Discussion

By Richard Romano on Sep 07, 2010

As part of this video series, we went to an Idea Alliance meeting at RIT and interviewed a number of color measurement and management experts, and one of the questions we asked was, why is color consistency still such a challenge for the printing industry, even after all these years? Perhaps it's a bit obvious in retrospect, but among the myriad reasons is that new papers, especially those designed for digital printing and the increasing use of "optical brighteners" that give designers and printers a nice white sheet make everyone making color measurement hardware and software--and standards--have to play catch up.

By Eddy Hagen (VIGC) on Sep 07, 2010

The basic reason for this new standard can not be stressed enough: different measurement devices will provide different figures.

It is extremely important that the industry (both printers and print buyers, but also the vendors) are aware of this problem. Many (most?) companies only have one instrument and this is considered as The Truth. But my Truth is not the same as yours...

Spectrophotometers need to be calibrated on a regular basis (and not only on a white tile/spot). They also need to be maintained by the manufacturar on a regular basis (how often do you send in your spectrophotometer to give it a check up?). This will cost you money, but wrong measurements can be much more expensive...

These deviatons between devices (even if they are the same brand, the same type!) have been studied by multiple organisations. We at VIGC made our study on this topic public just two years ago. You can check it here: http://www.graphicbrain.com/studies-by-vigc/spectrophotometer-nightmare/ These were tests in 'real life': the devices were not brand new, but were used in printing companies. Colors were measured as most people do: just one measurement (you can discuss, as some people did, whether this is the 'right' way to measure color, since making 3 measurements and average the result is more accurate, but be honest: how many printers do take 3 measurements and average it?).

If this new standard will help to obtain a better inter-instrument agreement, that's a good thing. But then also the other vendors of spectrophotometers should adopt it. X-rite is only one of the vendors. There are also others. And their instruments have the same issue.

By Erik Nikkanen on Sep 07, 2010

Why worry? Why not just ignore this source of lack of colour control as the industry does for so many other causes of lack of control and accuracy of colour reproduction?

Why a new standard? Is there not already a defined meaning of what a Lab value is?

As Eddy has said, the lack of instrument to instrument agreement has been known for many years in the printing industry and has been a little secret in the colour science field for much longer.

I remember reading posts on a colour science newsgroup a long time ago related to this issue. They knew that instrument agreement was a problem but what they started to discuss was that if you really wanted accurate colour measurements, you were not going to get them from commercial instruments. Then the discussion lead to the need to go to a specific instrument at a university, oh but that is not good enough. It ended that one had to go to some government institution where there was an instrument that cost maybe a half a million dollars in order to get accurate colour measurements.

The point here is that you are not going to get accurate colour measurements with the kind of instruments you now have. Calibrate them every day and that will still probably not be good enough since the instrument is not designed to be capable of getting accurate results. Good enough results but not accurate results.

If the instruments you have now been using were capable, then there would not have been any need to get them to measure the same. And just because two instruments measure the same, that does not mean they are measuring correctly.

IMO there is too much concern about colour accuracy. There should be more concern with colour consistency. Colour is not something that you can be very accurate with.

By Eddy Hagen (VIGC) on Sep 08, 2010

Erik: good points. But there is one thing you shouldn't forget: customers want some kind of 'objective' way to know whether a job has the quality they desire. And a stick to hit the printer when the quality is outside the tolerances they want (and therefore can ask for a discount or just reject the order and ask for a reprint). So color measurement is more than just process control. It's also about 'job validation' by the customer. And that's even a bigger issue (also due to lack of knowledge of many customers).

Judging a print job just by the looks is probably even worse (because not objective) than using instruments that don't give the same results...

By Greg Imhoff on Sep 08, 2010

I have two (2) points to contribute on the goal of improved communications, with a conclusion.

1) At a recent Print Properties Committee Meeting when I first heard of this plan asked and we were assured this would be a "open source" - a ISO based initiative open and available for all.
The goal was for all instrument makers to simplify the "compliance goal."

2) There always has been another phonemenon know as "Non Inter-Instrumet Linarity".
This means multiple varying instruments form differnt makers may agree to calibrate to a Density patch point ie: Cyan = 1.35.
However as true production related Density points diverge from this point (up or down), the units will not then agree. This is mainly due to individual instrument maker optical differences in their manufacturers pick up design.
These differences manifest as the instruments respond to thier own individual physical response curves, from the (eg 1.35) calibrated points. So as one moves up or down from a proscribed Density agreement point, the varying instrument units will not continue to agree, as they did at 1.35. All instrument makers see (pick up) Densities differently.

Conclusion:
If this is truly a ISO type - open source solution - one mathematically driven open to all instruments true production needs,then perhaps, this is a solution - but only if available to all instruments and only if simple to apply, by users in the field. Otherwise it is not.

By Erik Nikkanen on Sep 08, 2010

Eddy, I don't have any big disagreement with an effort to improve the apparent agreement of the instruments. I am always for some improvement in the practical use of technologies.

My point was actually more in line with what you have brought up which is related to the arguments that people have about colour. If people think that colour is something that can be measured exactly, then there will be trouble.

Let's put things into perspective of what the reality is. Probably all of this is something you already know. Colour does not exist in Nature. There is no colour in light and there is no colour in the universe. Colour is a perception which is only in our minds.

Colour science is more psychology than it is physics. Colour can not be measured directly. Colour science is very helpful and it is good to understand its limitations.because it is not an exact science.

Back in 1931 a group of people were tested on how they saw colour matches of pairs of targets. The data from that testing lead to the mathematical construction of the functions that were described as the Standard Observer.

Modern instruments are designed attempting to measure light and then calculate colour based on the functions of the Standard Observer. The Standard Observer is an average of how that small group in 1931 saw colour. The point here is that it is an average. It is not how everyone sees colour.

Different people can see colour differently. This brings up the known issue of Observer Metarmerism. One normal person can see two samples that visually match and another normal person can see that they don't match. If an instrument says that they match but a print buyer does not see a match, it can be a problem. I have read that the range of observer metamerism is about a delta E of 2.

Then there is the problem of the exact lighting conditions. One can say that not only instruments don't agree but I would bet that the lighting conditions in booths are not always equal or at the required specification.

Now think of the press. Allowable density variation is about +/- 0.05 pts. Presses have mechanical ghosting even though it is not directly visible and that is one reason Heidelberg thinks the Anicolor press is good because it tends not to have this issue. Also conventional presses don't ink the plate evenly from gripper to tail. With mechanical ghosting and this uneven inking of the plate, how can one have an accurate profile to provide to prepress. One can't.

Specifying a tight tolerance on colour measurements does not mean that consistency will be obtained. It just sets up for disagreements between printer and buyer.

Process predictability and consistency needs to be improved and can be. When that happens maybe buyers can back off on their demands that are not obtainable anyway. My guess is that buyers have gotten to the point of demanding something because they were not getting a reasonable result before.

Just because it is called colour science there is this expectation that it is exact. One has to think of it in the context of its capability to deliver results. Its a good guide but not an exact guide.

By Danny Rich on Sep 08, 2010

XRGA will be a major step forward but it is still only one step in a long road. The state-of-the-art in commercial colorimetry and color science is much more advanced than is described here. In the textile industry they no longer send proofs, only digital spectral data from retailer to garment maker to dyer and back. In automotive interiors they will pay you a premium if you use instrumental color control and not visually second guess the product in production.

X-Rite, through acquisition, has become an amalgam of instruments and technologies obtained over many years. Each technology and former company had a source of traceability and a chain of transfer standards. But those standards are not in agreement because there is no international agreement on the scale of diffuse reflectance with an uncertainty that is acceptable to the reproduction industry. Every instrument and every instrument maker has a bias just as VIGC has reported. Sometime the bias comes from the primary standard used with the instrument, sometimes it comes from the engineering compromises made to keep manufacturing costs down and sometimes it has come from ISO/CIE specifications being too loose so as not to offend any vendors. I gave a paper the CGIV Conference in Spain two years ago that demonstrates that if you assume errors on the order of those reported by VIGC at each stage of the reproduction process, finger printing the proofer, profiling the scanner, the press, qualifying the inks that the total uncertainty can be just as high as it was when we used only visual methods. Each step in the process can be reasonably precise, though not as good as we tend to think, but the leap from one step to the next can be unbelievable. ISO TC6 on papers have solved this problem to a great extent. They wrote a standard that essentially states, if you make measurements of the optical properties of paper you MUST use an instrument designed exactly like the one defined here and no other and you must buy your working standards from one of three authorized laboratories in the world and those standards must be replaced monthly.

What will it take get color in graphic reproduction under control? It will take dedication, lobbying of national research centers and money, none of which the graphic reproducition industry has shown any real commitment for in recent years.

By Eddy Hagen (VIGC) on Sep 08, 2010

Erik: maybe you've read my previous comment a little bit different than I intended it. :-) Instead of "you shouldn't forget", I better had said: "we shouldn't forget". This wasn't personal, I agreed with what you said before. It was just to bring a element into the discussion which is probably not that meaningfull for color scientist and color techies, but which is too often the hard reality for people who have to deal with print buyers.

To give a real life example of the lack of color knowledge: a few years ago we were contacted by one of our members. He had big issues with a really large customer. Always (= with every job) discussions about the color quality, meaning: always asking for discounts because the color wasn't within specifications. They were printing on brown corrugated. And the customer demanded a max delta E of 2. Everybody who has ever looked a bit closer at brown corrugated will know that this substrate is covered with dark spots... So the substrate itself had a deviation between different spots that was much, much higher than 2. As a reference they had a color sample (btw: it was a 1 color job), 'printed' with an ink roller on a bright, glossy paper... Because it was made with an ink roller, the ink/the color was absolutely not homogenuous. So which color, which density was the reference??? This was a large FMCG, with a dedicated print buying department... They ought to know about color, color measurement...

How imperfect color science may be, a print buyer wants something to quantify the color quality. He wants a number. And the lower it is, the better. Even if it doesn't make any sense. Even if he doesn't know what that number really means, what kind of difference in color perception it represents.

By Erik Nikkanen on Sep 08, 2010

Eddy, I didn't take anything you said as being personal. Dealing with colour is a real practical problem and having high expectations is a problem.

I am glad to hear from Danny Rich that the XRGA will be helpful. That's good.

But in my view, there are problems in the press which need to be corrected even if the colour science side is improved. Also it is only my opinion but I think the methods used now in prepress are not deterministic enough to get desired results.

Solving a problem can be difficult or easy depending on what method one uses. There is usually not just one method to solve a problem. And potential solutions do not have to be expensive. Poor methods make a problem look a lot more difficult than it has to be.

Much can be done but it won't be done. Things move much too slowly and there is no leadership. The leaders now might say they want improvements but they also have fought any suggestion in the past that things are not done correctly.

How can one win improvements when all the experts are more interested in protecting their expertise? IMO, this is more of a problem in the offset world than in the inkjet world.

Sorry, I digress. Getting back to colour. Colour is a tricky field.

By Michael Jahn on Sep 08, 2010

I agree with Erik - Colour is a very tricky field.

Few people know that there are quite a few different standard formulas being used to determine "color differences" when two samples are measured and compared.

You may have heard of CIE DE76, CIE DE94, CIE DELTA-E 2000, CMC(1,1) and DIN 99 - if not, you can visit this site;

http://www.colorwiki.com/wiki/Delta_E:_The_Color_Difference

I have had many discussions on this 'which standard' problem with several people far smarter than I, one being Dr. Ed Granger. As I am not a color scientist, I have nothing to say about that really, just that we use DeltaE 76 when setting up our Compose customers color proofing systems.

As an educational aid, IQColour (Alex Lelievre and Dr. Granger) designed a tool to test the 'differences' between a few of the popular formulas - you may find that this is a nice tool to visualize differences.

http://iqcolour.com/res_colorcompare.html

Years back I created an explanation of the tool which might help understand it.

http://docs.google.com/Doc?id=dfwf37zj_1077dfc3qxhq

Like Erik said, measuring color is tricky, I would like to add that it is pretty complicated as well.

As someone who works for a partner reseller of X-Rite products, we hope that XRGA this gets more traction than CxF.

By Jim Raffel on Oct 12, 2010

I fully support the concept of XRGA but like Mr. Imhoff agree that if the technology is not open source and easily available to all players in the color game then we should not be calling it a standard.

It's also very easy to track the stability of you own fleet of instruments with a Color Reference Material. One that is traceable and stable. Simply measure the reference monthly with each of you instruments and track the results. You'll know when the individual devices are drifting.