PrintColor

Densitometry as a means of SPC

2008-11-03T17:20:00.000-05:00

http://www.flexoglobal.com/flexomag/08-November/flexomag-toc.htm

Density & Dot Gain

2008-07-08T17:43:00.001-04:00

Please, submit any comments on the article in this post.

http://www.gravurexchange.com/gravurezine/0804-ezine/ploumidis.htm
http://www.flexoglobal.com/flexomag/08-July/flexomag-ploumidis.htm

Quality Control & The Role of Color Management

2008-05-04T21:07:00.002-04:00

Hello all,
I throught to provide this area in case you would like to comment on my articles published on GravurEzine and FlexoGlogal.
http://www.gravurexchange.com/gravurezine/0803-ezine/ploumidis.htm and/or
http://www.flexoglobal.com/pdfs/01-May%202008.pdf

I chose the topic of Quality Control as a means to emphasize the importance on color management and to provide an outline of the topics i'll be discussing.

Regards,
-D

Inkjet Proofing

2007-10-16T00:19:00.000-04:00

Hello again,

The most interesting area I've been involved during these months of absence from the blogging community was inkjet proofing. There are plenty of variables involved in this process.

The simplest and most reliable way to inkjet proofing is continuous tone proofs. I am not very interested in discussing continuous tone proofs, as there are plenty of software packages that are able to provide a reliable proof with certain pre-engineered color management algorithms. They certainly need their mastery and a critical visual evaluation, but the calculations, after approximately three iterations of optimization, can provide results that are really close to a visual target. In that, either EFI's rip or ORIS Color Tuner, provide very good results (i'm not all inclusive here... this is just what i tried). Then, certain adjustments for visual match give a very good proof.

Where do the complications of inkjet proofing start?

1. The white point of the target paper and absolute colorimetric intent.
2. Banding.
3. Halftone proofs
4. Hexachrome proofing
5. The color of black

To begin with:
1. WHITE POINT
Mapping the white point of the target paper through absolute colorimetric intent is not always perfect. Introducing the LAB value of the target paper into the proofing media alters the color of the proof. This is obvious especially in the highlight regions where the yellowness or blueness is introduced in the light areas. This problem is dealt with certain software packages where the user is able to affect only a small region of the color gamut. Relative intent works much better, but at cases where the target white points differs significantly from the proofing medium, there is a discrepancy that might be confusing to the press/customer. It is nevertheless astonishing that the profiling software is not able to map the white point with relative ease. I could make a wild guess and say that this is being caused due to chromatic adaptation, but it would be wild and uneducated.

2. BANDING
I've seen certain banding problems that can only be associated with the color algorithms of certain packages or with the ICC profile itself. I would guess again that the more you tweak a profile the more likely it is banding to occur. Then again, the way the software packages work, by trying to minimize the color difference from the profiling target to the media profile, some clustering of the target LAB values is not unlikely.

3. HALFTONE PROOFS
This is where the party starts. The rest is manageable. There isn't a software that is able to effectively color manage halftone proofs. EFI's rip offers a color management option with halftoning, but my experience is that the moire you get is unsaleable! Then again, you can separate the digital file to 1-bit TIFFs and try to work them around. How is color management applied to four channels that are a. already separated into 1010101 dots and b. translated into 7 or 8 eight channels based on the proofer? Only solution that we came up with was through editing the source profile. Results were decent, but unless my vision is all messed up it takes about 40 iterations, and then again... at some point... you just cannot get any closer to the target, since the above-mentioned clustering that would bring all the colors closer to the target LAB values is impossible through the naked eye. This might be a thesis topic for my valuable friend Arvind (http://cias.rit.edu/printmedia/03_programs/grads/karthikeyan.html) who's been helping me with proofing.

4. HEXACHROME
Ha! Forget about it, especially if you want halftones or spot colors as 7th or 8th channels. Then again, even halftone offset printing is kind of weird...

5. BLACK INKS
This is a colorant issue. Again, depending on the software, the color of the black might be unmanageable. I don't think that EFI has a good solution, as I was getting the cast of the inkjet proofers inks, and that was different from that of the ink supplier. I haven't seen such an issue yet with ORIS software.

In conclusion, inkjet proofing can have very good results in continuous tone proofs. My belied is that inkjet is the way of the future though, and that film based proofs with significantly lose their value, even if they provide higher stability and larger gamut (plus halftoning!!!) at this point of time. It is up to smart color scientists and engineers though to come up wth the algorithms that would eliminate all or some of the factors I discussed above, and -my belief again- that this is only a matter of time.

-D

PS> I would like to mention and thank my friend Rob Seckendorf of Color Correct Solutions, who has been providing his support and expertise with inkjet proofing (www.colorcorrectsolutions.com)

It is what it is

2007-04-10T15:34:00.000-04:00

When I visited TAGA I saw lots of concern about the G7 method of calibration. I only saw some results, and one paper saying that ISO actually outperformed G7.
Still, I was frustated by the general feeling of discontent about the G7 process. Sometimes it sounded like unreasonable, and sometimes I was hearing people that I overall trust to say that it doesn't work; yet I saw no valid research.

There, I made up my mind: I would try it out for myself and see if it would work or not. And I did.

I setup my roller stripes, drained my fountain solution and cleaned up the 64" beastie and there we were: ready to run with our hybrid ink set... shooting for grey.

To make a long sorry short, the results were disappointing. As with everything in litho, there can be numerous reasons.
1) idealliance software on shadows
2) measuring devices & idealliance software
3) natural press variation
4) ink set

I'm going to proceed from bottom to top:
4) We used the hybrid process ink set and the hybrid hexachrome ink set. On the color analysis tab we were getting the LAB values for paper grade #1 stock and the respective solid CMY. However, is the hybrid ink set supposed to be matching the NPDC curve, and whatever calculation thereafter? My speculation would be that the curves and calculation thereafter were done based on press runs with conventional ink sets. I might be lacking in theoretical understanding of the G7 calibration, but how can we use one set of curves for all the processes when there is a significant difference in dot gain between hybrid process, hybrid hexachrome (just the CMYK out of it), and conventional process inks? Should I be expecting to get my gray balance? I didn't... and this might be one reason. Even still, wet trapping our inks would allow us to be close to conventional TVI, and thus the software should work.
3) You know ;-)
2) I was getting significant differences in the suggested plate curves when measuring the targets with different measuring devices. Does the idealliance software work only with DTP70 and Eye-One? What prohibits it from working with x-rite 530 or spectroscan. Why in this earth do i get a difference of 20% on the yellow midtones between different devices? Praise the Lord! I guess the problem is that x-rite bought Gretag or WHAT HAVE YOU (i admit... i'm frustrated).
1) OFF.

SO... after such knowledge what forgiveness...

There are consultants who vouche that it works. I believe them.
There are other people who vouche that it doesn't work. I believe them too.

Thing is, at about 11:00 pm yesterday (and after missing a girl's birthday due to working late), I decided to get rid off all the software, match my densities, and calculate the curves on my own. I got my gray balance nice and neat! Am I G7 certified? I am 'Dimitri' certified. Why? I didn't use the software. Is grey balance tied up to G7? We were using the Brunner system before G7 and we were grey balanced already.

There is a long way to standardization... a long damn way and billions of variables to take into account. There is no magic; there is no beauty! Even if someone comes up with the perfect solution, nothing assures that we have the variables of the process under control. We are drifting and misjudging the effect as the cause... basic human fallacious reasoning... the dream within the dream... (i used to read Nietzsche instead of doing my homework at highschool).

What is the main problem? Calibrating something for certain conditions, or fixing these conditions? And I mean ink-water balance...

Here's my frustated plan of action:
1) trust noone
2) set up ISO densities and possibly TVI
3) get grey balance
4) train our operators
5) get device link software or other software with smart color management options.
6a) and MOST IMPORTANTLY: prioritize press maintenance
6b) and optimize support systems and supplies.

-D

PS1> Good luck to all... let me know if I am easily getting frustated...
PS2> Some people don't like feta cheese even though they have never tried it.

The Devil and the Virus

2007-04-05T23:52:00.000-04:00

We've got an old saying in Greece -maybe people have it in other countries as well- that says that there is a devil in every printshop, meaning that there are countless things that can go wrong.

Funny or not, I'm feeling that in order to be a printer you've got to have religious feelings: they are forced into us, and there is a certain amount of give-and-take with a higher deity when we are setting up a job. Are my prayers going to be answered, I ask myself time and again? is it going to work? is there a God above? What NOW? Is there a secret cord that David played to please the Lord?

Am I being sarcastic? Kind of!

Has anyone heard of having a malicious virus hitting the press? It CAN happen if your windows 2000 server is connected to a network. I guess, I could pose this question in reverse order, asking what is the least likely cause of press downtime. A damn virus! Presses connected to a network need Spyware and Virus protection I guess.

So be warned my fellow printers: don't put your pin in your press, and don't pay your bills online through the press console, and first and foremost, don't visit adult content sites between the scans.

-D

And one more thing...

2007-03-25T22:51:00.000-04:00

... now that I think of it directed to TAGA administration peoples.

Please, publish the TAGA Proceedings online!!! Please!

This wealth of information is going to waste in this digital age and it is a shame not to utilize the technologies available to help researchers further their work and professionals get answers. I became a member hoping to be able to access these issues and I cannot (unless I am doing something wrong). Not even the abstracts!!!

I'm not strong with economics and market prices, but you could sell each article for like $12-15 to non members, and $8-10 to members... or something similar. IS&T has a model website for it, where we can review the abstract and go on and purchase the article.

I would love to have the TAGA proceedings online. Please do something about it. I volunteer to help.

-D

Summing up TAGA 2007

2007-03-25T01:35:00.000-04:00

Well... I had a nice time myself. The hotel bar was pretty good and I hit Pittsburgh on Saint Patrick's day!

Seriously now:
I would love to see more attendants next year in SF. We were about 170 people. We can be more. There are lots to learn. Technical people can have a serious benefit in attending this meeting: Networking! Not so much in terms of making friends, but in terms of meeting the people that can provide an answer to their questions when the need arises. As Mr. B. Blom put it, throughout his years at TAGA that was the main benefit: the people! Well, you got to be a printing nerd too to enjoy it.

The students were an enthusiastic crowd as well (getting lost in conversations about density at the bar is absolutely awesome). Compliments on Cal Poly on winning the brochure award - of which I am particularly proud since my company happened to design and produce their packages; which I only became aware at the conference, mainly because I am knee deep in our litho department at work.

Top topics of what I have retained:
#1 There is a lot of research going on on RFID. I hope it works out.
#2 There is another definite tendency to move to colorimetry for process calibration, and why not control. It shall need a huge effort on technical and managerial level, but it can work and provide solutions were density fails short (workflow solutions and color portability, as well as spot colors I would say on top of my head).
#3 I retain the research on the low VOCs. It is important to be prepared for it... if possible!
#4 The discussions and presentations on colorimetry issues and ICC profiling (including mine) however -I have to say- are falling short of the research that is going on in imaging and color conferences. We cannot expect the printing industry to lead in this field, but we can definitely expect to publish some research on the applied elements of ICC profiling and colorimetry for the printing processes.
#4a. we need to work on color difference equations
#4b. we need to establish a linear relationship between IFT, density, dot gain and CIELAB.
#5 Paper companies were criticised -and forgiven nonetheless since it is realized that they are bound to price demands- for UV brighteners, standardization of white point, and to a lesser extend paper properties that affect printability for various processes.
#6 G7... well... read previous post!
#7 Process variation issues discussed and researched and found awful. Well: welcome to litho... and as professor Chung would say: variation is the 'nature of the beast'. Pretty much: 0.10 density units is something very common as natural variation, and it results in about 5.0 DEab... reminding me of William Burroughs novels, where drunk and angry press-boys wonder on saliva and the loose limbs of grey crocodiles and pink elephants (huh?!?)

Farewell,
-D

G7 on TAGA

2007-03-25T01:19:00.000-04:00

There was lots of discussion regarding G7 on this years TAGA conference. There was a session devoted on it. Everyone could chip in a comment. Most of the comments were against it. Why? There was only one published paper on it. Apart from this, there was no published research to substantiate the unfavorable comments. The research community should be the first to know that we need to back up our claims with research.

Being on the field right now, I NEED to be able to read published results regarding any new technology that can help me be more productive, and I need it from this community. Going into endless discussions about how and why it doesn't work is totally worthless, if not damaging.

The main argument against G7 is that it doesn't take into account the color of the inks. Profiling does. Even more, device link profiling is able to calibrate the presses and in the same time take into account the colorants. I buy that. I have tested device link profiles and I know that they are able to do an amazing job. Within the next month, I will test out G7 as well and see for myself if it works.

I have to comment on Dave McDowell's efforts (who was awarded for his immense contribution to the industry) to close the gap between the group that is for and the group that is against G7. Dave, and the particular ISO committee are trying to put G7, ISO 12647 and device link profiles into agreement taking the strenghts of each method into account. I would love to see the ISO document and I hope that it comes out soon enough.

Characterization of Conductive Polymer Inks Based on PEDOT: PSS

2007-03-20T16:19:00.000-04:00

by Erika Hrehorova, WMU

Erika started her presentation commenting on the need to continue the research on printed electronics, stating that there is a future along the line.

The goal of printed electronics is to create structures and devices that are similar in functionality with conventional electronics, but at greater speed, lower cost and less production complexity.

Ink conductivity is achieved by mettalic or other conductive particles incorporated on a non-conductible polymer matrices, or by using intrinsically conductive polymers in suitable solvents.

PEDOT:PSS based polymer inks are water soluble polyelectrolyte systems with good film-forming properties, high visible light transmittance, and excellent stability and it commercially available as aqueous dispersions. The main problem is substrate wetting and ink spreading. In this work it was modified with addition of alcohols and non-ionic sirfactants.

It was found that ethylene glycol decreases the surface tension. Surfactants and alcoholic co-solvents were used as additives to further reduce the surface tension of water based systems. It was found that ethanol was more effective in the reduction of the surface tension, as well as in the improvementof the surface wetting under both dynamic and static conditions. Addition of ethylene glycol increases the conductivity and improves the uniformity of the polymer layer.

losing momentum towards the end

2007-03-20T16:11:00.000-04:00

So... someplace here the 'live' blogging sort of ends. I think I lost momentum today... missing a number of speeches, plus having to present my own paper.

I will make a couple of more summarizing posts within the next few days. I know that what I wrote was incomplete and most likely focuses only on the results... but that was the best I could do. To get more information: become a member... it's not that expensive and there is a wealth of knowledge available. Next year it's in SF.... providing an opportunity for me to get out of the Valley and family-oriented Modesto, and for those willing to come, to see SF. It's really a great city!

-D

Analysis of Color Rendering in Digital Cameras

2007-03-20T12:25:00.000-04:00

by Jason Lisi, Ryerson

Lack of standardization of raw data. Only the camera vendor knows what the camera is doing, and companies like adobe have to reverse engineer the images to find out what is going on.

so the path is: CCD Data -> Raw Data -> RGB rendered data

The question is: can this color rendering step be isolated, mapped and predicted?
They used 6 rendering s/w, 5 raw format and one JPEG and assigned the sRGB profile.

Neutral (as well as quandrant) rendering is good from different vendors, but color matching overall is not as good.

Wiggles

2007-03-20T12:01:00.000-04:00

by Franz Sigg, RIT

This is an experiment Franz run across while I was still in RIT. Apparently, what Franz found out by chance in a first press run intrigued him, and he repeated the test focusing only on the 'wiggles'. The days at RIT were really nice.... when I was working late at the labs, every other month would come to me, all excited at 10 or 11 in the night, needing to communicate what he found here or there, every time about something new and interesting.

in this experiement, he found large deviation from the target value of a large printed target across a target with small incremental value increases. He assumed it was press nonuniformity of the inking, or the press, I guess mostly on the circumefential direction of the printing cylinder, and even more specifically at the lead edge of the sheet. He suggested to use two randomized targets and average them.

And a final remark was that the tolerances for printing are very hard too keep.

Color Difference Equations and the Human Eye

2007-03-20T11:38:00.000-04:00

by Martin Habekost, Ryerson University

Discussion on the color difference equations. DEcmc introduced the weighting factors and was developed for the textile industry. It is used widely for color matching in the ink manufacturing industry.

Second revision was on 1994, again with weighting factors that are adjustable as well. It is even more realistic than cmc, but it has weaknesses in the blue and violet region.

So CIE made DE2000, which is considered to be the best and is recommended. It is not used in the industry though. What I have heard is that it moves the color space around and it creates a moving color target.

Martin suggests after the experiment that DE2000 or DEcmc are to be used, and preferably DEcmc because the data were more coherent.

Q&A
Important considerations were raised during the Q&A section. First, Dave McDowell focused on the need to establish the weighting factors that are more fitting for our processes. After all, the developers of these equations reccomend each user to develop the factors that suit their processes best. Why don't we do so? One might say: lack of awareness or knowledge, or time to test out what is more fitting. I've seen people using a DEcmc(2:2) as QC. You cannot imagine what comes through as OK with these factors. I have seen some others using factors that narrow down the variation of hue, but allow variation in chroma and lightness. This should be very useful for the ink making process, where hue is the most critical parameter. However, the variations during the printing process would be on IFT, and if C* (being sensitive to IFT variations) has a wide tolerance, it is likely to be very forgiving... more than what a customer can forgive.

Densitometric and Planimetric Measurement Techniques for Newspaper Printing

2007-03-20T11:21:00.000-04:00

by Maria Wroldsen, Peter Nussbaum, Jon Hardeberg

an investigation of the statistical relationship between densi and planimetric measurements on newsprint. Is it possible to convert planimetric to densitometric and vice versa?

They used regression analysis with 2nd order polynomials. Each color was studied individually. They used a training set to establish the model and a test set to evaluate its performance. They did all the repeatability tests as well.

The model is not good enough for prediction, since it exceeds 2% tolerance that is the target for dot area meters.

It's interesting that they did this experiement because of the increased use of measurements in Norway

Method for Calibration of a Printing System with digital Data using Near-Neutral Scales

2007-03-19T17:57:00.000-04:00

by Dave McDowell, Consultant

There is a change in the game, cause more than 95% of content data is exchanged digitally and there are many options to manipulate the data... BUT no standard method to do so.

Dave created a model workflow: we need to determine a stable operating position AND match both aim targets and overprints. The system should be adjusted for in-gamut colors.

so how can this be done: adjust the process parameters or adjust the data.

Platesetter curves, rip curves, TVI curves, NPDC curves, device link profiles...

At ISO TS 10128 calibration standard meeting the committee focused on agreeing on the last three. One point that Dave makes thepoint that in effect there are no densities that are defined outside a CIELAB color space that is not based on some other characterization data. There are no densities as standard targets that are device specific. As I quickly understand it, and I might be wrong, this TS will help narrow down the specs by standardizing the calibration and using the strenghts of these approaches.

A Theoretical and Experimental Comparison of GRACoL and ISO approaches to Press Characterization and the Data Set Produced.

2007-03-19T17:32:00.000-04:00

by Greg Radencic, PIA/GATF

There is admiteddly great discussion for different color controls on press. This paper attempts an objective evaluation.

ISO 12647-2 defines colorimetric values for CMYK solids and densitometric control of tone reproduction. Grey balance is a secondary control in ISO.
G7 is a calibration method for proofing and printing and uses colorimetry through grey balance measurements.

When they run the G7: the calibration software doesn't perform a good calibration to grey below 30% and above 60% dot areas. G7 has admitted the problem and they said they will fix it. (I jut bought the s/w by the way... a bit disturbing to know that I bought sthg that might be wrong!?!)
When they run the ISO: it was succesful with nice grey.

The Effect of Certain Process Parameters On Inherent Color Variations on Press

2007-03-19T17:15:00.000-04:00

by John MacPhee, Baldwin Technology Company Inc.

John asks the question what is the expected upper bound of inherent type density variations for the lithographic process and provides an answer.
Then he suggests to reduce the density variation by changes in substrate, screening, use of waterless process and use of gravure inks. His study showed that none of these 4 parameters made any difference.

The main cause of variation is random IFT variation due to transfer.
There are three types of density variation: a. factor or color (Y being the smallest), b. screen area with picks of shadows, and c. location in the sheet.
The answer to the question what causes the variation is that it is variations in IFT or slurring.

The Benefits of Concentric Screening for Offset Lithography

2007-03-19T16:47:00.000-04:00

by Mark Samworth, Artwork Systems

A concentric dot is a regular AM dot divided into rings.

Mark describes differences between AM, stochastic and concentric screening. Well... concentric has the best of both worlds... Organic (2nd generation stochastic by Artwork systems) has better registration, low visibility, fine detail, press stability, color saturation, ink savings..... disagrees however that has better registration. Mark claims it is actually worse than AM screening. Main difference is the savings in ink, because of the rings that decrease the amount of ink film that builds up on the surface.
The benefits of stochastic screening is the smaller dot size. And concentric screening has a smaller size.
The benefit of AM screening is that uniform patterns appear smoother visually.
So... concentric screening combines both.

Plus, concentric screening has higher chroma and is adjustable on press.

Effects of Rendering Intents in Color Management

2007-03-19T16:15:00.000-04:00

by Anthony Stanton, CMU and Mark Bohan, PIA/GATF

Focus is to compare the color renderings associated with the applications of different rendering intents in a color managed workflow.

4 vendors / 4 rend intents +1 with no color mgmt / 5 samples

Suggest the use of DeltaE2000 as a better, more accurate metric. The older DeltaEab overestimates color differences, and depends on hue, saturation and color levels (non uniform in CIELAB color space).

The study was mostly descriptive.

How many DeltaEs are there in a DeltaD?

2007-03-19T15:26:00.000-04:00

by John Seymour, QuadTech

Kind of attacks the ISO 12647-2 standard. There are three reqs:
The first req is that we make a proof and we have to match the target LAB values. Then, the second req is that the OK sheet must meet the LAB values of the proof. Then the press run is the same, having to match the LABs of the OK sheet.

John's thesis is that it is possible to maintain a given DeltaD throughout a press run and n this way maintain DeltaE.

L* and density have an almost linear relationship. You can predict L* based on density when paper is a given.
Another conclusion is that we cannot use density to compare proof and press run, mainly because of the impurity of the inks (plus different pigmentation).

The Impact of Changing Fountain Solution Chemistry and Press Consumables on the Runnability of a Web Offset Press

2007-03-19T15:06:00.000-04:00

by Dr. Mark Bohan, PIA/GATF

Need for standardized fountain solution with types of paper, inks, presses and printers (well.. of course)

The project evaluated impact of process variables on a web press when fountain solution chemistry changes.
They used a traditional NA fountain solution and made variations of it.

Well.... there were 42 press trials 'successfully' completed. The results are pretty much broad and hard to sum up here.
One I retained was that stochastic screening takes more time to stabilize. Water levels change throughout the press run, but do so rather consistently. The chemistry of fountain solution affects the performance of the press run and can be optimized per process.

The Printing Problems Created by the Use of Reduced VOC Washup Solvents and Their Effect on Press Conductivity

2007-03-19T14:45:00.000-04:00

by John MacPhee, PIASC Solvent Task Force

Regulations cause south CA printers (all of CA printers actually) to use solvents that contain low VOCs (no more than 100 grams/liter).
What are the problems:

1. Longer makeready times following a roller washup. Mostly due to print defects. The problems didn't show up immediately. Solvent accumulates on the rollers, then it would be released and dilute the ink. Terrible effect. Ink tack is dramatically reduced.
REMEDY: Follow normal washup with an aqueous fluid (???). But it takes more time to washup the rollers.
2. Some low VOC solvents become bonded to the rubber when absorbed, causing the rubber in the long term to swell and become harder. Swelling is not uniform.
3. Some older washup systems apply more solvents than needed, which results in drips of fugitive solvent that can cause print defects and/or produce a safety hazard. But with high VOC this evaporates. with low VOC the amount of 'fugitice solvent' it collects someplace and creates problems. The solution is a third generation of systems that we got the solvent applied on a pre-soaked cloth.

Conclusions
1. When less than 100 gram/liter the problems created are very real and involve productivity deficiencies and also equipment and material costs.
2. CA will have less productivity and will be less competitive.
3. It should be stopped before it goes all over the US.

Honestly, doing some research on my own... it is a horrible thing and I have to find a way to solve it for my company.

The Effect of Ink Film Thickness Variation on Color Control in the Circumferential Printing Cylinder Direction of Offset Presses

2007-03-19T14:15:00.000-04:00

by Manfred H. Breede, Ryerson University

'Imperfections of the ink system design' as an opening statement. It goes straight back to our discussions with Eric over previous posts of this blog.

Gravure has direct transfer... no ghosting.. less variation.... very uniform amount of ink in x and y directions.
Flexo is consistent as well, due to replenishment of anilox rollers.
Screen printing is extremely uniform as well.
In digital printing the digital file is electronically linked with the imaging device.

And then.... welcome to litho:

Large roller train creates variation. The problem lies in the ink free zone of the cylinder gap. Moreover, the form rollers need several revolutions in the circumferential direction, and it is not easy to deliver the same amount in every revolution. In addition, the ductor roller oscillates and disperses a very think and unequal ink film. Finally, disruptive factors also include the image area configuration.

They printed a test form (0-100 tints) four times, starting with G7 standard density values and decreasing them every press run. they compared it with electrophotography.

There is density variation is about 0.10 points for all the process colors. The DE measurements show that the less ink resulted in more color variation (up to .15), but when the standard was run at the standard densities it was pretty much within 5 DE.

Manfred claimed that this was due to problems of the colorimetric system.

Q&A
First the DI press was a bad press.
The test form might cause ink starvation.
These were serious objections to the results of the test.

Liberty Inks

2007-03-19T12:15:00.000-04:00

by Steve Miller, Sun Chemical

They require less water... I can say that they have a smaller ink/water balance tolerance. so true... Steve also mentions that they need to change the mentality of the printer.
Drainage to the surface of the substrate varies a lot due to the substrate properties.

The resin system requires less water... if you make it right it is more stable.

ROLAND InlineFoiler Prindor

2007-03-19T12:01:00.000-04:00

by Hal Stratton, MAN Roland

The ability to foil in line and print on top of it.

1. varnish type of glue applied prior to the application by an extra unit
2. blanket
3. foil is applied to the glue

Like it too. Sheet fed: ROCK ON!

Multiple reel allows different configurations (less waste too I assume) and different colors or types of foils

Q&A
They are looking for RFID. I'm tired of listening about Walmart. Higher gold conductivity is a solution as Henry Freedman says. There are arguments about of the costs of RFID... not again the walmart story....

Genius 52 UV

2007-03-19T11:48:00.000-04:00

by Chris Travis, KBA

Small press with nice design. Waterless UV, first time I saw it it was in Drupa.

It uses a single impression cylinder with up to 5 colors. It uses a chambered doctor blade duct, an anilox roller, then an ink forme roller, the plate and last the blanket cylinder (I like it!!!). Then you get the usual huge KBA impression cylinder, common to all the printing units.

The wide variety of substrates it prints on is impressive. They also claim that you get makeready in 10 sheets, since you get away from ink water balance (no water, no ink keys, plus the anilox roller should give a better control of the flow of the ink).

As I said, I like it.

Q&A
How do you take the ink out of the anilox roller. (Is there a catch?) They have a temperature control unit that practically adjusts viscosity of the ink (different kind of inks too) and in effect controls density (within a narrow window... i guess there was a catch there!) So the anilox roller is the key. I would love to know the window. For sure, taking away the ink keys is huge, but it decreases the control of the pressman over the process.... so... if things are not going PERFECT on the press... you have no room to 'save' the job on press.

No ghosting because all rollers are the same size... clever!

Zero Slip Nips

2007-03-19T11:34:00.000-04:00

by Dan Perdue, GOSS INTERNATIONAL

We are talking about folder nips "transferring multiple webs more reliably, consistently and professionally". The difference is on different mechanical properties of the new folder nips.
I guess better print-to-cut registration and decrease in makeready time, less maintenance and splice sensitivity are the advs. Less waste too.

Technology Awards

2007-03-19T11:25:00.000-04:00

#1 Adobe PDF Print Engine
Touchless workflow.... minimizing 'touch' points => maximizing productivity
The less pdf to postscript conversions the less the issues/ problems.

#2 Microsystems Microcip Software
Facilitation of automation with CIP3 data

#3 DALiM MiSTRAL
Web based ... lost it... too fast

#4 WebCenter
3D modelling for packaging, plus communication tool between depts for complicated workflows

#5 Mx6 Magnum (GFI innovations)
Ink mixing equipment for higher accuracy

#6 DeskDirect
Hybrid mail solution

#7 XMPie ulmage
Fancy personalization technologies in Adobe CS2. Self programming

TAGA 2007 in Pittsburgh, PA

2007-03-19T11:16:00.000-04:00

In case you haven't figured out, I'm at TAGA, taking notes of the various presentations. I guess I should have started right away with an intro on what I'm doing, but I preferred to cover the first topic rather than write the intro.
Now it's the first break, so I got time to do that. I'll be editing typos and thoughts later tonight or during the week.
So... this is going to be on note style. It's not going to be complete coverage of the topics, just what captures my attention. But it's going to be live and I hope that you'll find it interesting.

-D

The Joy of Coating - A review of the latest trends in coatings, varnishes, and primers used within the Graphic Arts Industry

2007-03-19T10:20:00.000-04:00

by Graham Battersby (Business Consultant)

Funny guy... starts with a joke.
So gloss is an important effect for packaging, labels, publication covers.
He goes on.. gloss is just an aesthetic effect... but there are benefits of functionality, increasing the COF for example, or abrasion resistance... very true.
Comment about high level of interest in special effects that is not matched by sales volume.

Differential (matte, high gloss) gloss gains popularity.
Reticulating coatings are fancy. He shows a video that gives a 3D view to the water drops on a printed apple.
Low or no odor coatings are mentioned as well.
Coating units at the beginning of the press is a new idea I've heard yesterday and today.

He makes a good point that in order to increase the sales volume we have to discuss the design of the package before going into the technical parts. That would definitely increase the sales, and that's how it should be done, but only with collaboration with the printer... some of these printing effects have a high risk of failure when put on press.

Sustainable packaging is brought to caution, that we all must consider for coatings as well.

The Q&A ends up where i expected. These coatings are tough to use. I would say we (as printers) need support from the suppliers for specialty coatings such as these. Good communication is a must when you deal with such specialty products that have special needs.

Printable Foils - New options for packaging designs

2007-03-19T09:51:00.000-04:00

Presentation by Ray Verderberger, Eckart America.

Ray starts with an outline of the manufacturing of printable foils and the fact that they are useful for creating nice effects.
What interests me is that they are using the term brilliance to describe the metallic gloss. The bigger the particle (flexo) the higher the brilliance due to less light scattering from the edges of the particle, the less the opacity. Conventional metallic pigments for offset are exactly the opposite, having the constrain of the ink film thickness that goes through the press. However, the VMP pigments, in comparison to conventional pigments, are even finer but, can be applied in a high concentration and moreover be aligned... in this way they can achieve both high gloss and good opacity. They work excellent with flexo and gravure, but not as good with offset printing processes. The smoothness of the substrate is another factor (the smoother the better). An interesting point is made that since printable foils are applied as inks, they have much less waste than hot foil stamping, where you throw away the foil that doesn't remain on the product. I guess from an economical standpoint, they are a sub mostly for foils and not conventional inks. I might be wrong.
Another term that Eckart is coming up with is leafing, referring to where the metallic pigments fall within the ink layer.

Density Standards

2007-03-03T15:49:00.000-05:00

What is the best status density for controlling a process? First of all, what do we mean by 'best'? I would say that for process control purposes, 'best' should be the status density that allows the most sensitive response to ink film thickness (IFT) variations. Depending on the application (metallic colors, fluorescent colors, process colors, wet or dry readings) this might vary. With that in mind, we should be looking for the density response that amplifies the particular wavelenghts that correspond to the spectral reflectance of a particular color. Le'ts focus only on process colors.

Examining status T and status I densities, it was obivous that status I densities are far more sensitive to unveil variations in IFT than status T. Status T densities were developed in the film era to make the separations. Thus, they needed to be broadband, and capture a good deal of the wavelengths involved. For process control however, the spectral reflectances of process inks are pretty much known, and thus this 'spread out' is not sensitive enough to amplify the response and allow the user to see IFT variations, at least not as much as status I does, where we have a much higher peak. Status E densities are almost the same with status T densities, with the exception of a higher sensitivity of the yellow filter.

Another issue is polarization. Polarization was developed in order to be able to make the dry densities look like wet, by taking out the diffuse reflectance of the light from the substrate. When the ink film is wet, we have much less diffuse reflections, and higher gloss... the beam of light that returns to the densitometric sensor is more focused. The problem with polarization is that it is complicated in its calibration. It's not always convenient to have a wet sample as a standard.

Why am I going here? US is on status T, Europe is on status E. We are talking about global process control standards (ISO, GRACoL7 to name a few)... but we have different measuring practices. In that sense, having the 'best' status density would mean to be using the status density that adheres to the regional standard. In this way we would be able to communicate efficiently with our suppliers and customers.

The situation gets a little more complicated when we are dealing with press manufacturers. A lot of them are in Europe, yet they are selling their presses in the US. Do they, or their subcontractors, adhere to the US standards? Should they? In order to answer this question (without wanting to go to the obvious, long-term solution of finally setting global standards) they should ask and answer the question what is best for their customers.

If I was the President (as we used to say in my village), I would setup either status I, or status E densities for all. I am a UV printer, so I do not care that much about polarization -easy way our, right ;) But we are status T in the US, and I am not, neither want to, be the President. So, that necesitates me having status T, unpolarized, even if it is worse that status E or status I.

To be more particular, and after having briefly talked about the different technologies, in my company we have online press controls that are status E, polarized, and they are pure old densitometers (no spectral densities here). The reason we do not have spectral readings converted to densities is that we have to be able to take fast readings of the color bar for process control. A spectral reading, sampling the entire range of wavelenghts at 10 or 20 nm would be slow. Our software (which is in every other aspect absolutely amazing and the best I've seen in the industry, check out www.lithec.de) offers us the opportunity to calibrate our polarized density readings to the spectral density unpolarized readings of our handheld instruments (500 x-rite series). Confusion builds up.

Overall, it works just fine, but when we have to move to a different range of densities, it fails... especially for the yellow and black filter readings. The easy solution is to recalibrate not from a given standard, but from the printed sheet itself. In this way it would be somewhat wet (to take care of the unpolarized-polarized confusion) and in the right range of densities. What do I mean by the right range of densities? Without having done any research on the topic, I would think that since density is a non-linear response, when you calibrate to a 1.30 dY, and then you print at 0.90 dY, the conversion factor in the software does not simulate the nonlinearities of the varying IFT. I guess there wouldn't be a nonlinear conversion factor for it either. The difference between spectral and filter density makes it even tougher.

(Standardizing the readings of the spectrodensitometers is an entire new chapter... even less probable to be solved. If I remember correctly, the falacy of color theory is in lacking a standard white point reference... I shouldn't be stating this without opening my books cause I do not remember exactly, but I think that that's it).

Anyway, explaining all these to our manufacturers' we were able to get them working on some status T, unpolarized density filters. Will they work? I hope so... but I do not know so. Would we be able to adhere to the US standards better, with the downside of having less sensitive process control? Might be.

We have to take our risks, do our research, and find out... otherwise we would still be using our eyeballs for process control...

-D

PS> Don't get me wrong, first and last for process control is the nice, old uncalibrated eyeball and these undetermined pathways to the V1 and V4 cortex areas and then to perception! We just need systems to support all these!

Art, craft, or science?

2007-02-28T01:31:00.000-05:00

So what's the business we are in? Art, craft, or science?

I was talking to one of our most capable pressmen the other day: he considers himself working in the field of Graphic Arts. He implied that he is an artist, realizing in the same time that coming across with such a bold statement (himself an artist) would be a weak argument. So he held back, just making a comment of himself as being in the 'Graphic Arts'. Same argument I got today from our ink manager. He thinks he works in arts.

Every day, I wake up (rather late I admit), make some coffee, check my emails, and drive my car to work, organizing my thoughts on the way of what I have to do for the day. Approaching our establishment, I see a gray building erecting where it used to be dairies and fields. And I realize: I work in a damn factory.

So what's wrong with these people? Don't they see the same building I do? Don't they see that there is software counting the number of impressions, their downtime, the color coordinates, the temperature of the fountain solution? Don't they see that 8-color beastie that needs greasing every other week?

I was more of an artistic kinf of person back in Greece. I also liked to philosophize about stuff in general. One of the unresolved issues in philoshophy is 'what is art'. Postmodernism theory states that art can be anything. I might be a work of art sitting here right now and typing in these words. Aristotle would stand up and say that art is a form of representation. I like to believe that art is the creation of something nonexistant that we conjure out of our inner thoughts, emotions, or what have you. Unless our pressmen, managers and ink technicians are postmodernists, there is no definition of art that they fit into it.

Are these people disillusioned? I have to go back at work at 11:30 tonight to establish the MEK tolerances for our new UV nonskid coating. Will these people fail to print if I don't? Will the coating come loose after the Asitrade? I bet you that it won't. But what if one of our delivery lamps is off and the energy density falls beyond 60 mJ/cm^2? What if while they are testing it they hit the black at 2.10 density?

I am not a scientist. I would be a scientist if I was in some labs doing research religiously to find the linearities between changes in IFT and C*. I consider myself a technologist or engineer. I apply science in my work. I need concrete data to implement the procedures and technologies needed to quarantee that things happen.

Likewise... these people are not artists. They take 2 to 8 plates every day that carry an image and try to match color.

To match color though... to handle the 8 headed hydra... they need to be craftsmen. They need to check the fountain solution for the particular ink, check their roller settings, their IR, their ink and water balance... and it goes on and on. It's tough. There are times that all the rules that science sets up for them cease to matter. There are times that DE formulas fail to predict color perception. The craftsman has to understand and consider all the variables involved and make a call with the customer waiting in the lounge for the press check.

To conclude: the printing process involves variables that with the current technologies are not adequately measured and when this point is reached, when science fails to provide and answer, craftmanship is needed to make things right.

PS> My old man back in Greece uses his eyeball as a densitometer... and he is succesfull... most of the times

-D

UV curing

2006-09-21T14:25:00.000-04:00

Curing of UV inks, hybrid inks and coatings is an issue that hasn't yet received a good solution. There is no objective way to determine whether inks or coatings have cured. The industry relies on several methods that are either empirical or indirect.

Empirical methods
MEK test examine the resistance of a coating to chemicals. The user has to rub the coating with MEK and note down the number of rubs that it takes for the coating to break down. Results may vary depending on the user however. Moreover, only resistance to chemicals is tested, and not actual curing. Furthermore, when a printed product has to go through the entire converting process, the chemical properties might change and the printer might end up with unexpected results.
The reticulating pen techique involves making a stroke with a standardized pen and observing how the ink behaves on top of the coating.
Potassium Permanganate Testing involves dropping a stain and seeing the color change on the coating.
For inks, one commonly testing method used is IPA, quiet similar with MEK.

It is advised to use all three tests, and if the results of two of them are the same, then this might provide an answer as to whether there is curing, overcuring, or undercuring.

We recently did a UV adhesion test at PSC testing all possible light configurations, two speeds, and two sets of inks and coatings. What impressed me was the the same coating behaved differently when used with different inks. I am still trying to figure out why; what are the variables involved. Yet, I haven't come up with a good answer.

The best a printshop could do is to establish its onw standards for the inks and coatings. Then, whenever a job is run, conduct a simple test to determine whether curing takes place.

FOGRA has come up with ACET. A device that tests varnishes and inks, and possibly even coatings. The tests are the same with MEK and IPA, only that it takes out the subjectivity of the operator, since it's a device that makes the test.

On the category of indirect tests, we talk mainly of testing the lamps. This is a big issue, and yet not quite objective. UV lamps' performance deteriorate over time, shifting the wavelengths of their light emissions. Since different wavelenghts have different curing capabilities, there is a need of a system that monitors the life of the lamps. Moreover, the lamps are not performing equally in the center and on their edges (imagine you are having a 64 inch press). Still, there are systems (LM9000) that allow constant monitoring of the performance of the lamps and they allow you to determine a warning level based on whether the light emissions are sufficient for curing. In this case as well, each printshop has to determine it's own standards.

Second thing to do is to put some strips that change color when they receive UV radiation. These can be measured spectrally, with a radiometer, or even visually. This doesn't tell us if there is cure or not, only if the UV lamps reach a point that emit not sufficient UV radiation. I measured spectrally some strips from the test we run. I think that the prefered wavelength is at 600nm, since there is where I observed the highest variation from the strips (there is no point in the documentation of the strips -UV FastChecks- that says in which wavelength to measure).

There are also several rub tests, that test adhesion. This two involves a printshop to determine it's own standards.

These are the methods I have tested so far. If I come up with a better one I will let you know. If you know of a better method, please let me know.

There is information on the web about this issue. www.uvprocess.com has a great number of products you might wish to check out.

-D

Stochastic moire

2006-08-14T19:50:00.000-04:00

Hello all,

I run across an interesting case of moire the other day at work. We had to print boxes for a customer that was selling loud speakers. Loud speakers, as you might imagine, have pretty much a halftone pattern on themselves, which is usually grey.

We tried outputting the job with AM screening and we got terrible moire. We figured we had to try FM screening, since it is known for eliminating moire since the dots are randomly dispersed within the addressability grid. Well, we got somewhat less of moire, but there was still there.

It is a case of subject moire, or moire that is embedded in the digital file, due to the halftone pattern of the digital camera's CCD sensor. Hence, since it is on the digital file and not on the halftone screen of the printed sheet, the moire is much less likely -if ever- to go away.

What we tried to do and got successful was to increase the resolution both of the RIP and of the digital file. We figured that a more detailed addressability grid and higher pixel definition would generate FM dots that are more dispersed, thus reducing the likelihood of moire. It worked!

-D

PS. Foveon X3 sensors offer a distinct advantage in avoiding subject moire in the captured image. Check it out: http://www.foveon.com/article.php?a=69 and for a more technical reading: http://www.foveon.com/files/FrequencyResponse.pdf.

coming back to life...

2006-07-28T18:49:00.000-04:00

Seeing the latest entries from Adam's blog I decided that I should try to find some time and maintain this blog that I started as a student. I still consider what it would have been if i weere to go for a PhD in Color Science and stick with research.

From Gerontion (TS Elliot), I quote:
After such knowledge, what forgiveness? Think now
History has many cunning passages, contrived corridors
And issues, deceives with whispering ambitions, 35
Guides us by vanities. Think now
She gives when our attention is distracted
And what she gives, gives with such supple confusions
That the giving famishes the craving. Gives too late
What’s not believed in, or if still believed, 40
In memory only, reconsidered passion.

Oh well! I am working as a Color Assurance Engineer for Pacific Southwest Container. We are the leaders at the west coast for high quality sheetfed packaging printing. The company has 4 KBA presses, 3 of them with interdeck UV lamps, and ranging from 40 to 64 inches.

I can share that most of my work right now doesn't involve technical knowledge; rather, it is building relationships that can make things work.

My main responsibilities are to assure consistent reproduction of color. This involves lots of training for the press operators so that they can understand and employ the various tools that are available. Awareness of the different variation (I can define tham as temporal/over time, spatial/across the sheet and over batch) that are inherent on the process is of paramount importance. Then comes utilization of color bars in order to control variation. I try to break away from the concept of craftmanship. Even the best pressman who claims that he/she is able to control the press by eyesight, there will be some point that eyesight will betray him/her. Printers are proud for being able to be craftsmen, and they deserve all the respect for it. This respect should be maintained at all costs, otherwise they will ose their interest in their jobs and their craftmanship will be replaced by closed loop systems, that however good they are fail to check the conductivity of your fountain solution (for example) and other press variables. It's all about adding a little bit of science to complement craftmanship. Well, as they tell me... good luck with that!

I am actually looking forward to that day that most of the print buyers will be demanding quantifiable color. This will force printshops to become more 'scientific' in their processes. Even if you are able to get the job right, quantifying color and measuring variation helps as a preventive measure.

This might have sound like vocalizing of my internal thoughts.... Just warming up guys. It's been a while.

-D

PerfX Color Management Pro

2006-06-09T13:00:00.000-04:00

Hello everyone,

A few months ago I was contacted by Louis Dery of TGLC (http://www.tglc.com). TGLC is a profiling s/w company from Canada and they provide an ICC color management solution that “uses artificial intelligence for the color calculation”. I was in the middle of testing the IT8.7/4 profiling targets and it appeared as a good opportunity to perform the same tests with profiles made from TGLC’s software.

As I understand TGLC’s solution, it is a profiling software that uses an advanced (intelligent) statistical approach for the construction of the LUT, allowing for more accuracy in the prediction of the color behavior of the profiled device. Additionally, there is a different approach regarding the generation of the black (K) channel. TGLC’s software –and the statistics applied therein- seems to cope better with the nonlinearities of color modeling.

ICC provides the framework for the application of color management technologies, allowing each profiling software developer to come up with a better solution. From a scientific point of view, it is fortunate that there are companies that take advantage of this freedom and implement technologies that are traditionally used in other fields of science.

The graph below displays the cumulative relative frequency (CRF) curve of ∆E*ab from the software I used for the main testing. The further to the left the curve, the least the color difference. The grey curve displays the spatial variation of KPG Approval (which is very good).

As you can see, PerfX software has a very good performance. The green and blue curves are made from the profiling s/w that was used for the rest of the experiment. The blue curve displays the results of the profile made with the profiling target with 6mm and the green with the one constricted at 4mm. Both have large color differences after the 80th percentile. Closer inspection showed that these color differences occur at the dark neutral region (four color overprints). PerfX profiles manage to minimize the color difference at the dark neutrals, and this is the reason why they display less color difference after the 80th percentile. The results were nevertheless almost identical with the performance of the randomized target, which, in general, had better results from the visual one.

Taking these into account, a wish to provide the space of my blog to increase the awareness of the market for TGLC’s software. Any s/w that has something new to offer is welcome, and the choice relies on the end user. I consider increasing awareness to be beneficial for competition, and therefore for the entire market.

Spot Color #3

2006-05-20T16:14:00.000-04:00

My good friend Jorge has some interesting questions about spot colors. I choose to post them as a blog, rather as a comment. I know that I haven't been very serious about maintaining my blog lately, because I am closing up my studies at RIT and prepare to move to Modesto, CA, to work for Pacific Southwest Container as a Color Assurance Engineer (www.teampsc.com).

I invite everyone to take some though on Jorge's questions and initiate a round of discussions.

>>>
I have the following questions/comments:
1. The whole reason on reproducing spot color by 4,5,6 or 7 process colors is productivity. The largest downtime of a print shop dedicated to spot colors is the preparation and matching of them for each job. Having a press with only process colors loaded will avoid ink changes.

2. What does the Spotless technology from Creo offer that cannot be done with the normal Adobe Software? Is it the color separation? Are there any other benefits? (Their web page is very limited on information).

3. Kodak or Creo made a test for offset and determined how many Pantone colors could they reproduce. Does someone have access to these results?

4. The Spotless technology is linked to their Stacatto FM Screening technology and Square Spot. Does this bring large benefits on the consistency of the reproduction of the spot colors using process colors?
>>>

Spot Colors #2

2006-05-04T15:30:00.000-04:00

What are the benefits of printing spot colors instead of process (CMYK) colors?

1. Spot colors are easier to reproduce, requiring a single printing unit and without the need of transparency or trapping that process colors have.
2. They are more capable to achieve accurate color matching, since their hue is more or less defined during their formulation. Assuming that the process is stable then that hue will be reproduced consistently, and the variability will depend upon chroma and lightness. Chroma and lightness are affected by the thickness of the ink film, which is adjustable during the process.
3. Spot colors are reproduced from a single printing unit and a single plate. This, combined with the less makeready time due to lack of trapping issues, make them more cost effective.
4. They are preferred by print buyers who wish to achieve brand recognition, by defining their company or product images with color cues. This is why sometimes they are called 'brand' colors.

Spot Colors #1

2006-04-26T18:04:00.000-04:00

I am starting a series of posts dealing with spot color reproduction.
First, the basics.

Definition:
Special spot colors can be defined as premixed inks that are printed as solids from a single printing unit in order to achieve a visual or a colorimetric match.

There is no available literature on spot colors. Likewise, there is no standard or specification that deals with spot color formulation or reproduction.

There are two categories of spot colors. Those that are specially formulated on the request of a print buyer, and those that refer to a color naming system like the one published by PANTONE.

The only official reference I came across was the gravure industry standard for Color Reproduction and Process Control for Packaging Printing that provides substantial information on special spot colors or brand colors in its Annex B (ANSI CGATS TR012-2003).

The gravure standard defines Brand or Special Spot Colors (SSCs), as colors other than the normal process colors.

One question that several people asked me was why 'Special' spot colors. The only reason I can think of that explains 'special' is that they are 'specially' formulated by premixing different inks for a particular customer.

2006-04-18T17:46:00.000-04:00

I would like to thank Will, Scott and Adam for their contribution to the pdf print engine post.

I hope in the future we will get even more useful discussions!

Thank you,
-D

Adobe's pdf print engine

2006-04-10T17:26:00.000-04:00

Adobe announced the release of PDF Print Engine at IPEX.

I seems to be sthg really important, but I do not think it attracts much attention. If someone knows that there are some ongoing discussions about it, please make a comment about it and i will publish it as a post.

So...
what we have here is a pdf print engine that incorporates pdf and jdf (supporting ICC v4) into the workflows.

Some interesting links:
1. http://blogs.adobe.com/loridefurio/2006/04/adobe_pdf_print.html
2. http://www.atford.co.uk/
3. www.adobe.com/products/pdfprintengine
4. http://www.ondemandjournal.com/specialfeatures/sherburne1.cfm

Will, shares with us his article at:
http://english.ohmynews.com/articleview/article_view.asp?article_class=4&no=236662&rel_no=3

An overall remark: The power goes to the designer (it is going there already though, isn't it?)
Designer's and printers usually speak a different language (RGB vs CMYK). It is not the designer's fault but one has to understand the limitations and desires of the other. The value that printers will offer would be to color manage the chaos that most likely will come forth, and to automate the process. Regardless, the main problems with pdf might be eliminated, namely spot colors and transparencies.

-D

Device link profiling

2006-03-31T11:32:00.000-05:00

Device link profiles are included in ICC.1:2004-10 Specification (profile version 4.2.0.0) as a separate profile type that “contains a pre-evaluated transform that cannot be undone, which represents a one-way link or connection between devices.” A direct conversion from source to destination is achieved by storing the digital values of the transformation in the AToB0Tag and then -based on the sequence of the profiles that serve as source and destination- apply these values to an image. In other words, device link profiles create a direct link between the source and the destination profile that is saved as a separate file, which is then applied to an image and converts it permanently. What is special in this case is that the conversion doesn’t occur through the PCS, allowing for the purity of the inks to be preserved, because the color is not converted from a 4-component color space (CMYK) to a 3-component color space (i.e., CIELAB) and then again to a 4-component color space. This helps avoid printability issues like registration of rich black text or line art and unsharp borders; additionally, ink savings may be realized.

Device link profiles are available as options in standard profiling packages like ProfileMaker 5.0 or as stand alone profiling applications. For my testing I used Alwan's Link Profiler. The results of the testing were published in Test Targets 5.0 publication of the School of Print Media. Alwan's software contains a number of other 'Smart CMM' features that add extra value to the package.

What I tested was whether achieving ink purity by performing a direct CMYK-to-CMYK conversion resulted in colorimetric inaccuracies. More specifically, if a device link profile maps an 100% C, or 100%C&100%M overprint from source to destination, would this result in a colorimetric mismatch that would be perceptible? If the conversion occured through the PCS then the colors would be slightly altered during the conversion to LAB, ending up with contamination and resulting to sthg like 99%C and 3%Y (for example) depending on the white point of the paper, or on architectural shortfalls mainly of previous ICC specifications. This issue is refered as 'scum dots'.

I used paired comparison tests to see if there were any perceptual differences, and measured the basic IT8.7/3 target to see if there were any quantitative differences. The results overall were that device link profiling provides the benefits of maintaining ink purity without any drawbacks on the perceptual or colorimetric quality of the profiles. Also, increases in chroma for each primary and secondary color were maintained at a rather constant hue angle and a slightly larger color gamut was achieved.

Device link profiles are good solutions when there are a number of different workflows and printing devices in a production environment that need to be linked as closely as possible, without realizing problems due to the differences in the printing processes that might be involved.

Optical brighteners

2006-03-28T17:52:00.000-05:00

A topic that is gradually coming into the focus of our industry concerns the optical brighteners that are used in papers.

Optical brighteners are added in the pulp and their effect is that they absorb the light from the ultra violet end of the spectrum (370-400 nm) and in turn give it off in the visible range (400-700nm), usually toward the blue range (430nm), resulting in a whiter and brighter appearance. A white paper treated with an optical brightener emits more visible light than shines on it, making it appear brighter by hiding the yellow and brown tones. Different requirements are placed depending on the paper grade.

However the amount of optical brighteners used is not easy to be measured since the light sources of spectrophotometers usually starts at the 400nm. The question is how can the pulp manufacturing process be controlled, and how can the effects of the optical brighteners be measured over time and under different lightning conditions.

Paper manufacturers are required to produce paper at the lowest cost and sell it at the lowest possible price. This means that they have to use more optical brighteners instead of bleaching which is more costly. In addition, bleaching might effect the other physical properties of the paper in an adverse manner.

A very interesting demonstration to see the variation in the use of optical brighteners would be to take a number of sheets from a given paper batch that has optical brighteners and inspect it under a black light. What you will see will be most likely astonishing.

How consistent is the use of optical brighteners in a single batch of paper I do not know. It would be useful if there was a way to measure the effects of optical brighteners. As I see it, the problem lies in being able to measure and monitor the manufacturing process in order to produce pulp that has a known amount of optical brighteners and to communicate it to the printer.

A good white paper on the issue, especially in relation to proofing, is found at:
http://www.cgs.de/news/whitepapers/WP0511_OpticalBrighteners.pdf.

CIEDE2000

2006-03-27T16:10:00.000-05:00

The DE00 (or CIEDE2000) equation was created in order to correct for the inaccuracies of DEab. It seems that the responsible CIE committee realized that CIELAB was not perceptually accurate in certain regions of the color space (mostly the dark blue and violet) -meaning that the quantified color difference did not correspond with the visual perception of the average observer. Instead of creating a new color space, they came up with color difference equations (DEcmc, DE94 and lately DE2000) that introduced certain weighting factors in the basic deltaE equation. It seems that this solution is working. (There is a lot more of literature on the topic)

CIELAB however is 3-dimensional. This means that when we evaluate the measurement data for, lets say, cyan we have to look at all three dimensions (L*,a*,b*) and most likely also C* and hue.
This constitutes the implementation of colorimeter in process control rather difficult, due to the demand of quick decisions and lack of knowledge regarding color spaces of press operators.

Apart from this, a topic worthy of discussion would be whether DE2000 (which provides significantly smaller DE numbers than DEab) is suitable for process control. The DE2000 data are not that sensitive to slight changes in color, and color shifts that might be important in terms of process control might go unnoticed.

Have a nice trip to Vancouver

2006-03-17T16:15:00.000-05:00

For those of you fortunate enough going to TAGA I wish you a nice and interesting trip.
As for myself, I will be waiting for the Proceedings to find out what's new.

Take care,
-D

and now work...

2006-03-14T11:58:00.000-05:00

The press run for my thesis is done!!!
And now its time for the measurements and then data analysis!

Time for the fun to begin!

Shooting for grey balance...

2006-03-08T19:42:00.000-05:00

Last week I was glad to see an experimental press run at RIT administered by an industry group that -at least as I understood it- had the objective to implement the GRACoL 7 methodology with the SWOP target densities.

As a sidenote, the publication of GRACoL 7.0 has initiated a lot of testing at RIT, and a number of professors, researchers, and students are working towards reproducing the GRACoL methodology, or implementing one that is slightly differentiated but still focuses on grey balance. Grey balance is therefore, a "hot" topic here. We might see results of this work in the following Test Targets publication (for more info, see Adam's blog: http://printmode.net/blog/archives/category/test-targets/)

Going back, allow me to share my thoughts on this methodology, without claiming that I have significant experience in seeing it work or being tested:
For the implementation of a standardized procedure for achieving grey balance you need a very (and I mean VERY) detailed grey balance chart, with A LOT of CMY combinations that cover the tonal reproduction curve. The issue that arises therefore is that depending on 1) the quality/inconsistency of the raw materials, and 2) the nonlinearity of the CMY combinations that can result in a grey, it can be very difficult to achieve it for every press run with the same set of curves. As such, the procedure is not easy to be implemented as a standard.

Moreover, during my limited experience out in the field, I got the impression that most US printers do not have the time and resources to do two press runs for testing.

One might ask then: wouldn't it be easier to go with the traditional method of adjusting your densities to make a visual match of the 50% grey??? No testing, no training, no anxious test press runs. And if a company is willing to implement a quantifiable approach, then it could teach its press operators how to work with CIELCh, measuring the actual grey balance swatch.

PS1. On defense of this methodology I would say that ICC profiling may have the same obstacles, as the created profiles most likely will be different from press run to press run.
PS2. For web offset presses there might be another solution: Closed loop color control systems that adjust the densities based on readings of a grey balance patch.

Refractive index (plus some thoughts on quality control)

2006-03-03T09:31:00.000-05:00

The question that has been unanswered in the literature review of my thesis was how can the refractive index of the spot colors be determined. The refractive index is the change in the speed of light that occurs when light passes from one material to another (think of the distortion that happens when we view an object that is emerged half in water). In the case of inks we have two refractive indices that are of relevance. One is from the air to the ink, and one is from the pigment to the vehicle. I am interested in the second case. If the refractive index is close to unity, this means that there is no refraction of light from pigment to vehicle, and the ink is transparent.

The only references that I found were from personal conversations with professors from RIT and Cal Poly. Their opinion was that the refractive index of spot colors does not need to approximate unity since spot colors do not need to feel the requirement of transparency (they are usually printed as solids). However, it is undetermined how more opaque they should be with relation to process inks.

I am going to test my inks for transparency (as per ISO 2846), but I expect to get just an informative value that I would compare with process inks.

As a final note, I feel that doing this work I go into details that really do not help people make money. I mean that who cares how transparent his spot colors are? Why spend the time and the money to find out something that the customer does not pay you for (I mean as a printer)?

If there is any significance in this level of detail, it would be under the scope of quality control and evaluation of raw materials. Let's say that I make all the tests and at the end determine which are the critical ons in terms of achieving consistent reproduction or even color matching.

As I am saying throughout all this time that I am waiting for my thesis to be completed (it is usually out of my hand): we'll see...

-D

DE distribution

2006-02-28T12:59:00.000-05:00

The lack of normality in the distribution of color difference is a topic that has been discussed by several scientists. Two good references:

1. Berns, R. (2000). Billmeyer and Saltzman principles of color technology (3rd edition). New York: John Wiley & Sons Ltd.
2. Fisch, R., & Bartels, S. (1999). Characterizing an ink on paper four color print process. 1999 TAGA Proceedings. 433-451).

The problem lies in the fact that there can be no prediction with regard to the standard deviations, since the probability that a value lies above or below the mean or median is undetermined. However, it can be assumed that the distributions are always skewed towards a lower value and under that assumption parametric statistics can be used if we are able to determine the value of skewness. In any case however, if that value indicates a near normal distribution, then parametric statistics can be assumed to provide accurate results. HINT: Do check the skewness value of your samples.

Another approach is suggested by Chung and Shimamura (2001, Conducting a press run analysis. Advances in Color Reproduction: Proceedings of the 28th IARIGAI Research Conference, section 2.2) that propose the use of cumulative relative frequency distribution that use rank order statistics to evaluate the results. An identical approach is used by RRDonneley which terms it %CumSum.

These are sufficient solutions for the printing industry. However, if we are called to perform a difference of means test, these approaches cannot provide a reliable solution. In that case we would have to rely on nonparametric statistics, namely the Wilcoxon test that uses rank order to determine whether two populations are identical.

An area that I am still trying to find a solution is the use of sigma statistics (CP & CpK) for colorimetric evaluation. CP& CpK can be used with density measurements, but I do not know if they can be used with color. I am expecting the results of the press run for my thesis at RIT, and I will let you know if there is a solution to this.

-D

3-D CIELAB plots

2006-02-28T12:53:00.000-05:00

CIELAB and CIELCh are 3-dimensional color spaces. Yet, there is a lack of tools that are able to display the coordinates in three dimensional graphs. Usually, we see 2D a*b* plots that display the color difference between two, for example, two patches. However, the achromatic dimension is omitted, and unless we diplay a separate L* plot, L*C*, or L*h plot we are not able to see it. In any case, it is diffucult to evaluate a 3dimensional color difference visually in a 2dimensional plot.

A solution would be to use the 3dplot function in MATLAB to plot 3D graphs. Any other ideas?

-D

Did you know (about COLOR) that...

2006-02-24T17:30:00.000-05:00

(1) ... if our eyes were able to see at longer and cooler wavelenghts we would be able to see ourselves glow due to the temperature of our blood?

(2) ... the sensitivities of our cones developed when our fruit gathering ancestors had to be able to discriminate color differences between ripe and unripe fruit? Better see it before you test it, right?

(3) ... ghosts are not hiding in the dark, but that they are always there. It is just that our rods (scotopic vision) are more sensitive to discriminate differences in low illumination levels (shadows), and when the lights are turned on, the ghosts are still there, but our cones are not able to discriminate their silhouettes? Still afraid of the dark?

(4) ... the color of our skin is orange? Looking at an a*b* plot we can see that there is a basic hue for orange and brown; thus, we could say that we are: orange, Caucasian!
...I still know of a number of people though, who have pink chicks and ruin my theory ;-().

to be regularly updated...

Perceptual vs. Relative Colorimetric rendering intents

2006-02-24T15:17:00.000-05:00

There is some slight confusion as to which of the two rendering intents should be used for pleasing reproduction of pictorial images.

The confusion arises from the fact that in ICC version 2 profiles the perceptual intent is the same with the relative colorimetric with black point compensation checked on (black point compensation /BPC). In ICC version 4, the two rendering intents are supposed to be different. In terms of the structure of the profiles, the difference is that the colorimetric intent tries to match the in-gamut colors, whereas this is not a requirement for the perceptual intent. The out-of-gamut colors are clipped in the colorimetric intent, and scaled with the perceptual. (if BPC is checked on there is scaling at the low L* with the relative as well, and as such the two profiles are almost identical).

Actually, the s/w developers have all the freedom in the world to develop a perceptual intent that results in pleasing reproduction. The only constrain is how 'pleasing reproduction' is interpreted by each s/w developer, or how it can be achieved for each device and/or workflow (pleasingness is different for office documents than posters). There are ongoing discussions in the ICC working groups as to how is pleasingness best tested and evaluated. In other words, which perceptual intent is the best. The main difficulty is the lack of a common reference against which the evaluation could be done. It should be mentioned that this is not against the wishes of ICC, since ICC wants to allow this kind of freedom to the s/w developers.

Available information is found at: http://www.color.org/whitepapers.html

-D

Spot Color Variation

2006-02-22T15:21:00.000-05:00

This blog is an answer to Erik's comment, whom I thank for the information!

I am dealing with spot colors specially formulated to match a certain LAB value withing tolerances and printed from a single printing unit. All the comments about the variation of the press from Eric are valid (valid is an innocent term... they are my nightmare at the moment). Variation depends on temperature (of the room and press), humidity, dampening solution conductivity, hardness and pH, press speed, pressure, paper, water balance... what else... anything. They affect the ink film thickness on the printed paper.
My main problem is that ink film thickness (IFT from now on) is impossible to measure accurately, either because when you test it on a printability tester the ink is not emulsified, or if you test it on a trial press run there are environmental variables and pressure that are unaccountable. Most importantly, even ISO 2846-1 does not provide an accurate means to measure IFT, because it does not take into account the absorbency of the paper. it provides a measure of ink transfer, which is still rather inaccurate. Dryback is another aspect of the whole nightmare. As such, I am shooting in the dark in terms of matching color. What is even more frustrating is seeing the practices of the industry at this point. Ink manufacturers rely on 'quick draw downs' using the Little Joe. Fast and easy, right...? but inaccurate. It is impossible to calculate a range of IFT over DE (as per the standard ISO 2846-1) so that you have an estimate of the suitability of the ink you are using.

I am questioning the industry practices in terms of quality control of raw materials. Even if I fail to provide a solution with my thesis (which would be a set of realistic colorimetric tolerances for spot color reproduction), at least I shoot to provide a methodology by which colorimetry could be used not as much for process control, but for evaluation of raw materials. This is the foremost step in consistent reproduction.

I will keep you updated with my results (hopefully I will have my press run withn the next month), and I will post some other comments related to spot colors and proces control. hopefully, you can pick up some of the points above to start a new conversation.

-D

The number of patches in ICC profiling targets

2006-02-22T14:47:00.000-05:00

How many patches are necessary for an accurate ICC profile? The latest IT8.7/4 has 1617 color patches. To what extend does the profiles created by such a huge set of data is superior than the IT8.7/3 Basic set with 182 patches, or IT8.7/3 Full set with 928 patches.

The question is interesting and raises questions as to the sampling of the color space of the device. While I was doing this research I talked with a number of consultants, professors, or people dong research in the industry, and noone had the opinion that an icrease in the number is important. First, the CMM is important in performing an accurate interpolation, and can account for lack of information at certain regions of the color space. Then, too many patches might create a nonlinear sampling of the color space, confusing the data in a way. Having equal increments might bias the sampling. Another point is that it akes a lot more time to measure the targets (even though the latest spectrophotometers by Gretag and X-Rite shorten the time significantly). A fourth comment was that the color space should be sampled inteligently, meaning that ther should be more patches at the critical regions of the color space (ie. skin tones).

I tested the IT8.7/4-V, IT8.7/3-Full and IT8.7/3-Basic, TC3.5 by Gretag, and a customized version of IT8.7/3 Basic with 91 patches and the inclusion of a CMY grey balance bar.

I did a paired comparison test comparing SCID images with a reference image printed in Test Targets 5.0 publication of RIT to see if there are any perceptual differences between the profiles. I found out that perceptually there is no statistically significant difference no matter what profiling target you use (only the customized target I created betrayed me and was significantly worse than the rest). What is more interesting was that the performance of the profiles was image dependant, meaning that different profiles provided different results for each image. From this single experiment, it could be said that there is no target that is best for all purposes.

To access the quality of the created profiles quantitatively I applied the profiles to 5 a*b* slices and measured the deviation from the target value through the roundtrip approach. Statistically, no difference whatsoever. However, if you take out the confidence intervals you see that IT8.7/3 Full and IT8.7/4 are better. They have more results at the 0.00-2.00 DEab range, and less at the higher DE ranges. I also observed a bimodal distribution, most likely due to the conversion of the out of gamut colors that results in DEab of 3-4.

As a conclusion, I would say that depending on the demands of your workflow you might use even IT8.7/3 (basic) for decent profiles. It does a very good job, and if you are not trying to match critical colors it is excellent. For more demanding workflows one might have to use IT8.7/3 or /4. Betwee these last two however, there is no difference.

-D

Simulate Paper White

2006-02-22T14:15:00.000-05:00

My first post will be about the Simulate Paper White option of Photoshop.
I am currently working at the School of Print Media labs and students come to me with all sorts of questions. Most of the times I can only assume an answer. So what I write in this first post is an effort to answer a question that I was asked today and it is just an assumption of the simulation function unsupported by any experimental results. If anyone has some more concrete thought, please share them here.

The question is how does the Simulate White Paper work. Generally it gives better results for proofing purposes, but what is the function involved?
My guess is that it must perform a chromatic adaptation from D65 (monitor) to D50 (paper / standardized illuminant for printing). The question was that since the absolute intent matches the white point of the paper, shouldn't that be enough for proofing on the monitor. I would say no, because of the difference in the viewing conditions and the illumination. My guess is also supported by the fact that it the function does not work for RGB profiles, which do not need to support D50 for paper.

Here is a link Adam (http://printmode.net/) sent me for some more detailed information on the topic by Andrew Rodney: http://www.digitaldog.net/files/200410rodney.pdf

-D

About myself

2006-02-21T15:11:00.000-05:00

I am currently finishing my thesis for the MS in Print Media at RIT. My topic is on the consistent reproduction of Spot Colors; I am actually testing how different papers, ink key settings and press speed affect colorimetrically the reproduction of color.

I got involved in printing from an early age as my father was running a small printing company in Athens, Greece. Starting off as a printer I got involved in design, publishing and photography, ending up making promotional campaigns for various customers.

One and a half years ago I came to the US to study printing, and I got involved in color management and measurement. My main interest is ICC profiling, and I have been testing various aspects of ICC workflows. Recent research has been on the quality of the profiles relevant to the number of color patches included in the profiling target and on device link profiling.

In the few blogs to follow I will try to summarize the findings of my research so that I can provide topics for converstation or comments.

As a general remark, I consider that color management -and ICC in particular- provides solutions to many problems, but there are many limitations to its application in the real world, either because of the time and cost that it requires for its implementation, or because of certain limitations that are imposed by colorimetry -or by the ICC architecture- as it is today.

I expect that this blog can be used as a communication tool to answer questions or to provide insight to some of the issues that make us all wonder...

-D