Dan Margulis Applied Color Theory - Printing Newspaper to a Gray Bar

From: John Denniston, INTERNET:john_denniston@bc.sympatico.ca
Date: Thu, Apr 20, 2000, 2:53 AM
RE: [ColorTheory] Printing to a greybar

Our new presses continue to have problems printing colour.

The theory is that the pressmen should print to a greybar that runs down the fold of each CMYK page rather that adjust colour for the pictures on the page. When ever the pressmen do this, ie get a good greybar, the pictures suffer, mainly because they're over inked. They look flat at best but usually the pictures are too dark and flat with no highlight at all.

The pressmen say they have to print for what's on the page, the techs say the system will only work if press adjustments are made according to densitometer readings from the grey bar.

I'm in the middle of all this, sending out separations without any idea of which group will win the battle today on how to print the next days newspaper.

So I'm wondering, is printing by the greybar is an achievable goal considering we are a newspaper, normally printing a hundred pages a day at about 75,000 copies and hour?

Regards, John

John Denniston,
Photo Editor, The Province
Vancouver, BC
john_denniston@bc.sympatico.ca
jdenniston@pacpress.southam.ca

From: Chris Murphy, lists@colorremedies.com
Date: Thu, Apr 20, 2000, 3:50 PM
RE: Re: [ColorTheory] Printing to a greybar

>The theory is that the pressmen should print to a greybar that runs down the
>fold of each CMYK page rather that adjust colour for the pictures on the page.
>When ever the pressmen do this, ie get a good greybar, the pictures suffer,
>mainly because they're over inked. They look flat at best but usually the
>pictures are too dark and flat with no highlight at all.

The press is not optimized. You need to optimize for ink density, print contrast, and dot gain. Use a gray balance test target (you can make one yourself) to find out what gray balance is for the press under optimized conditions, then maintain it.

There is an erroneous contention that gray is always 50, 39, 39. Gray is whatever combinations of CMY come out neutral once the press is optimized. Maybe for you it's 52, 35, 45. It could be totally bizarre and that's not something to freak out over if you want the best possible printing from this press. If you want the best possible printing, your press will be different from a printing specification as well as Joe's press next door and likely even a conventional analog proof.

If your goal is to force this press to match some other behavior, you might want to get a copy of GRACoL and/or hire someone to spend a day with your pressmen to straighten out the problem. Regardless you'll save money by getting this problem solved ASAP. I can provide a referral for someone who can train in process control if you like.

Chris Murphy

From: Dan Margulis, 76270.1033@compuserve.com
Date: Fri, Apr 21, 2000, 4:42 PM
RE: Re: [ColorTheory] Printing to a greybar

John Denniston writes:

>>The theory is that the pressmen should print to a greybar that runs down the fold of each CMYK page rather that adjust colour for the pictures on the page.>>

No, the theory--and it's a good one--is that it will be better for everyone if the press is run in a consistent way. What that way is is up for discussion.

>>When ever the pressmen do this, ie get a good greybar, the pictures suffer, mainly because they're over inked. They look flat at best but usually the pictures are too dark and flat with no highlight at all.>>

OK, this demonstrates that the inking values that the techheads advocate are considerably higher than what the pressmen have typically been using. It doesn't prove which group is correct.

>>The pressmen say they have to print for what's on the page, the techs say the system will only work if press adjustments are made according to densitometer readings from the grey bar.>>

These are two extremist positions. There *should* be a general set of numbers that represent the press's default. OTOH, the pressmen should not feel like they will be shot if on certain pages the color looks bad and they decide to change it.

>>I'm in the middle of all this, sending out separations without any idea of which group will win the battle today on how to print the next day's newspaper.>>

That's an untenable position. It's going to be very hard for anybody to argue against consistency in printing. There needs, however, to be agreement as to what the best way to print is, so that you can try to derive numbers from that to guarantee that you will print that way in the future. Your print quality needs to be somewhat similar to that of other newspapers or advertisers will start to get upset. However, with new presses and with (I assume, given your geographic location) excellent newsprint quality, you don't want to just blindly follow a standard when even better quality is achievable.

Certain practices will make the press easier to run. They may also cause workflow difficulties for those making the separation. Photographers should not dictate how the press should run and neither should techheads. Neither should pressmen, for that matter. The three groups should get together, present their views, and come up with a consensus method. *Then,* and only then, will the control bars make any sense.

>>So I'm wondering, is printing by the greybar is an achievable goal considering we are a newspaper, normally printing a hundred pages a day at about 75,000 copies an hour?>>

If the question is, will they eliminate variability, no. If the idea is to get a better result than before, yes.

The idea of using a grayscale is probably more appropriate for a commercial printer than a newspaper. I'd prefer to see patches of solid inks, maybe with a couple of gray patches off to the side. The reason being, in a commercial printer if they can't figure out why the grayscale is looking silly they'll stop the run until they can. In a newspaper, the press won't stop, so you want to give the pressmen swatches that are a little easier to decipher. If there's something wrong with the magenta coverage it will show to the naked eye in the magenta swatch. In a gray bar it's hard for most people to tell whether the problem is in the magenta or elsewhere.

Dan Margulis

From: Olivier Herzock, OHerzock@x-rite.fr
Date: Fri, Apr 21, 2000, 10:58 AM
RE: Re: [ColorTheory] Printing to a greybar

Hi John,

You need first to generate the Calculation of the real print contrast of each ink per each group, the print contrast give to you the best result for contrast in the dark area of the images, at this time you can generate your own control bar (Grey if you want).

The print contras use a specific formula, these is present inside all the real densitometer , like the 518 or 528 from X-Rite.

The ATS or ATD from X-Rite can help you to generate automatically some kind of report about the PressOptimsation

Contact your X-Rite Dealer to get all the information about the Press Optimization.

Regards
Olivier


******* ******* ******* ******* ******* ******* ******* ******* *******
Olivier HERZOCK
Specialiste d'applications Industries Graphiques & PAO
X-Rite Mediterranee (France)
35 rue du Saule Trapu
91300 Massy ( France)
Tel: 33 (0) 1 69 53 66 20
Mobile : (0)6 09 56 90 56
Fax: 33 (0) 1 69 53 00 52
OHerzock@x-rite.fr
OHerzock@xrite.com
http://www.xrite.com

******* ******* ******* ******* ******* ******* ******* ******* *******

From: John Denniston, john_denniston@bc.sympatico.ca
Date: Sat, Apr 22, 2000, 5:48 AM
RE: Re: [ColorTheory] Printing to a greybar

>There is an erroneous contention that gray is always 50, 39, 39. Gray is
>whatever combinations of CMY come out neutral once the press is
>optimized.

Chris,

We are planning a comprehensive press test in the next few weeks which will be set up to, if not optimize the press, at least get us enough information that we can begin that process.

Your answer, however, is interesting because I've always assumed our greybar was accurate and that, when printed, if it looks grey and of about the proper density, the problem was with what ever happened pre press. In other words, it's all my fault. :-)

I take it that you are saying if the greybar is created with a set of theoretical numbers or from a press test with different inks or some other variable it would make the images out of sync with the greybar on the negative?

If that is true I will make that a part of the upcoming test or at least find out how the greybar is created.

If I sound a bit naive about this it's just that my background is photography and my current responsibilities came about because of an interest in computers and the fact that the people who really knew about this stuff were laid off during the great tech change revolution that occurred everywhere in our industry about 5 years ago.

Regards, John

John Denniston,
Photo Editor, The Province
Vancouver, BC
john_denniston@bc.sympatico.ca
jdenniston@pacpress.southam.ca

From: Chris Murphy, lists@colorremedies.com
Date: Sat, Apr 22, 2000, 9:09 AM
RE: Re: [ColorTheory] Printing to a greybar

>Your answer, however, is interesting because I've always assumed our
>greybar was accurate and that, when printed, if it looks grey and of about
>the proper density, the problem was with what ever happened pre press. In
>other words, it's all my fault. :-)

There are several things going on that could be causing the confusion. First there are basically two ways to print, and if you're a wise ass like me you'd say there are three ways to print (on printing presses):

1.) Per a standard or specification such as GRACoL, SWOP, SNAP, or FIRST. This means you force your press to behave per that specification, not better. Now the catch is that it's not 100% possible to force the press to behave per those specifications by themselves, GRACoL is easier than SWOP because SWOP is more specific. Separations are made generic, including gray balance. The job of the printer in this case is to use the specifed inks, paper, and press condition in order to achieve this. To fine tune it they might need to use transfer curves - but as far as you are concerned this is irrelevent. You separate per a standard spec, and they print the same spec, everything comes out fine. Naturally they need a proof to match because the specification aren't very specific (except for SWOP which is quite specific).

2.) Optimize the press for its best possible behavior. It's a good idea to start out with an applicable specification and then do better. In this case you really need to fingerprint (characterize, or profile) the press in order to find out how it behaves because generic separations aren't going to work as well (note I'm saying AS WELL, the pressman is going to be able to fake it pretty well considering the separation isn't ideal for his press); but to take advantage of a press's best behavior you need custom separations. In this case gray balance could be totally different than for a specification because you aren't forcing the press to behave that way.

3.) Wise ass part of this is that the pressman doesn't do either 1 or 2, they do something different entirely and no one really knows what's going on. While those exist I think you find a large number of printers that are somewhere between 1 and 2. They try to get better press performance yet they use analog proofs that attempt to force them into some other standard behavior and that's where problems come up. Using a proof that isn't designed for the press condition.

Now the other aspect of all this is prepress vs printer. Prepress I will consider those who are making separations. They need to know the gray balance of the press being used in order to make separations. If they are guessing they're asking to redo the work (the separations).

If your gray bar is made up of 50, 39, 39 (or whatever) and you are able to make gray out of it, then you need to ensure that prepress is separating with that gray in mind. If you are having problems with that gray, you need to do a gray balance test to find out what values make up gray then tell prepress what those values are so they can make separations with that gray in mind. Even if it's whacked like 60, 30, 50 - as long as it prints gray and you're getting wonderful saturation, the separation compensates for all of the other differences including gray balance.

While you can take generic separations and print them on such a press, you will need to use transfer curves in order for them to have even a hope of printing out reasonably well. However nothing beats a separation made for those press conditions. It's just a matter of whether it's good enough to just use a transfer curve and get the job out the door (and the customer is happy with it), vs. going back and having to make new separations.

>I take it that you are saying if the greybar is created with a set of
>theoretical numbers or from a press test with different inks or some other
>variable it would make the images out of sync with the greybar on the
>negative?

Yes. When the scan is made you either scan in RGB or CMYK. If you are scanning in CMYK you MUST know what makes gray on the press that will be used. If you are scanning RGB, you only need to know gray is equal amounts of red, green, blue. When you convert using a separation table or ICC profile, it knows what makes gray on that press and will make separations accordingly.

Then the job on press is to make sure that the gray balance test patch remains gray. And if this is a problem, then a gray balance test run needs to be performed to find out what makes gray, then inform prepress of the new numbers so they can make seps with that gray in mind.

Chris Murphy

From: Chris Murphy, lists@colorremedies.com
Date: Sat, Apr 22, 2000, 2:04 PM
RE: Re: [ColorTheory] Printing to a greybar

>The idea of using a grayscale is probably more appropriate for a commercial
>printer than a newspaper.

I might disagree. I think it might be more realistic for commercial printing than newsprinting, but I definitely think it's more valuable for newsprinting.

With newsprinting you're hoping you don't introduce color casts more than anything else. So you want to make sure that images are gray balanced (customer's responsibility) and separated with a target gray balance in mind; the press should also be graybalanced with those same numbers in mind. While blues might be more or less saturated than the customer was expecting, hue should be mostly preserved and color casts won't be introduced if the press is gray balanced. Whereas with commercial printing the customers there are usually thinking they can get exactly or very close matches. With newsprint I think this is asking a lot, but I think gray balance would be very useful for avoiding the introduction of a color cast.

Again, a classic example is a magenta cast on the sunblock ad. That's one reason why you don't see a lot of four color sunblock ads in newspapers. Four color cosmetics ads aren't as common as black and white likely for the same reason.

> In a gray bar it's hard for most
>people to tell whether the problem is in the magenta or elsewhere.

I think it's valid to print a 50% black next to what should be 50% gray using CMY and visually analyze that if measuring is going to be a difficulty.

In a way I would say more focus on the neutral gray would avoid the truly distasteful color matching problems from happening, and perhaps make it a secondary priorty to control SID and dot gain - once make ready is complete.

Chris Murphy

From: Dan Margulis, 76270.1033@compuserve.com
Date: Sat, Apr 22, 2000, 10:09 PM
RE: Re: [ColorTheory] Printing to a greybar

Chris writes:

>>With newsprinting you're hoping you don't introduce color casts more than anything else.>>

This is exactly backwards. What you are discussing is the never-ending war between color and contrast. Certainly we want both, but if we can only have one, as output conditions get worse, retention of contrast becomes more important and color fidelity less important.

>>Again, a classic example is a magenta cast on the sunblock ad.>>

This is indeed a big problem for newspapers. A much larger percentage of newspaper images involve fleshtones than would be the case at a commercial printer. This would be an argument for printing patches of typical fleshtone value in addition to, or instead of, gray patches on the color bar.

>>I think it's valid to print a 50% black next to what should be 50% gray using CMY and visually analyze that if measuring is going to be a difficulty.>>

It's certainly valid, but you have to remember that there is a very limited spcace in the fold. It would be better if we could run a whole set of Brunner bars but we can't. We're limited to maybe half a dozen small swatches.

>>In a way I would say more focus on the neutral gray would avoid the truly distasteful color matching problems from happening, and perhaps make it a secondary priorty to control SID and dot gain - once make ready is complete.>>

Nobody is against having good gray balance but one has to be real-world about it. A newspaper press does not stop for any technical reason other than a web break. Decisions during the run are going to be made not by machine but by human eye, like it or not. So, the control swatches should be those that the human eye is most comfortable reading. If there's space for gray swatches, fine, but they won't be the primary way of maintaining color.

Dan Margulis

From: Chris Murphy, lists@colorremedies.com
Date: Sat, Apr 22, 2000, 10:45 PM
RE: Re: [ColorTheory] Printing to a greybar

>>With newsprinting you're hoping you don't introduce color casts more than
>anything else.>>
>
>This is exactly backwards. What you are discussing is the never-ending war
>between color and contrast. Certainly we want both, but if we can only have
>one, as output conditions get worse, retention of contrast becomes more
>important and color fidelity less important.

OK I don't think I stated or implied that I would sacrifice contrast at the expense of color fidelity. I'm already accepting that I'm not going to get the exact color I want if I'm printing newsprint. Yes you want as much contrast as possible, but if I'm printing an ad in a newspaper four color, I would want to make sure my humans look human, not martian (no offense to martians), and my products don't have undesirable color casts. This doesn't mean sacrifice contrast.

So I would amend my previous statement to say, with newsprinting you're hoping you don't introduct color casts or sacrifice contrast more than anything else. My point of saying what I did is that I'm less concerned about dead on color than I am about the introduction of color casts on press; hence I feel it's at least as important if not more important to maintain gray balance on newspaper presses (even if it's going to be difficult, and not realistic to do it as often as the ideal).

>>>Again, a classic example is a magenta cast on the sunblock ad.>>
>
>This is indeed a big problem for newspapers. A much larger percentage of
>newspaper images involve fleshtones than would be the case at a commercial
>printer. This would be an argument for printing patches of typical
>fleshtone value in addition to, or instead of, gray patches on the color
>bar.

Interesting idea.

>>>I think it's valid to print a 50% black next to what should be 50% gray
>using CMY and visually analyze that if measuring is going to be a
>difficulty.>>
>
>It's certainly valid, but you have to remember that there is a very limited
>spcace in the fold. It would be better if we could run a whole set of
>Brunner bars but we can't. We're limited to maybe half a dozen small
>swatches.

New technology will provide solutions; although it likely won't help us with existing equipment.

>Nobody is against having good gray balance but one has to be real-world
>about it. A newspaper press does not stop for any technical reason other
>than a web break. Decisions during the run are going to be made not by
>machine but by human eye, like it or not. So, the control swatches should
>be those that the human eye is most comfortable reading. If there's space
>for gray swatches, fine, but they won't be the primary way of maintaining
>color.

Newer presses and the increasing use of CIP 3 systems will likely make this moot where ALL of these critical colors can be controlled constantly, within seconds of being printed, without the pressman needing to be involved (the system measures and makes automatic adjustments to ink keys and items normally modifiable only by humans.)

Chris Murphy

From: GMattei@aol.com, GMattei@aol.com
Date: Sun, Apr 23, 2000, 7:39 AM
RE: Re: [ColorTheory] Printing/shooting to a greybar

From reading these posts I cannot determine if I were to shoot a photograph with a digital camera, would it or would it not be helpful to include a MacBeth Color Checker in the photo to give the printer a point of reference for color?

Thanks
George

From: David Riecks, david@riecks.com
Date: Sun, Apr 23, 2000, 1:34 PM
RE: Re: [ColorTheory] Printing/shooting to a greybar

At 11:18 AM 04/22/2000 EDT, George Mattei wrote:
>>From reading these posts I cannot determine if I were to shoot a photograph
>with a digital camera, would it or would it not be helpful to include a
>MacBeth Color Checker in the photo to give the printer a point of reference
>for color?

George:

I had heard that Gretag Macbeth (http://www.gretagmacbeth.com) had come out with a new version of the color checker that's better for digital photography. It has more fleshtone patches, etc. However, I can't seem to find it on their site. Latest press release is 4-12-2000, but no word on this product. Of course I can't recall where I read about it now<grrr>.

I did leave a message off the website, so if I hear anything back I'll post it to the list.

I've used the Kodak Color Separation Scales (Q-13 and Q-14) before and they help if the firm is familar with what these scales are, and has a set on hand. Then again if they don't have these it'll probably be of little value to the printer, but may help you to get your initial RGB values more in line...which should lead to more consistent (and one would hope better) results.

David
David Riecks * david@riecks.com
701 W. Washington St * Midwest/Chicago ASMP
Champaign, IL 61820 *
ph/fax 217-239-FOTO(3686) * http://www.riecks.com/

From: Dan Margulis, 76270.1033@compuserve.com
Date: Sun, Apr 23, 2000, 10:59 AM
RE: Re: [ColorTheory] Printing/shooting to a greybar

George writes:

>>From reading these posts I cannot determine if I were to shoot a photograph with a digital camera, would it or would it not be helpful to include a MacBeth Color Checker in the photo to give the printer a point of reference for color?

Having objects in the original image that are known to be gray is helpful. Anything more than that is normally of little value. A lot of lip service is paid to the idea of "matching the art". In reality, in printing one never wishes to match the art precisely. Thus, the fact that we know exactly what a certain color is in the original photograph is usually irrelevant.

The exception to this is a gray object. If something in the photograph is known to be 50% gray, there is again no guarantee that it should reproduce as 50% gray. However, it should reproduce as *some* kind of gray most of the time, even if lighter or darker. This can be a valuable aid to whoever is color-correcting the image.

The bottom line, therefore: including a grayscale in an original photo is a good idea if convenient, including other colors won't hurt but rarely helps.

Dan Margulis

From: Andrew Rodney
Date: Sun, Apr 23, 2000, 2:20 PM
RE: Re: Dan-Re: [ColorTheory] Printing/shooting to a greybar

on 4/23/00 11:34 AM, David Riecks at david@riecks.com wrote:

> I had heard that Gretag Macbeth http://www.gretagmacbeth.com had come out
> with a new version of the color checker that's better for digital
> photography. It has more fleshtone patches, etc. However, I can't seem to
> find it on their site.

Not due out for another few weeks. I have one (pre-release) and it works quite well. I've been using in their ProfileMaker 3.0 software and getting very good results. There should be two versions available; one that will have an average TDF file (costing about $250 from what I've been told) and one costing $100 more with custom measurement data for that exact target.

Andrew Rodney
http://www.digitaldog.net

From: Alan Kimrey, alan2@wakpictures.com
Date: Mon, Apr 24, 2000, 11:55 AM

George wrote:
> From reading these posts I cannot determine if I were to shoot a photograph
> with a digital camera, would it or would it not be helpful to include a
> MacBeth Color Checker in the photo to give the printer a point of reference
> for color?

We shoot tabletop digitally with the mini colorchecker, a tiny Macbeth chart. It goes in every shot per printer request. It's used only for the greyscale, as this stuff gets printed on freezer packaging with a very low dynamic range. We set the black and white points per press specs within the camera software, neutralize tones along the scale and tweak gamma within those limitations.

The printer's getting shots of the same stuff from different studios, so they've figured out this maintains some consistensy and assures they have a fullest range of tones possible for the end use. It also gives us the opportunity to see just how limited the press specs are and to do what we can to optimize the image data at capture. Getting decent contrast with an 80% black point and a 10% white point is a neat trick.

When we started with this client, the printer sent us a sample of previous shots, telling us to match the Macbeth chart. Man, I wasted so much time trying to match the color values, until they finally clued me in that only the greyscale was important. Live and learn.

Color wise, they toe the line with "shoot for the grey and the color will follow."

__
Alan Kimrey
WAK Pictures, Inc.

From: GMattei@aol.com, GMattei@aol.com
Date: Mon, Apr 24, 2000, 11:18 AM
RE: Re: Re: [ColorTheory] Printing/shooting to a greybar

In a message dated 4/23/00 3:00:13 PM, Dan wrote:

>>The bottom line, therefore: including a grayscale in an original photo is a good idea if convenient, including other colors won't hurt but rarely helps.>>

Dan, Thanks for the input. One other related question. If it is impossible for me to include a grayscale in my photo due to space, can I shoot 2 photos, one with the grayscale visible and another photo without the grayscale. The printer could take his readings form the photo with the grayscale than switch to the photo without the grayscale to do his conversions, printing etc.

Thanks George

From: Dan Margulis, 76270.1033@compuserve.com
Date: Wed, Apr 26, 2000, 4:27 PM
Re: [ColorTheory] Printing/shooting to a greybar

George writes:

>>If it is impossible for me to include a grayscale in my photo due to space, can I shoot 2 photos, one with the grayscale visible and another photo without the grayscale. The printer could take his readings form the photo with the grayscale than switch to the photo without the grayscale to do his conversions, printing etc.

This is an interesting idea that is probably more trouble than it's worth. Obviously, the two shots would have to be with absolutely identical exposures and lighting conditions, meaning studio work only.

This seems like a lot of effort for both you and the prepress folk to go through. It can't hurt, however, to talk to the person handling the corrections to see what his views are.

Dan Margulis

From: Raymond Cheydleur, rcheydleur@xrite.com
Date: Wed, Apr 26, 2000, 4:36 PM
RE: RE: Re: [ColorTheory] Printing/shooting to a greybar

>>This is an interesting idea that is probably more trouble than it's worth. Obviously, the two shots would have to be with absolutely identical exposures and lighting conditions, meaning studio work only.>>

Actually shooting a separate grayscale or Mussel chart is relatively common for some newspapers doing event coverage like a football game, convention etc. where the lighting is reasonably consistent, and they are shooting with a digital camera. They use the shot of the target as the baseline correction for the event.

RayC
Raymond Cheydleur, Senior Application Support Specialist,
Imaging and Graphic Arts, Developer Support
X-Rite Incorporated
Application Support 888-826-3046, Option 1
Customer Support 888-826-3059

From: DMargulis@aol.com, DMargulis@aol.com
Date: Fri, Apr 28, 2000, 5:02 AM
RE: RE: Re: [ColorTheory] Printing/shooting to a greybar

>>Actually shooting a separate grayscale or Mussel chart is relatively common for some newspapers doing event coverage like a football game, convention etc. where the lighting is reasonably consistent, and they are shooting with a digital camera. They use the shot of the target as the baseline correction for the event.>>

Unless the football game is played indoors and the photographer never changes his location, this is a poor idea. In a game played in sunlight the colors of the uniforms will not be the same in the second half as in the first.

Similarly, in a convention hall where there are typically banks of strong lights, lighting will not be consistent from area to area. Therefore, if the photographer shoots everything from one location, he can use this method, but if he moves about the hall, it won't work well.

Dan Margulis

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