . .

a usable Sep Set Up is as follows:

CMYK SetUp:
Dot Gain: Curves, enter curves ala Dan M

Sep Options: GCR

Blk Generation: Custom: 100% > 91%
90% > 78%
50% > 45%
Put Tot'l Ink limit where you need it

A much better method is as follows.
This will give you identical Grayscale opacities as your b/w image,
5% b/w will= 5% 4c > 100% b/w will= 100% 4c

This has been successful for me-

Having your b/w document document open,
duplicate it 3 times for 4 identical copies, then in Channel Palette Options, select merge channels and create a cmyk file. . .(thanks Lee)

This gives a very dark muddy image, not to worry.
You then open a 'Curve AdjustmntLayer' and load curves i can send you if you
want to explore beyond the first suggested Sep SetUp above.

i've done a few of these curves, to give neutral, cool, and warm b/w 4c's

regards,
Floyd Rominski

 


 

Date: Thu, 25 Jul 2002 10:57:53 -0400
From: Dan Remaley
Subject: RE: Re: neutral CMYK - excellent!

Stephen writes:

> Can you comment on my other belief/feeling - that a heavy GCR type CMYK
> grayscale repro could benefit from some small extra dot gain being factored
> in over the regular gain for these conditions as in say a regular colour
> image (say 5% or similar for all plates or perhaps just more in the K)?

There are many positive things about GCR and a few negative ones. GCR is mainly for neutral and gray image issues (the worst, a 4/c Black & white) because gray is the first color to change on press, also the hue shift of gray impacts all the colors (gray balance in scanning). So, the belief that you can 'color correct' at press isn't quite accurate. When you 'push' any of the colors at press, your improving 'that color' in 'that image', but globally your moving other colors in other images. The use of GCR 'limits' how far the press can 'move' the color. Now, if I had a good proof (measurable) and liked the color - I don't want the press to 'change' the color and with high GCR they can't! The second advantage of GCR is that the color stays 'purer'. Remember how grays, purples, browns change at press? When GCR is applied to a Red Apple, for example, the gray component is Cyan. If I use a high amount of GCR then the Cyan is replaced by Black, there's no hue shift with Black in could become lighter or darker but no hue shift as Cyan would make if not printed correctly. One study, done in Europe, they scanned images and proofed them, some of the colors were a Delta differences of 100 or more. When scanned again with the use of GCR these differences were reduced to 50 in the same colors. The negative things - you can't change the color on press! Or is that a 'good thing'??? On the printed sample I have offered to the group, there is a conventional scan (4/color as a black & white) with a +6% magenta 'push'. Next to it is a GCR scan with the same +6% magenta 'push', the GCR scan has little change! The other negative, is if the black is run too heavy the color will be dark and lifeless. I agree in taking some weight out of the black to help with GCR printing. The two 'secrets' of color printing are gray balance at press and GCR!

Dan Remaley/GATF
412.741.6860x450

 


 

Date: Thu, 25 Jul 2002 20:28:25 -0500
From: David Riecks
Subject: Re: Re: GCR

Stephen Marsh wrote:

>> When is it advisable to use Heavy GCR?
>
>When you really need to hold neutrality and don't trust the press to hold
>gray balance or when you get a really tough image that may need more K than
>CMY to form neutrals and to perform the role of the opposing ink/s.

Stephen:

I've got an upcoming job where I need to print some historical images in a four color booklet. Some of the scans are RGB, and some are grayscale. I would like to introduce a slight color cast to each (not the same for each) to "simulate" old print conditions (sepia, iron toner, albumen, etc) so what I'm wanting primarily is a heavy black plate with some tones imparted by the CMY.

If I'm primarily working in RGB, and going to CMYK at the end, and we assume similar conditions, 150 linescreen, sheetfed commercial press, white coated paper, would you suggest going with Heavy or Medium GCR?

Also, would you suggest going to CMYK directly from RGB and then imparting the colors, or sticking to RGB to do that work?

David

David Riecks
http://www.riecks.com, Chicago Midwest ASMP member

 


 

Date: Fri, 26 Jul 2002 18:21:49 +1000
From: Stephen Marsh
Subject: Re: neutral CMYK - excellent!

Dan Remaley writes:

> On the high end scanners we could add GCR as a percentage of black
> replacement (i.e. 70%). Photoshop is not as good, even it's 'med' setting
> does little in the quarter tone area, compared to a high end scanner.

Very true. A few months ago list member Lee Varis wrote with similar concerns and I mentioned that I have used a layer with a true greyscale version in the K channel and layer option blend if sliders/opacity to add some manual GCR into the highlight/quartertones to beef up the greater CMY component in this area (the undercolours may need slight reduction so that luminosity does not change). There are other ways too - this was just the inital method I came up with at the time to get the job done and I have not really explored things since...

> There are many positive things about GCR and a few negative ones.
> GCR is mainly for neutral and gray image issues (the worst, a 4/c Black &
> white) because gray is the first color to change on press, also the hue
> shift of gray impacts all the colors (gray balance in scanning). So, the
> belief that you can 'color correct' at press isn't quite accurate. When you
> 'push' any of the colors at press, your improving 'that color' in 'that
> image', but globally your moving other colors in other images.
> The use of GCR 'limits' how far the press can 'move' the color. Now, if I
> had a good proof (measurable) and liked the color - I don't want the press
> to 'change' the color and with high GCR they can't!

Good points.

> The second advantage of
> GCR is that the color stays 'purer'. Remember how grays, purples, browns
> change at press? When GCR is applied to a Red Apple, for example, the gray
> component is Cyan. If I use a high amount of GCR then the Cyan is replaced
> by Black, there's no hue shift with Black in could become lighter or darker
> but no hue shift as Cyan would make if not printed correctly.

Again, I agree - the black plays a much greater role and this is where there seems to be a lot of variation.

I have a good real life example of this which I might post to the list as it makes a very good lesson for a number of reasons. Although different subject matter - the issue was overinking of cyan and how a VERY heavy GCR would have saved the day (although that does not excuse the printer, as many pages matched the proof perfectly - so someone fell asleep half way through the run).

> The negative things - you can't change the color on press! Or is > that a 'good thing'???

Depends on the job and the separation/print conditions. Due to many printers not providing decent separation aimpoints for customers - they are given many separations which are not ideal for their conditions. Although to be fair to some of the printers, since prepress is becoming a lost art and a dead-end business, people who would prefer not to make separations are making them - this is why separation info is important to both the printer and their customers. The upshot of this is that without the ability to 'change' colour on press to a limited extent, the less than ideal seps may not do the job by themself.

Again, a small flatsheet printer who is printing a custom job can sometimes alter things on press, for the better...while in a publication doing this may work for one image but throw off many others. I think it depends on your print setting as to what can and can't work on press, but there are limits and this is obviously highly dependent on the page and sheet content.

For the good part of a year I worked for such a small printer - who gave little upfront concern to prepress unless there were major problems on press. CTP can get a corrected plate out in 10 minutes and this seemed to be a 'viable' option for this printers less than ideal press input workflow.

> The other negative, is if the black is run too heavy the color will
> be dark and lifeless.

And this seems to be a wide issue. Photoshop's old custom CMYK defaulted to a medium GCR and dot gain which was probably a bit too low for many users conditions with a high K limit. Many canned profiles seem to offer heavier black plates (say the Adobe v2 SWOP) than say a 'traditional' light skeleton UCR style K - but ICC profiles only offer 1 dot gain per profile which may not be ideal for all users. It depends on the image content if this becomes an issue or not .

> I agree in taking some weight out of the black to help with GCR printing.

Thanks, upon reflection I probably forgot to apply the Dan Margulis 'black cat*' principle and have lesser weight undercolours than for a full colour image in the same conditions when separating a four colour greyscale with heavy GCR, in the example that I remember coming out muddier than intended.

> The two 'secrets' of color printing are gray balance at press and GCR!

That still covers a lot of ground!<g> Add in dot gain and it get's deeper...

Thank you Dan, your thoughts, time and offers of literature are greatly appreciated.

Regards,

Stephen Marsh.

 


 

Date: Fri, 26 Jul 2002 18:45:43 +1000
From: Stephen Marsh
Subject: Re: Re: GCR

> I've got an upcoming job where I need to print some historical images in a
> four color booklet. Some of the scans are RGB, and some are grayscale. I
> would like to introduce a slight color cast to each (not the same for each)
> to "simulate" old print conditions (sepia, iron toner, albumen, etc) so what
> I'm wanting primarily is a heavy black plate with some tones imparted by the
> CMY.

When a subtle cast is required - it is wise to get all parties on the 'same page' understanding that casts are actually required and that this is not bad seps or proofing.

> If I'm primarily working in RGB, and going to CMYK at the end, and we assume
> similar conditions, 150 linescreen, sheetfed commercial press, white coated
> paper, would you suggest going with Heavy or Medium GCR?

For a number of reasons, I would go with Heavy GCR (the #1 reason being the K weight vs undercolour weight).

> Also, would you suggest going to CMYK directly from RGB and then imparting
> the colors, or sticking to RGB to do that work?

David, I personally would do it in RGB and just leave everything to the sep algorithm for consistency...get each grayscale right with endpoints and other key tones, impart a cast that does not impact on tone and then batch to CMYK while you sit back and have a break. This is not to say that doing things in CMYK is not viable, and the RGB workflow does not answer the K channel which may need some custom curving for each image after separation (avoid CMY edits in this situation if you are not sure of things).

If you are not reproducing strong corrected images but only the faded historical images - then the custom created sep that I mentioned in an earlier post may be better as the luminosity is entirely made of K and does not have a CMY component as in heavy GCR in custom CMYK in Photoshop.

Hope this helps,

Stephen Marsh.

 


 

Date: Fri, 26 Jul 2002 12:40:31 -0700
From: Pete Masterson
Subject: Re: GCR

David Riecks wrote:

>I've got an upcoming job where I need to print some historical images in a
>four color booklet. Some of the scans are RGB, and some are grayscale. I
>would like to introduce a slight color cast to each (not the same for each)
>to "simulate" old print conditions (sepia, iron toner, albumen, etc) so what
>I'm wanting primarily is a heavy black plate with some tones imparted by the
>CMY.

I'd suggest converting the image to LAB color -- it will hold the black relatively unchanged while you can fiddle with the overall color effect as you wish. I discovered this trick after reading one of Dan's articles in Electronic Publishing -- I have a client who has a series of art and architecture travel books where a black and white photo is modified to have a very heavy color cast to match other colors used in the design. When "fiddling" with the color in RGB and/or CMYK I'd lose much of the contrast in the base photo. Using LAB protected the blacks and let me modify the color to any extreme. Once I was satisfied with the look, then I'd convert the image to CMYK for printing.
--
Pete Masterson

================================================================
Aeonix Publishing Group - Publishing Production and Consulting
Cover and Book Design, Layout, and Production
Voice & FAX: (510) 222-6743 -- 9AM to 5PM M-F Pacific Time
Visit my web site at http://www.aeonix.com for publishing hints
P.O. Box 20985, El Sobrante, CA 94820-0985
================================================================

 


 

Date: Fri, 26 Jul 2002 23:28:32 +1000
From: Stephen Marsh
Subject: Re: Re: GCR

Hi again David - my suggestion of Heavy GCR obviously presumes a more neutral shadow in your casted images. If the historical process you are reproducing did not have a neutral shadow then lower GCR may be required so that the darker tones have a chance of remaining casted. It would probably depend very much on the image and cast colour and strength in question on the K generation used. Medium may be a safer bet for the cast over the neutral in the long run.

Also, depending on the subtle casts being reproduced, you will probably want to tweak the caste once in CMYK - I usually prefer wider spaces to introduce or remove casts and then CMYK to finally hone things. As noted earlier, you may be making K curve tweaks so each image is probably going to need some finesse anyway.

Stephen Marsh.

 


 

Date: Sat, 27 Jul 2002 09:53:59 +1000
From: Stephen Marsh
Subject: Re: LAB (was GCR)

Pete Masterson writes:

> I'd suggest converting the image to LAB color -- it will hold the
> black relatively unchanged while you can fiddle with the overall
> color effect as you wish. I discovered this trick after reading one
> of Dan's articles in Electronic Publishing -- I have a client who has
> a series of art and architecture travel books where a black and white
> photo is modified to have a very heavy color cast to match other
> colors used in the design. When "fiddling" with the color in RGB
> and/or CMYK I'd lose much of the contrast in the base photo. Using
> LAB protected the blacks and let me modify the color to any extreme.

Pete, you might want to try some additions to your LAB workflow...

i) Try simulating LAB edits while in RGB or CMYK by using the fade command, or setting your channel mixer or curve adjustment layer or paint tool to COLOR blend and not normal for colour work - or LUMINOSITY if you are sharpening or altering tone but not colour. Thinking in LAB while working in RGB can often get the desired results without the minor penalties of the mode change to LAB (quantization errors, loss of levels - which may or may not be visible or an issue to you). LAB is obviously great for radical changes and it is really easy to push colours out of gamut for many devices - so I find the color blend edits in RGB offer a bit more finesse than in LAB, particulary so for curves.

I think the 'ultimate' application of Dan's curving techniques is when you can enhance both colour and luminosity in each separate channel while in normal blend mode for your curve adjustment layer. Sometimes I have problems reaching this goal - and I have to cheat by making two curve adjustment layers, with one being set to color and the other luminosity blend mode. Although not as good as the one curve adjustment in normal mode, separating the move into two edits and making LAB/HSB type edits while in RGB or CMYK often works well for me when regular methods do not. Some scanners offer this approach with LCH edit controls or LS curves or whatever.

https://www.ledet.com/margulis/ACT_postings/ACT-LABLoss.html

http://ep.pennnet.com/Articles/Article_Display.cfm?Section=Archives&Subsection=Display&ARTICLE_ID=76110&KEYWORD=dan%20margulis

(free rego req'd) - Dan's early Make Ready article on separating colour and contrast which seemed to evolve by his current PP6 book and is quite different in the curve approach, unless I am missing something (ch3).

ii) If you do need to go to LAB because other modes do not offer the required results - then perhaps do this in a dupe of the current file. If making changes to the AB then do so and then blend this over your original data in COLOR blend mode so that the original files luminosity values are preserved. Most luminosity based tonal or USM edits seem to easily simulated in RGB or CMYK with a LUMINOSITY blend - so I personally do not go to LAB just for L channel edits. The color blend or the layer options to only include CMY data are also options for CMYK work.

Retouching in color or luminosity blend modes while in RGB or CMYK can have benefits too, although not for the more extreme retouching where truly separating colour from tone can help as in LAB.

I think part of the issue with Photoshop's LAB mode is that it uses the ICC LAB model and not the CIE LAB model, which means that for the wider gamut RGB spaces (Adobe RGB and wider) you can loose unique colours in a trip to Photoshop's LAB mode. Perhaps if APS used CIE LAB this would not be an issue. There is also the bit depth issue, which does affect the data but may not make any visual difference to results. What happens to the data and what produces a pleasing image are often very different things, but it can be wise to know the issues with LAB edits as it is not a free ride.

LAB is great - but the more I experiment the less I seem to find true uses for it. This is probably becuase most of my edits are usually small moves to finesse a close existing colour, and not radical departures which LAB seems to better suited.

Sincerely,

Stephen Marsh.

 


 

Date: Sat, 27 Jul 2002 10:23:48 EDT
From: Dan Margulis
Subject: Re: Re: LAB (was GCR)

Stephen writes:

>I think part of the issue with Photoshop's LAB mode is that it uses the ICC
>LAB model and not the CIE LAB model, which means that for the wider gamut
>RGB spaces (Adobe RGB and wider) you can loose unique colours in a trip to
>Photoshop's LAB mode. Perhaps if APS used CIE LAB this would not be an
>issue.

First, if one had an RGB file that contained such psychedelic colors, there would be no reason to enter LAB with it, since the usual purpose of LAB is to richen the colors.

Second, if such a psychedelic color existed, it couldn't be displayed on the monitor or output accurately in any conventional way.

Third, if even if such a psychedelic color existed and could be output, small variations in A and B at such an eyepopping saturation would be invisible to the human observer. They would only be visible if they were also found in the L, which is unaffected by this purported problem.

Moving to a wider-gamut LAB would be a big mistake--it has no upside, and it would make correcting in LAB even more volatile than it currently is.

Dan Margulis

 


 

Date: Mon, 29 Jul 2002 19:16:48 +1000
From: Stephen Marsh
Subject: Re: LAB (was GCR)

Dan replies:

> Third, if even if such a psychedelic color existed and could be output, small
> variations in A and B at such an eyepopping saturation would be invisible to
> the human observer. They would only be visible if they were also found in the
> L, which is unaffected by this purported problem.

Thanks for the points from the real world Dan.

Since I usually only go to LAB for unusual AB edits and then reblend them in color mode into the original file I avoid edits to the L component. I simply use luminosity blends for any L work while in RGB or CMYK when colour correcting.

> Moving to a wider-gamut LAB would be a big mistake--it has no upside, and it
> would make correcting in LAB even more volatile than it currently is.

I can only imagine how much harder this would be -- as things currently stand there can only be so much movement to a curve before gamut issues come into play, having finer increments would only make things harder with the current tools. I was simply thinking along the lines of mode transforms, more so than the actual edits while in the LAB space - I am guessing that the PCS LAB in ICC profiles is CIE based and that is why CMYK > CMYK is said to be better than CMYK > LAB > CMYK.

This is related to some CMYK > CMYK issues that I am trying to sort out in my head.

I am stumbling to see what the difference is - since ICC profiles either go to XYZ or LAB behind the scenes...and many profiles use LAB as the PCS.

So if using the old Photoshop versions before ICC profiles or even with ICC profiles in later versions - CMYK > LAB > CMYK, does it use ICC LAB?

If using C2P or even better - P2P, then the workflow is CMYK > PCS (xyz or lab) > CMYK, if LAB is the PCS does it use CIE LAB?

After some recent discussions on other lists and reading the ICC primer PDF's from Apple and GretagMacbeth I am still confused (yes, shocking as it may seem - this profile stuff is often confusing).

Regards,

Stephen Marsh.

 


 

Date: Mon, 29 Jul 2002 19:26:15 +1000
From: Stephen Marsh
Subject: Re: LAB - Correction

>>If using C2P or even better - P2P, then the workflow is CMYK > PCS (xyz
or lab) > CMYK, if LAB is the PCS does it use CIE LAB?>>

Sorry - that was obviously a mistake. Convert to profile is much better than the v5.x profile to profile transform. I presume that both methods are the same in regards to the PCS. Does the same hold true for both v5.x and later versions which fully embraced ICC profiles?

Stephen Marsh.

 


Adobe Photoshop training classes are taught in the US by Sterling Ledet & Associates, Inc.