Applied Color Theory - Ink Limits

From: James Steincamp, INTERNET:tomomi@erols.com
Date: Mon, Oct 25, 1999, 7:55 AM
RE: Ink Limits...

Thanks to everyone for providing enough information about RGB set-up and Gamut to keep me preoccupied for a long while.

On a different track, i have a CMYK file that i have questions about. It's another promotional card for my photographer friend (this time he got a drum scan done as opposed to using the Kodak slide scanner he has access to at work). It will be printed on Kromekote at one of those places that gangs up every job on the same press run... i imagine that it's a web press.

The photographer wants to include the black frame around the image from the original chrome (you know the part that says Kodak EPP 6005 with the frame number as has been all the rage in photography for the last decade). It registers as the darkest area in the image, weighing in at: 60c 49m 53y 94k, though some shadow areas in the "exposure" area also register the same.

My instinct is to fiddle with the numbers as the total ink value is only 256 and i should be able to hold 300 on a web/coated press, right? -- but i'm pretty new at all this so i don't know if i should be trusting my instinct. Doesn't 94 in the black channel seem a bit heavy? Did the scanner operator know something that i don't?.....

The other half of the question is that he wants to include a drop shadow in the image and as i see so many horribly done drop shadows i wanted to know how to do one right (not the photoshop technique), should it be generated from black ink only? Why is it that i see so many that have a halo effect... where the image tapers off nicely and then, all of a sudden, the halftone dots stop just short of the last few percentage points of black. Sorta hard to describe verbally. I see this much more often on uncoated papers, but i've seen it on grade one coated papers as well. Is there a technique to reduce the risk of this?

thanks.


From: Chris Murphy, INTERNET:lists@colorremedies.com
Date: Mon, Oct 25, 1999, 12:02 PM

James Steincamp writes:

>It registers as the darkest area in the image, weighing
>in at: 60c 49m 53y 94k, though some shadow areas in the "exposure"
>area also register the same.

Sounds like UCA to make sure the frame shows up black enough; the scanner operator may realize that the black used on this particular press isn't that black (or black enough).

>My instinct is to fiddle with the numbers as the total ink value is
>only 256 and i should be able to hold 300 on a web/coated press,
>right? -- but i'm pretty new at all this so i don't know if i should
>be trusting my instinct. Doesn't 94 in the black channel seem a bit
>heavy?

Not if the frame is black. I'd also not be too concerned with heavy black generation (if that's what this is the result of) because the scanner operator may be doing this intentionally. If he's familiar with this printing process, he might be thinking about how more black generation will reduce the visibility of slight misregistration. If misregistration is a concern, then increasing black generation will sacrifice shadow area detail in exchange for reducing the visibility of the misregistration.

300% TAC (total area coverage, sometimes called total ink limit) on a web/coated press applies specifically to SWOP. Specific inks, specific paper, specific dot gain, etc. If this process isn't exactly like SWOP, then the 300% TAC doesn't necessarily apply. To determine total ink limit for a printing process that doesn't have an applicable printing spec, you can print a total ink limit target to compute what the total ink ought to be; some of those targets also help you figure out the black ink limit.

>Did the scanner operator know something that i don't?

Ask him. If he knows the process he's scanning to, then all of the above may be for a reason. If he's never scanned for this particular kind of output, who knows.

>The other half of the question is that he wants to include a drop
>shadow in the image and as i see so many horribly done drop shadows i
>wanted to know how to do one right (not the photoshop technique),
>should it be generated from black ink only?

Depends on what's in the image. If you've got an image that contains something gray (like a tin can, or silver bullet - whatever), if you use a black only drop shadow then you risk the possibility that on press the oscillations on press will cause hue shifts to occur on your image, but not the drop shadow; assuming the image is also not 100% black. So in this case it's better to make a drop shadow in a similar method to the neutral areas in the image; that way if a hue shift occurs in press, most people won't notice it that much because the drop shadow will have the slight hue shift also; so it won't look out of place and tip the view off that the shadow is gray, but the image content has a cast to it.

Otherwise a black only drop shadow, I think, looks better. You don't see little magenta and cyan dots in the highlight areas; sometimes you can also get rosettes in the drop shadow unless it's 100% black.

Making a good drop shadow in Photoshop isn't that difficult if you start out with an RGB image, copy the shadow to a new document, change CMYK Setup for Maximum black generation, then convert to CMYK, now copy it back as a layer behind your image and flatten.

The filters in Photoshop are stupid when it comes to CMYK. They all think that gray is C=M=Y so you end up with muddy, non-gray balanced results when using any filter. The assumption is correct if you're working in RGB - gray is always equal component RGB.


 

From: Dan Margulis, INTERNET:76270.1033@compuserve.com
Date: Mon, Oct 25, 1999, 1:38 PM

James writes,

<<The photographer wants to include the black frame around the image from the original chrome (you know the part that says Kodak EPP 6005 with the frame number as has been all the rage in photography for the last decade). It registers as the darkest area in the image, weighing in at: 60c 49m 53y 94k, though some shadow areas in the "exposure" area also register the same.

<<My instinct is to fiddle with the numbers as the total ink value is only 256 and i should be able to hold 300 on a web/coated press, right? -- but i'm pretty new at all this so i don't know if i should be trusting my instinct. Doesn't 94 in the black channel seem a bit heavy?>>

For an average image, yes, but you have to ask yourself *why* it's high. Black values that high are discouraged because you can't be sure that the dots won't close up and destroy detail.

Since, in the sprocket area of the film, there *is* no detail, why should you care whether you lose it or not? So, select the area and jack black all the way up to 100%. Forget that you're under the total ink limit, this will still give you a very dark black.

<<The other half of the question is that he wants to include a drop shadow in the image and as i see so many horribly done drop shadows i wanted to know how to do one right (not the photoshop technique), should it be generated from black ink only?>>

No. Doing it in all black often results in overly obvious edges, especially if it's surprinted over a light area of the image that wouldn't normally contain black. Also, going all black is dangerous in that black is not only the strongest ink, it's also the one most likely to vary in density on press. So, not only is a press error more likely, but it's more damagingn when it happens.

The reason that people erroneously use all black is that they are worried that if they create the shadow as they would a normal light gray, with CMY only, there may be a color shift if something unexpected happens on press. This is a very valid concern.

Addressing it by making the shadow *entirely* black, however, is swatting a fly with a sledgehammer. Black ink is so potent that even if you use half that much, there's no way that the shadow can take on an off color. And if you do, the whole look of the shadow will be smoother, the edges less pronounced, and much less chance of a press error making the whole thing too dark.

You can't handle it like a normal separation, either. Making the shadow with Heavy GCR is ideal--the black value will be about as high as the magenta and yellow. This will make a color shift impossible without any of the drawbacks of going all black. It sounds like your picture is printing pretty large, and if so, there's another consideration. If that sprocket area you mentioned is more than about an inch wide, it's an absolute lock that black ink will come down too heavily in the interior of the image. So you should probably even reduce the black component of the shadow because you know it will pick up on press.

The very good point is made by Steve that you should be adding noise, which is the enemy of banding and harsh edges, to the shadow.

And the very sophisticated point is made by Chris that if there is some other prominent light neutral object in the picture, you're better off making the shadow CMY only, and if color shifts on press that's how it goes. This eminently correct advice would not occur to about 99.9% of imaging professionals.


From: Chris Murphy, INTERNET:lists@colorremedies.com
Date: Mon, Oct 25, 1999, 9:43 PM

>And the very sophisticated point is made by Chris that if there is some
>other prominent light neutral object in the picture, you're better off
>making the shadow CMY only, and if color shifts on press that's how it
>goes. This eminently correct advice would not occur to about 99.9% of
>imaging professionals.

I don't know about that. If your gray has only a relatively small amount of black in it, it's possible for gray balance on that press to be different from what you're expecting. I see presses all over (I'm not claiming this is the right way, I'm saying it's reality) that have some hue cast to a square patch of 50c,39m,39y. So perhaps the last sentence ought to be revised to say, "The eminently correct advice would not occur to about 99.9% of imaging professional, so long as the press is gray balanced (i.e., proper process control procedures are being applied.)"


From: James Steincamp, INTERNET:tomomi@erols.com
Date: Mon, Oct 25, 1999, 11:23 PM

Thanks again for the advice...

>Since, in the sprocket area of the film, there *is* no detail, why should
>you care whether you lose it or not? So, select the area and jack black all
>the way up to 100%. Forget that you're under the total ink limit, this will
>still give you a very dark black.

Should i reign in the black plate for the useful image area to a more conservative number?

>The reason that people erroneously use all black is that they are worried
>that if they create the shadow as they would a normal light gray, with CMY
>only, there may be a color shift if something unexpected happens on press.
>This is a very valid concern.

And most of the "how to" books bear state that reason for using black only.

>Addressing it by making the shadow *entirely* black, however, is swatting a
>fly with a sledgehammer. [...]

>You can't handle it like a normal separation, either. Making the shadow
>with Heavy GCR is ideal--the black value will be about as high as the
>magenta and yellow. This will make a color shift impossible without any of
>the drawbacks of going all black.


As the file is already a CMYK file is it possible, or desirable to re-separate to use Heavy GCR? I'd ask the drum scan operator what settings he used, but the scan was done over a month ago...

>It sounds like your picture is printing pretty large, and if so, there's
>another consideration. If that sprocket area you mentioned is more than
>about an inch wide, it's an absolute lock that black ink will come down too
>heavily in the interior of the image. So you should probably even reduce
>the black component of the shadow because you know it will pick up on
>press.

The sprocket frame is roughly 1/4" around the 6x6 frame so it sounds as if i'm in the clear on that matter. I suppose that should i need a solid black background for an image that in the future it might be better to add a second black as the fifth color to avoid the issue altogether.

>The very good point is made by Steve that you should be adding noise, which
>is the enemy of banding and harsh edges, to the shadow.

There's another technique that most of the "how to" books overlook.

>And the very sophisticated point is made by Chris that if there is some
>other prominent light neutral object in the picture, you're better off
>making the shadow CMY only, and if color shifts on press that's how it
>goes. This eminently correct advice would not occur to about 99.9% of
>imaging professionals.

Being a designer i have always relied on having someone else prepare my scans and deal with the color management issues. But i've also come to the point where some of my clients are paying enough to deserve top-notch color and printing... and those are the jobs that make my end of the work most rewarding and profitable.

Poking around on the net, this list and reading a few books on the subject have given me a much better understanding of what goes on after i hand my files off to the service bureau/printer. Prepress folks seem a lot more like artisans who've learnt the rules of the trade by practise rather than scientists running around in white coats (as some printers would have us believe).

It's not an entirely reassuring picture since photoshop has clearlyput some of that responsibility in my lap as well... ;)


 

From: INTERNET:Jsweengatf@aol.com, INTERNET:Jsweengatf@aol.com
Date: Tue, Oct 26, 1999, 9:58 AM

In a message dated 10/26/1999 5:48:12 AM Eastern Daylight Time, rbean@execpc.com writes:

> Could you explain the difference between a colorimeter and a
> spectrophotometer (ie, how do they work and what are they
> actually measuring)?

SPECTRO measures reflectance all accross the visual spectrum -- typically at 10nm intervals - 400-700 approx. ROYGBIV, the rainbow....

COLORIMETER by definition, like a densitometer, also measure reflectance, but at THREE point, R,G, B. (Tri-stimulus)Status-T and Status-E are densitometer response. (also Status M and A forPhoto)

CIE X,Y,Z are the basic colorimter response; from that we calculate L*a*b*,etc.

Lots of good reading on this, published by X-Rite, Gretagmacbeth, GATF, etc.

NOTE: A Spectro IS a densitometer and a colorimeter, with the right math, thus the popular term "Spectrodensitometer" used in the Graphic Arts for many instuments -- X-Rite 500 series comes to mind, and my org's on-web Spectro, ColorQuick, which measures a color bar on the fly.


From: "Ron Bean", INTERNET:rbean@execpc.com
Date: Tue, Oct 26, 1999, 5:48 AM

 

Chris Murphy <lists@colorremedies.com> writes:

>I don't know about that. If your gray has only a relatively small amount
>of black in it, it's possible for gray balance on that press to be
>different from what you're expecting. I see presses all over (I'm not
>claiming this is the right way, I'm saying it's reality) that have some
>hue cast to a square patch of 50c,39m,39y.

How do you know it's the press and not the ink? If it's a perfecting press, do you see the same effect on both sides of the press sheet? What happens if you run the inks in a different order on the same press?

As long as I'm asking questions... Could you explain the difference between a colorimeter and a spectrophotometer (ie, how do they work and what are they actually measuring)?


 

From: INTERNET:jglazer@after-image.com, INTERNET:jglazer@after-image.com
Date: Tue, Oct 26, 1999, 11:09 AM

Another reason not to use process shadows is if the job is being run at a lower lpi (<133), where the rosette will be too strong. The dots of a K only shadow often are more subtle than its process equiv. Of course, this does not apply to higher line screens.

>The reason that people erroneously use all black is that they are worried
>that if they create the shadow as they would a normal light gray, with CMY
>only, there may be a color shift if something unexpected happens on press.
>This is a very valid concern.


From: Dan Margulis, INTERNET:76270.1033@compuserve.com
Date: Tue, Oct 26, 1999, 2:56 PM

James writes:

<<Should i reign in the black plate for the useful image area to a more conservative number?>>

Unless you have very critical detail in the shadows and are fairly proficient in color correction, no. The scanner operator gave you a balanced, dark shadow. That's what he's supposed to do. It's higher in K and lower in CMY than I recommend for the majority of images. However, it's usually a workable approach, and now that he's done it that way I wouldn't change it unless the circummstances are really exceptional.

<<As the file is already a CMYK file is it possible, or desirable to settings he used, but the scan was done over a month ago...>>

You are speaking here of the shadow. What you want to do is take a copy of the whole image and reseparate it using Heavy GCR. Then, carefully blend *only* the shadow of that copy into the original CMYK image.

<<The sprocket frame is roughly 1/4" around the 6x6 frame so it sounds as if i'm in the clear on that matter. I suppose that should i need a solid black background for an image that in the future it might be better to add a second black as the fifth color to avoid the issue altogether.>>

That's correct, and is the general approach when you have a large solid area of any color.

<<Being a designer i have always relied on having someone else prepare my scans and deal with the color management issues. But i've also come to the point where some of my clients are paying enough to deserve top-notch color and printing... and those are the jobs that make my end of the work most rewarding and profitable.

Poking around on the net, this list and reading a few books on thebsubject have given me a much better understanding of what goes on after i hand my files off to the service bureau/printer. Prepress folks seem a lot more like artisans who've learnt the rules of the trade by practise rather than scientists running around in white coats (as some printers would have us believe).

It's not an entirely reassuring picture since photoshop has clearly put some of that responsibility in my lap as well... ;)>>

This is all quite true. More and more designers and, especially, photographers, face the same issue. I can offer the sweeping and somewhat inaccurate generalization that of the people I teach, the prepress folk often can execute almost anything they want but really don't know what it is they want. Whereas, the creative types are more image-conscious. They know where they want to go to but not how to get there. You're actually lucky. It's a lot harder to learn what you want to do than it is to learn to achieve what your good judgment tells you you want.


From: Dan Margulis, INTERNET:76270.1033@compuserve.com
Date: Tue, Oct 26, 1999, 2:27 PM

jglazer writes:

<<Another reason not to use process shadows is if the job is being run at a lower lpi (<133), where the rosette will be too strong. The dots of a K only shadow often are more subtle than its process equiv. Of course, this does not apply to higher line screens.>>

There is some merit to this if we restrict the discussion to K only vs. CMY only--the dots will be the same size, and there'll be three times as many of them. However, a CMYK shadow using heavy GCR is the best of all worlds.

There'll be four times as many dots, but each one taking up an area less than half the size of what it would be before. So, it gives much more of an illusion of continuous-tone. This is why most newspapers who can print their B/Ws in 4/c in certain locations do so.


From: Chris Murphy, INTERNET:lists@colorremedies.com
Date: Tue, Oct 26, 1999, 8:43 PM

I wrote:

>>I don't know about that. If your gray has only a relatively small amount
>>of black in it, it's possible for gray balance on that press to be
>>different from what you're expecting. I see presses all over (I'm not
>>claiming this is the right way, I'm saying it's reality) that have some
>>hue cast to a square patch of 50c,39m,39y.

Ron Bean writes:

>How do you know it's the press and not the ink?
I don't. I'm referring to the whole thing as the press. It could be the
paper or the ink or packing or whatever. It's a common problem to have
lack of gray balance on press, that's all my point was. That gray isn't
necessarily 50,39,39 - you can get color shifts on press.
>If it's a perfecting press, do you see the same effect on both
>sides of the press sheet? What happens if you run the inks in a
>different order on the same press?

I'm usually seeing only single sided printing. As for print sequence,
that can certainly affect gray balance. However print sequence is a more
an issue of tack than anything else. If you change the print sequence,
you need an appropriate set of inks, not just the same one's you've been
using. Otherwise you'll get contaminated inks.

>Could you explain the difference between a colorimeter and a
>spectrophotometer (ie, how do they work and what are they
>actually measuring)?

I don't know the inner workings of a colorimeter really. Densitometers use red green and blue filters. Colorimeters probably have the same. The difference is that densitometers report reflected (or absorbed) light through one filter and then reports the value on a logarithmic scale (0.00 to 4.00 I think, where 4.00 is perfect absorption). Colorimeters probably use different types of filters, then compute colorimetric values, usually CIE XYZ, then use math to compute CIE Lab which is used in ICC profiles for output devices.

Spectrophotometers use a prism to split out the wavelengths of reflected light, and are able to measure around 32 bands (or slices) of that separated light. That is spectral data. Usually that is not used hoever.

There is a complex formula that I have from a Heidelberg book on color that requires the reference whitepoint (which would be D50 in this case, but the spectrophotometer could compute it for other illuminants), to generate CIE XYZ out of spectral data. From there it's converted to Lab. All of this is done inside the spectrophotometer, then reported back as CIE Lab to the profiling software.

Generally spectrophotometers are considered "better" just because they start out with spectral data to compute CIE XYZ, whereas colorimeters don't. It doesn't make colorimeters bad or inferior; it just depends on what you're doing. For example emissive colorimeters do a fine job on monitors because they are making assumptions about monitors that are correct. If you use that colorimeter on an LCD display, then you get invalid numbers from the colorimeter. The colorimeter has to be designed to work with the samples it's measuring or the measurements arenincorrect. This isn't the case with spectrophotometers. There is newer spectrophotometer technology being developed that won't require a prism (perhaps already being used) that should be much less expensive to develop, as well as miniaturize for inclusion inside of various devices. So maybe one day all of our devices will just profile themselves, and send the profile to the operating system.

 

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