Gordon Pritchard, Kodak's Value In Print Initiative Marketing Manager has a blog post up on Grow Your Biz in which he outlines 11 items to consider when using gray balance as a press control:
Achieving gray balance in presswork is the major mantra in today's print production world. It is the paramount metric - now defined in unambiguous CIE L*a*b* values - for the G7 calibration method to align press and proof color as well as for achieving presswork that conforms to GRACoL 7. That being said, one must keep a "balanced" view on gray balance as a metric in presswork. While it certainly has value - interpreting that value requires some understanding of the peculiarities of the press room as well as a bit of history.
Head over to Gordon's post for his laundry list of considerations. What do you think of the list?
Discussion
By Erik Nikkanen on Sep 24, 2008
Gordon, you are right on track while the rest of the gray balance crowd is derailed but don't know it.
You are much too nice with your comments but I can understand that. I would be inclined to say that the idea of gray balance of an image is a myth. It is pseudoscience like a lot of other accepted truths in the printing industry.
Rational discussion is quite absent in the industry but your comments help to point us in a more correct direction. Thanks for posting them.
By Joseph J. Pasky on Sep 25, 2008
I think that Felix Brunner's work over the past 30+ years has demonstrated that gray balance must be the primary consideration at press.
G7 is simply a new tool to achieve that same end. (Balance can be monitored with a densitometer or visually, btw. No spectro required). In order of importance: gray balance, tone and color. It's not that difficult to achieve on a well-maintained press.
By Joseph J. Pasky on Sep 25, 2008
Here are some additional thoughts on Gordon's piece:
GP: 1. As far as I can determine there has never been a formal study of gray balance targets measured in press sheet color bars and how they relate to the live image content of the press sheet.
JJP: I always thought it was pretty well accepted that any color system must be able to produce a neutral gray. In my video camera and digital SLR, it’s called ‘white balance’. I would think that proofing systems and presses being ‘color systems’ that the same requirement would apply.
GP: 2. Press operators do not "make color" on press. They concentrate on what a press is designed to do - lay down a consistent film of ink approximately one micron thick on the substrate.
JJP: Well, I think that is exactly the point. I want a pressman that will operate the press exactly the same way day after day after day. I don’t want an artist; I want a boring perfectionist. David McDowell wrote in a white paper someplace, that the ability of the press to be manipulated to match a out-of-spec proof has been highly overrated. (I think I paraphrased him correctly). The mantra of G7 (and good presswork in general, I believe) is that the press operation must be consistent, day after day. The prep guys can’t hit a moving target.
GP: 3. Press operators use densitometers - not to measure color, but to indirectly measure ink film thickness.
JJP:
If they use them at all.
But, a densitometer will tell you how much the ink-film has modulated the light coming from the white paper…and if this changes as the press is running.
GP: 4. The deepest study that has been done on gray balance was by System Brunner using densitometers rather than spectrophotometers to measure gray balance.
You have to consider however, that Brunner’s techniques were developed in a world of film and plates and where spectrophotomers were had the same price tag as a Porche 911. Densitometers were not very clever either. His course-screen/fine-screen technique to indicate dot gain in terms of a density target was cool at a time when densitometers didn’t do gain calculations.
GP: 5. The ISO specification (e.g. ISO 12647-2, ISO 2846-1) defines gray balance as "neutral" - a very vague term.
JJP: Yes, that is true, and also the reason why G7 has suggested a specific description of gray: a*=0; and b*=-2. And because spectros are quite common nowadays, we can easily measure the balance throughout the entire tone ramp.
GP: 6. A press lays ink down in a series of zones which run from the lead edge to the tail of the sheet. If a specific process color (C, M, Y or K) is not require in that zone because of image content, that ink zone is switched off and gray balance is no longer achieved in the color bar.
JJP: The point of monitoring gray balance is to achieve an accurate and consistent reproduction of 4-color process images. Even if there is no process image in that zone, we are still measuring and controlling SIDs.
GP: 7. Image content that requires a substantial increase in solid ink density in some colors (e.g. a sunset scene) will result in gray balance no longer being achieved in the color bar.
JJP: I disagree. The solid ink density as measured in the color bar, should NOT change as a result of the image on the plate. The amount of ink that is fed to that zone (by changing the ink-key setting) will supply the correct amount of ink required of the plate. If the ink key feeds the plate 0.001kg of ink, and the plate takes away 0.001kg of ink, the resulting ink film thickness at the color bar does not change. The readings of SID (and tone) across the press sheet will be the same. (I made-up those numbers).
If, in fact, you are fighting these ‘in-line’ problems, it suggests that the press/plate curves are not calibrated properly. G7 uses a P2P target to establish these curves; Brunner calls the relationship between tone and solid, positive and negative “conformity”.
GP: 8. The correlation of gray balance distortion (i.e. color bias) vs perceived color shift in quarter, mid and shadow areas has not been studied. It is probable that it is very non-linear. That is to say that the shifts we see in the gray balance target in the color bar may shift quite a it, however, we will not necessarily see an equal shift in highlight or shadow color.
JJP: That may be: Changes in shadow are less perceptible than in the highlight ranges, However, I would suggest that Brunner’s PCT (Print Contrast Theory) is instructive in determining what the range of tolerance is acceptable as defined by different types of picture images. A 2% variation in midtone gray balance (3.0 Delta E as I recall), is quite acceptable for most all picture categories.
More striking is the effect of overprint trap in color shift of the tonal areas as you go up the scale. These overprint trapping variations are tough to measure on-the-fly except with as system like Brunner’s Instrument Flight on Quad/Tech’s closed-loop color system. (this turns out to be a terrific color data collection and analysis tool…you can watch the dynamics of color change and press print performance as it’s happening)
GP: 9. Gray balance targets are likely too sensitive to normal color fluctuations on press to be of practical use. Just like a car's speedometer would be useless information if provided 3 decimal place readouts of speed (e.g. 50.392/kmh, 50.471/kmh, 50.148/kmh).
JJP: I read someplace that the measuring device should be 10 times more accurate than the thing you’re measuring. Nobody is saying that you must control balance to 3 decimal places, but it’s important to be able to see the variation before it affects the printed image.
In Brunner terms, the gray patch is a ‘class zero’ picture: no contrast; no color. The variation in the printing process makes it impossible to be reproduced consistently on the press…which makes it a very good measuring tool.
GP: 10. It is dangerous to apply gray balance principles derived from scanning and proofing to the pressroom since the mechanics of how color is achieved is radically different.
JJP: I’m not sure I would agree. Again, I would argue that every color system must be able to produce a neutral gray, which would include proofing systems. G7 is based upon ‘appearance’…how you get there is not the important question. I think the point has been made repeatedly in ongoing testing that a ‘certified’ G7 proof and a G7 calibrated press will match. There are some cautions, of course: standardized lighting is most important when evaluating proofs and press sheets.
GP: 11. The job of the press operator is to align the "live" image area of their presswork with the proof by achieving the appropriate solid ink densities since printers sell the live image area of the press sheet - the color bars go into the recycling bin.
JJP: And I’m sure that on many occasions, the entire job goes into the recycle bin along with the trim waste.
This often happens when the operator finds it impossible to ‘align’ the press sheet color with the proof and ends up compromising the color to his best judgment (sometimes: ‘whim’). Now I would agree that if the proofs are good (around 1 Delta E) and the profiles are correct, and the press/plate curve is correct/consistent, matching the proofed image should not be difficult, with or without color bars. With the new closed-loop systems that are available that read very narrow bars, that small piece of real-estate seems a small price to assure consistency.
GP: Gray balance targets in presswork color bars certainly have some value - but interpretation of the information they contain must always be considered in the context of the mechanics and chemistry of the press itself.
JJP: If the objective is to reproduce good color on a consistent basis, I think we have to make the assumption that ‘good presswork’ and a well-maintained press is a ‘given’. That would include mechanics, chemistry, rheology, temperature, humidity…etc. etc. The advantage of an ‘appearance based’ control method is that how you get there is not as important as getting a visual match.
To make the color bar more ‘user friendly’ to printers I work with here who are accustomed to visually matching everything (nary an instrument in sight), I developed a ‘print buyers’ color bar that I use when doing color ok’s for my clients.
The color bars I usually see on press have a lot of ‘mechanical’ stuff like slur indicators and plate resolution targets, dot gain patches, etc. (that I’ve yet to see an operator here use). These are fine and dandy to troubleshoot a press problem, but that’s not what I’m interested in. I believe gray balance is most important, so I replaced all of those mechanical indicators with simple midtone patches at 25/50/75 of both K and cym. Next in importance is SID, so I put enough solid patches to get a good indication of how even the ink distribution was across the sheet. Then I put in a couple of RGB patches…there is not much an operator can do in the middle of a zone to affect trap. This bar works very nicely because it is quite simple and visual: I can tell immediately if the press/plate curves are correct (‘conformity’) and if balance is achievable at normal SID’s. My instructions to the operators then becomes simple: make the patches gray, match the tones of the k and cym, make everything even across the sheet.
Here is nice summary of G7 I’ve posted at my blog site:
G7 ON A POST-IT NOTE
A guide to getting repeatable color matches on press, everyday
The G7 methodology has been criticized as being 'too complicated'.
It is not.
Here is a very nice summary of the process that was orignially published by Glenn Andrews of Schawk in Los Angeles. (with a couple of small changes of mine).
1. Use good paper and good ink.
2. Have the press in good running condition.
3. Include color bars with solids, overprints, a 50%C 40%M 40%Y neutral and a 50% or 53% midtone K patch.
4. Run to industry-standard solid ink densities.
5. Mid-tone density must be 0.54 plus the density of the paper.
6. 50%C 40%M 40%Y neutral must be exactly the same color and value as the mid-tone K.
A press sheet run to these specifications will match a good proof.
By Dwane on Sep 26, 2008
Hi Joseph,
Thanks for the G7 on a post-it note. I personaly like the idea of measuring gray balance. If it vary's the most, then it's what we should be looking at. Also the simplicity of taking two measurements for 4 colours, intead of 4 (which never happens). And a good visual match ability.
Your recommendations on the colour bar are also appreciated.
By Erik Nikkanen on Sep 27, 2008
Again I still say that Gray Balance of an image is a myth and has little to do with colour reproduction, but the usual rational reasons do not seem to be accepted so no point in going over them again.
Since G7 is totally based on the gray balance concept, what has the G7 group said about their own method. If we look at their earlier specification we find.
Page 50 of the G7 Guidelines and Specifications 2007
www.idealliance.org/g7global/specification/IDEA_finalG7.pdf
Limitations of G7
The G7 method controls gray balance and neutral
density, but not color reproduction. A properly G7-calibrated device should produce gray tones that are extremely similar to any other G7- calibrated device, but colored areas may still be different due to variations in ink colors, ink trapping, etc. (This is also
true of TVI-based calibration.)
Neither G7 nor TVI calibration alone can guarantee to simulate a reference CMYK color space. Some additional color management may be necessary to compensate for ink or screening variables.
Andrew Tribute also stated in an article http://members.whattheythink.com/evt/08/drupa/drupa08tribute1.cfm
that the following statement was in a G7 publication but I could not find it now. It said:
“Remember that G7 calibration only controls gray scale appearance, not the appearance of colored areas. Depending on the device you are calibrating, some additional color management (for example ICC profiles) may be needed to optimize the match to a specific reference print condition.”
I think these statements in earlier G7 publications are quite honest and clear. G7 can not ensure colour matching. The G7 method is only meant to match the gray scale between devices.
It does not seem that these kinds of clear statements of the performance of G7 are in more recent publications. The impression being given to the industry is that this is an accurate way to reproduce colour and this is not true.
The conclusion is that all of this effort is being made so that different devices can print CMY gray scales predictably. If there was a big market for printing these gray scales, that would be great but the market is really buying the colour of images, which G7 does not have control over. And that is according to G7 publications. Then add in that fact that most images probably have some form of GCR and any CMY gray is not even in the image.
Also let’s make something a bit more clear. G7 does not change the way the press prints. The idea that G7 calibrates the press is misleading. A particular press will basically print the way it prints. G7 calibrates the prepress software with the purpose of having the gray scale match other devices.
G7 might be marginally better than TVI, but both do not address colour reproduction. The damage that the well intentioned G7 hype is doing to the industry is that it is delaying the development of better methods based on a more rational and scientific foundation that will address colour reproduction that would be simple, consistent and predictable.
By Joseph J. Pasky on Sep 29, 2008
“Again I still say that Gray Balance of an image is a myth and has little to do with colour reproduction…” - Eric
If gray balance has little to do with color reproduction, it certainly has been a persistent myth:
“In photomechanical reproduction, [ ] neutral reproduction of neutral grays is required”. J.A.C. Yule, 1967.
By Dan Wilson on Sep 29, 2008
Firstly thanks Joe for an excellent summary of the G7 calibration technique, a Draft 2008 version has been posted at www.g7global.org for comment.
The ISO (International Standards Organization) has just endorsed (or looks likely to do so soon) the "Near Neutral" method of press calibration i.e. gray balance method of press calibration.
This will be along side the traditional PSO TVI method and another new method using Device Links.
Using ink and paper that matches the ISO Standard, the G7 methodology calibrates a press that will produce printed sheets who's TVI is within the ISO tolerances - no ifs, no buts.
Therefore the color matches what the ISO Standard 12647-2 requires.
Grays not assisting the matching of color . . .
Might pay to talk with one of the 253 Gracol Master Printers to see what things are really like in production. Presumably tens of thousands of print consumers are happily paying invoices for print produced using the G7 methodology.
Gracol run regular G7 Professional training sessions which offer not only theoretical background but hands on training.
By Greg Imhoff on Sep 29, 2008
What is GRACoL G7 and why is this useful?
G7 is a globally driven print reproduction specification based in ISO standards, with industry best business practices driven in demand by brand owners and printers alike, each desiring improved on press quality results profits and growth. G7 is a digital calibration method based in Offset born of CTP imaging benefits to improve any 4-color output and may also be used as a method and control tool for Flexo, Gravure, Inkjet or any 4 color process.
G7 is about process control so if certified by Idea Alliance as a G7 Master Printer you will see improved results. For SWOP or G7 gray balance control on press this is mainly about printing to Neutral Print Density Curves (NPDC) more efficiently based upon your individual press characteristics...
Gray Balance has long been recognized as the best metric to judge color because we are most sensitive to shifts be it warmer, colder, darker, lighter etc. On press the problem has been the tools to implement gray balance effectively were never present. Spectral values are a better metric than Densitometric however there was a disconnect with most press operators not “speaking spectral” now solved in my patent pending solution which uses the G7 specifications as “a bridge” to link both Densities and Spectral resulting in all values displayed including ink key directional moves in 1 measurement.
Erik mentions page 50 just above that on page 49 is a description of NPDC to TVI (dot gain):
>Snip>
Page 49: E.5 Relationship of NPDC to TVI
When correlated back to TVI curves (an inexact process) the G7 NPDC curves lie between the old Positive and negative curves in ISO 12647-2, but closer to positive than negative. To further align GRACoL 7 with other interpretations of ISO 12647-2, such as FOGRA and ECI in Europe and SWOP in the USA, the G7 black curve was lightened in February 2006.
End Snip>:
The main variables affecting on press color are primarily based in CMYK plate imaging curve shapes dot or spot geometries and, from the combined CMYK ink and paper color characteristics. So it stands to reason when most related variables are known managed and adjusted to do run the best results. G7 is about process control and printing to Gray Balance is a logical useful metric for profitable print manufacturing.
Actively applied G7 print operation does improve make-readies, extends your press gamut to hold and match greater proof gamut’s but more importantly, to hit color consistently which Brand Owners are driving you to.
By Bob Chung on Sep 29, 2008
Gray Unbalanced: Points and Counterpoints
Gordon Pritchard's article on Gray Unbalanced (Sept. 24, 2008) from http://growyourbiz.kodak.com/default.asp?item=226515 was circulated at RIT. Franz Sigg (FS; [email protected]) and Bob Chung (BC; [email protected]) offer their counterpoints.
FS: Obviously, a light tint in the "live" image is less affected by a change in solid ink density, so what? If you had that information, what would you do with it? Would you set different tolerances in inking for ink zones with only light tints for each job? if you can do this, go right ahead. What counts is, how the gray balance targets relate to ink film thickness, not how they relate to the "live" image.
BC: (a) Gray balance was studied by scatter plot of a*b* of 3C neutrals and ∆C* vs. time. It is demonstrated in the On-site Excel template; (b) correlation between gray balance analysis and color variation of live images is a function of the type of color images or color image contrast factor.
FS: Agreed, in other words, we are trying to print to the numbers.
BC: Repeatable color, i.e., same (color or density) numbers every time you print, requires that substrate and ink are defined. In addition, press is calibrated to lay down a uniform of ink every impression.
FS: Correctly so, because all we need is process control: constant ink film thickness and dot gain.
BC: (a) The goal of printing process control is to achieve conformance by detecting changes in printed results and adjust the process accordingly based on the process know-how; (b) Densitometer has been effective process control tool because density number is more sensitive to ink film thickness changes than colorimetric numbers.
FS: You do not need spectrophotometers to check gray balance, densitometers can be used just as well.
BC: (a) Brunner’s ‘color image contrast’ is a way to assign criticalness of gray balance to different classes of color images. So, it’s a quality assurance method, and not a process control method; (b) A RIT study, published in Test Targets 7.0, shows that spectral density where the maximum absorption of the ink occurs is the best process control signal for CMYK and spot color inks.
[Editor's note: Test Target 7.0 can be downloaded from http://cias.rit.edu/~gravure/tt/html/tt70.html" rel="nofollow">http://cias.rit.edu/~gravure/tt/html/tt70.html]
FS: Neutrality on the press sheet depends on the chosen tone values of the test patch, and on the inking and paper color. If the "wrong" tone values are chosen in prepress, then the print will not be neutral. A color profile made from the standardized printing condition can be interrogated to find out which CMY tone values will result in a neutral color on the print. This then becomes the specification for the gray patch in the color bar. If we assume standardized printing conditions (ISO), then we can use a standardized color profile (Fogra 39) to come up with the required tone values.
BC: (a) ISO 2846-1 defines the color of process inks in the can; ISO 12647-2 defines the color of he solids and their overprints as printed on defined paper substrates. Both ISO standards are meant to define the gamut colors, thus, not in-gamut colors. Thus, neutrality is not viewed as a primary significance. If, however, neutrality becomes a primary importance, the practitioners can meet the requirement via the use of the device link profile between the target press and his press.
FS: Even if ink is "switched off" there is still ink on the rollers coming in from the side. However it is not well controlled. If for example no cyan is used in a given ink zone, but magenta and yellow are used and important, a smart analysis system could ignore the cyan content for the no longer gray patch in question, and only control the redness of the "gray" patch. The same problem occurs if process control were attempted with single color solids and tints only, this is not specific to gray patches. If there is a small image that takes little ink, then, if it is important that it too has correct inking, it may be necessary to add an ink take-up bar at the end of the sheet that consumes some ink. But dont' hold these problems against the gray color control bar.
BC: We’re speaking about exceptions as opposed rules. Further more, we’re speaking about offset presses with ink keys that can be adjusted locally and not flexo, gravure, digital, and Heidelberg Anicolor offset press that are keyless.
FS: Image content should not require substantial increase in solid ink density. If the proof was made using proper color management and process control procedures, then the press can match it with normal, standardized inking. If out of gamut colors are required, then a different process is needed (spot colors, or another set of aim points and tolerances for SID's) which will have its custom output color profile that can be used to set up a proofing system.
BC: (a) GP’s statement is based on the old paradigm of ‘print to match customer’s implicit expectations.’ (b) Color management paradigm is based on the premise that what you see in a proof is what you get from a press run. Then, the color image adjustment decision should occur in the proofing/adjustment/approval cycle, and not in the pressroom.
FS: Highlights will be less affected by changes in ink film thickness, shadows will be more affected. That is why we do not use highlight dots in a color bar to control inking. All the operator can do is to try to keep one (dark) point on the tone scale as stable as possible by controlling dot gain and ink film thickness.
BC: I’m sure there are studies regarding the relationship between the degree of gray unbalance and their impact on color image reproduction. I would pay special attention to the issue of perceptibility, i.e., can I detect the difference between two gray patches by numbers or visually; and acceptability, i.e., can I accept the outcome when examining the reproduction on its own merit.
FS: Gray unbalance provides an early warning before the print is out of tolerance. It is desirable that it is more sensitive than the images.
BC: Gray balance patch is a quality assurance indicator and not a good process control indicator. The reason is simple – too many factors can lead to gray imbalance, e.g., differences in TVI of CMY and ink film thicknesses of CMY. So, when gray unbalance is detected, say, by excessive C*, one has to check the individual solid and tint patches of CMY in order to determine its root cause.
FS: We are not simply applying gray balance principles from prepress to press. (I do not really know what this would mean.) What we are doing is to recognize, that if one process ink is off the aim point, this is a more serious problem than when all three process inks are off the aim point in the same direction. In the second case, the gray balance is maintained while the overall image is a little darker or lighter, but this is still acceptable. In other words, a variation in L* only is less objectionable than a variation in a* only or b* only. Maintaining gray balance makes a lot of sense.
BC: GP’s comment is not clear regarding the meaning of gray balance between prepress and press. Today, wee see IDEAlliance promotes G7 as a method to adjust the gray balance of My_Press to that of a Standard_Press provided that the color gamut of the two printing processes is the same.
FS: The job of the press operator is primarily to set the inking so it conforms to the standard aim points, within the specified tolerances. Under these circumstances, the press sheet matches the proof, as long as the proof has been made with proper color management and process control procedures. We definitely do not want to do image editing on the press. If you do not know how to make a proof that represents the printing standard, then you get in trouble, with or without gray balance control. If you do not have a standardized printing and proofing system, a lot of sheets (not just color bars) will go into the recycling bin.
BC: Color control bar reflects how color printing behaves within a press run and from run to run. Color agreement between proof and press sheet is ensured via color management and process control. So, color control bar provides tangible evidence of printing process conformance in case there are disputes between print buyers and print suppliers. Printer should use color bar wisely.
By Raymond J. Prince on Sep 29, 2008
For a little interesting reading try -- Controlling Ink Color during the Press Run - Solid Ink or Gray Balance - TAGA Proceedings 2008.
As a side note - Bob Chung is right on target.
By Henk Gianotten on Sep 29, 2008
Erik mentioned GCR but all other contributors did not mention the important influence of GCR on press stability. I would like to get more info on that subject too.
Regards, Henk
By the way: System Brunner helps the printers to improve quality. But they don't only depend on densitometry. Several of their devices nowadays are based on spectral measurements.
By Erik Nikkanen on Sep 29, 2008
This discussion is very health. It allows some issues to be brought out into the light for consideration.
I don't think anyone is arguing that a gray patch can't be helpful on press to see shifts in colour. But the gray patch being gray does not mean that the whole image is printed properly. If one uses a gray patch just as some indicator of colour shift, that does not require a complicated calibration process. It mainly requires a gray patch that is a combination of CMY screens that prepress says should print gray at standard conditions.
But the idea that gray balance of that patch and a gray scale curve of some kind ensures colour reproduction is just not true.
The basic problem in the reproduction of coloured images is not gray balance but is the ability to reproduce the colour, from a specified area of the target image to the same specified area of the printed image.
Basically on conventional offset presses, the printing of CMY and also K is non linear and non independent. Mathematically, you can not solve that kind of problem with curves, be they NPDC or TVI curves.
One can not guarantee colour reproduction with NPDC or TVI but there is also a problem with reproducing colour with ICC profiles on conventional offset presses. The problem is that conventional offset presses do not have a specific profile that can be applied predictably for all images.
With gravure and flexo presses and with offset presses like Anicolor, Karat and Genius (not Cortina) one should be able to have a reasonable profile because basically there is consistency of ink application in any part of the image. With conventional offset, this is not the case.
With conventional offset it is well known that the colour patch (and gray patch) does not represent the actual inking conditions in the rest of the image that is in line with that patch. This is due to the random and cyclical variations of ink film on the form rollers that are a result of printing the image itself. In the worst case this can show up as a visible mechanical ghosting but there is always ghosting going on even if it is not obvious but it does affect colour.
There are also problems where the density is different at the top of the image than it is at the bottom where the colour bars might be. There is also starvation ghosting, which is not a image design issue but a roller train design issue.
Combine all these issues and it should be easy to see that you can not have a specific profile for these presses.
I am very glad to hear that ink film control is so important but I am still so surprised that there is no interest to take advantage of the solution to this problem.
There has always been this tendency to just brush over the numerous and very real problems and think they have minor influence. Just talking about an ideal situation that does not exist is misleading and delays understanding. This is one reason problems don't get solved. Lot's of so called minor problems add up to a big one. The lack of accurate colour reproduction and control on press.
By Erik Nikkanen on Sep 29, 2008
“In photomechanical reproduction, [ ] neutral reproduction of neutral grays is required”. J.A.C. Yule, 1967.
Yes required but not sufficient to ensure colour reproduction.
On your car you require that your tires are not flat. You also require that the gas tank is full and that the battery is charged. They are all required but individually they are not sufficient to get your car going. Just because your tires are not flat does not mean that you can drive if the tank is empty.
By Michael Eddington on Sep 30, 2008
"With gravure and flexo presses and with offset presses like Anicolor, Karat and Genius (not Cortina) one should be able to have a reasonable profile because basically there is consistency of ink application in any part of the image. "
Although I agree that offset has unique issues that Erik and Gordon have outlined, consistency of ink application in and of itself doesn't necessarily mean that a specific profile can be used representative of the entire category of presses. Take flexo for example. Currently there is no published data set or ICC profile for any portion of the flexo industry. Many who have tried, some participating in hundreds of press fingerprints, have indicated that no two profiles were alike. However, based on pressruns that I have seen using near neutral calibration and ink qualification, flexo can reproduce quite similarly towards SWOP/Gracol. The point I'm making is that qualifying presses, consumables and a consistent approach to press calibration can make a significant impact in quality and predictability, despite shortcomings of the process itself.
Not all proponents of gray balance and near neutral calibration (or any that I can think of) claim that G7 fixes all issues with offset press consistency, but rather can be used as part of an overall process control program to facilitate consistency and yes, reduce the overall color differences that may exist between an individual press and a published data set representing the process. I've seen many presses reduce the average delta E of their press toward the Gracol data set from 5-6 dE76 to 2-3 dE76 by simply using G7. Color management could likely take this value lower, to 1-2 delta E, but add in the more complex workflow, and admittedly, consistency issues with offset lithography, and one might not find it worth the extra effort.
Would a more consistent delivery of ink help? Certainly yes, but as I stated, even with consistent density, one still needs a target to aim toward, and a process to get there. Consistent ink delivery doesn't necessarily correct for different rheological properties of the inks or for paper inconsistencies. Meanwhile, those of us in the trenches need to work within the existing capabilities of our machines.
"Combine all these issues and it should be easy to see that you can not have a specific profile for these presses".
Yet, many are using the Gracol/SWOP data sets and profiles in tandem with a G7 calibrated press quite successfully, reducing make ready times and increasing predictability.
"There has always been this tendency to just brush over the numerous and very real problems and think they have minor influence."
These very real problems (and I have no doubt they are real) does not negate the benefit of having clear and unambiguous targets and procedures for press calibration. Take your flexo example...a process that can apply consistent ink application. Why then is the flexo process so variable and why are standards so far behind in this part of the industry? In my opinion this is due to a lack of a consistent approach to qualify the presses, consumables and process control techniques coupled with ambiguous specifications.
I certainly don't want to seem disagreeable with Erik or Gordon, as they make good points, and the industry should look at the root cause of consistency issues and address them at their source ...ideally before inkjet takes over the industry. ;) However, I whole heartedly support any initiative to improve Joe Litho's output by providing tools he can use now.
By Bob Chung on Sep 30, 2008
Gray Balance for Device calibration vs. CMY Neutral for Process Control
I received emails from Print CEO Blog readers wanting to know (a) the difference between gray balance-based device calibration and gray balance-based process control, (b) the difference between quality control (QC) and quality assurance (QA), and (c) why CMY neutral is not a useful QC metric, but good QA metric. I decided to pose my response here since these questions bear general interest.
Let me discuss the notion of gray balance as a device calibration first. Gray balance depends on how ink-paper-press behaves. By printing a gray balance chart in a repeatable CMYK device or querying the ICC profile of the device (or my_Press), certain CMY dot area combinations will render themselves neutral. We now have a gray balance graph of L* as a function of C, M, and Y dot area. If the gray balance curves are different from a reference gray balance condition (or std_Press), we can derive three transfer curves whereby the dot (in) and dot (out) for each of the C, M, and Y channel have the same L* value. We then apply these curves at the CTP stage, and print to my_Press. In other words, by altering the incoming data with the use of three one-D transfer curves, neutrals seen in the std_Press are also printed neutral in my_Press. The strategy to use gray balance for device calibration has been demonstrated by G7. Indeed, it requires that the same ink and paper are used; and it aligns neutrality between the two printing conditions well. But it does not address differences in two-color, three-color overprint areas, etc.
Quality control is about “doing the right thing” and quality assurance is about “Making sure that right things been done.” In process color printing, QC addresses CTP, selecting the correct combination of ink, paper, press, going through the press make-ready, and printing to specifications. Specifications can be density, TVI, color, or gray balance for that matter. QC also addresses the regulation of the press run by sampling, looking, measuring color bar on press sheets, and making press-related adjustments so that special-caused variations are corrected for and over-control of the press is avoided. On the other hand, quality assurance addresses sampling of the printed lot per MIL-STD-105D sampling plan, inspecting press sheets for possible print defects to make sure that defects are within tolerances. In this case, excessive color variation can be one of the defect definitions along with registration, ghosting, moiré, etc. So, quality assurance provides added confidence to print providers that quality products are printed. If print buyers are not as sophisticated and do not demand QA data (also known as the certificate of analysis), printers opt to store press sheet samples and do nothing until the payment had received. Today, if we perform process control by numbers, the resulting data not only are useful for real-time control, but also useful as documented evidences of process conformance.
The notion of using a 3C neutral as a process control patch addresses a totally different application than using gray balance for device calibration as discussed earlier. In this case, we assume that there are process drifts and ink-and-water balance issues that contribute to color variations. The question becomes, “Which color patch is sensitive to inking changes and is indicative of its root cause?” If we consider using a 3C neutral as the target, we need to know what to correct when the patch is not neutral. As mentioned in my previous posting, there are too many factors can lead to gray imbalance, e.g., differences in TVI of CMY and solid density of CMY. So, when gray imbalance is detected, say, by excessive C*, one has to check the individual solid and tint patches of CMY in order to determine its root cause. That’s why CMY (3C) neutral is not as useful as the individual solid and tint patches for process control. But, 3C neutral is a useful metric to indicate stability of the press run when multiple samples are measured and plotted as a a*b* scatter plot.
RIT conducted a case study, entitled "Color Control for Process Color Printing: Eyeball vs. SPC," in 1991 (some 17 years ago). We wanted to find out how good is visual assessment of a 3C neutral vs. measuring CMY solids in detecting and correcting (unannounced) process drifts. If interested, you can download the PDF at http://cias.rit.edu/~gravure/bob/personal.html#1991
By Michael Eddington on Oct 01, 2008
In response to the comment that gray balance is a myth, I'd suggest the following simple test.
1. Take a CMYK image into Photoshop, any CMYK image you'd like really, and place a patch of 50C, 40MY and a 50K patch some where in it. Assign the Gracol2006Coated1 profile to the image.
2. Now duplicate the image as many times as you'd like and assign different CMYK profiles to it. Offset, Gravure, Flexo, whatever you'd like. Notice the difference in how they are displayed on screen through the other profiles assigned.
3. Now, to simulate a G7 optimization, create a curve layer and edit the CMY curves until the LAB of the gray patch equals the LAB values of the image with the Gracol profile assigned to it. Now note how the visual differences have diminished, in some cases drastically.
Though this is a synthetic test, it does represent how balancing toward gray can improve the image as a whole, rather than just the gray patch itself. Its certainly not perfect, as in this case, we cannot qualify the paper and ink color up front, and yes, using profile conversion would obviously be superior when mapping color, but the point is this "myth" can be very effective at normalizing color casts, despite different processes or even ink hue to some degree. Balancing images toward gray is nothing new, and has been used reducing color casts in photography for years. I'd argue that applying gray balance principles in the pressroom is not "dangerous", but advantageous, even given consistency issues (and averaging would be helpful here).
If one is looking for numerical evidence of gray balancing benefits (delta E values), other than going to press, you can do a similar test with ColorThink Pro and Photoshop.
Taking an IT8.7/4 target, assigning a custom CMYK profile to it and extracting the Lab values (absolute rendering), you can compare delta E values for a given press profile toward Gracol2006Coated1. Now, open up the IT8 image in photoshop, assign a custom CMYK profile, add a 50C, 40MY patch and balance the Lab values with curves (per previous example). Extracting the Lab values in Colorthink Pro, you'll see numerically if there is a net benefit for gray balance adjustment on the given device.For example, using this technique to simulate G7 on an offset press of a client of ours, I can see that there's potentially an improvement from an average 11.52 dE76 (8.66 dE200) toward Gracol data before adjusting gray balance, to 4.77 dE76 (3.22 dE2000) when implementing near neutral calibration. Not too bad for 1 dimensional curves alone. I wouldn't expect this large of an improvement for processes using different ink hues, but for processes using ISO 2846 inks, there is often a net improvement.
These (admittedly unscientific) tests, and observing the actual press results, indicate to me that there indeed is a correlation between gray balance and the live area of the press sheet, although as ink hue changes, predictability diminishes.
By Erik Nikkanen on Oct 01, 2008
Bob Chung's post is excellent and gives a good view of gray balance and the use of gray target patches.
What I am not happy about is the description of process control, which is the traditional one used in the industry.
"QC also addresses the regulation of the press run by sampling, looking, measuring color bar on press sheets, and making press-related adjustments so that special-caused variations are corrected for and over-control of the press is avoided."
This is viewed as process control or even SPC. I would say that this is a poor form of process control and that ultimately process control and QA is more dependent on process capability. The fact that there is measuring and adjusting means that the process is not capable. I would also say that SPC is invalid for this process at this time because the variables are not controlled. An ink key setting is not directly related to the ink feed of that zone. It is a special-cause variation that needs to be addressed.
All the measuring and plotting of values is not SPC but is only a histogram of what happened. The control method being used is a crude closed loop with the operator closing the loop by using the feedback of the data to make adjustments.
The lithographic offset presses is not unstable but it is inconsistent. These terms are used incorrectly at times. The offset process is quite stable from a dynamics point of view. Stability has to do with how well a process wants to stay at an equilibrium point. Leave a press alone and it will run quite consistently. Adjust the water or speed or air movements and then the process (printing density) will move to a new equilibrium and sit there. Since the process sees a lot of these disturbances it is inconsistent.
To improve a process one would really like something that is better than just a stable process. One wants a consistent process and one with predictability. The goal is to remove the equilibrium nature of the process and make it a "fixed" process. A fixed process that is not affected by disturbances.
Having a positive ink feed fixes the process so that disturbances have very little affect on print density. Yes there are still variations but they are variation about a fixed point. Such a system would have much less need for continuous measurement looking for drift. It would be inherently consistent and that would also improve the consistency of the other related print issues. No ink water balance variation. Process control inherent in the design of the technology.
Michael commented:
"However, I whole heartedly support any initiative to improve Joe Litho’s output by providing tools he can use now."
I can understand the need to have solutions now. This tends to be the consistent view in the industry. But who is looking for future solutions?
I have not found anyone in the industry, in the 12 years I have been looking, who wants to solve problems for the future. Those people do not exist in this industry.
Even 12 years ago, it was possible to modify an existing press, at relatively low cost, that would allow it to get to preset density targets in about 50 impressions. It would then run consistently for its whole run. This would be for a sheetfed press or a web press. No one has been interested.
My point is that if the industry is to progress, then it has to re-look at what it is doing and take action to correct problems and methods. When one finds errors in technology or methods, that is where the opportunity is to get a competitive advantage. The fact that printers and press manufacturers have had a great difficulty in getting a competitive advantage is because they don't bother to investigate and take advantage of these opportunities.