Dan Margulis Applied Color Theory

On Preparing a Black Channel for Printing,
And Surreptitious Reseparation by Printers

K in commercial printing
Posted by: "Jeremy Stephenson"
Thu Aug 14, 2008 9:28 am (PDT)

Hi everyone. I have a whole bunch of questions. Sorry!

1)I am preparing a book of photos for printing in Singapore. My printer has told me to use the FOGRA coated sheetfed profile. I would like to use Custom CMYK for different black generation possibilities. Does anyone know what inks the FOGRA profile uses? Are they even among the options listed in the Custom CMYK ink options, or will I have to create the inks myself in the Custom Ink options?

2)As all of Dan's recommendations in PP5E deal with SWOP inks and conditions, I don't think I can rely on any of his settings for anything. I've already discovered that the colour relationships are markedly different than the SWOP profile, but I'm particularly interested in learning more about the black generation.

The FOGRA profile has a black limit of 100% (and total ink of 350%). I notice that the black generation is a lot heavier overall than the SWOP defaults are. Dan recommends a fairly light black ink for general work in North American (SWOP) conditions. Are there any reasons why I should NOT follow that guideline in creating a Custom CMYK profile for the Singapore inks?

3)Also, Dan states that it is advisable to have a black ink limit of 85-90% if I remember correctly. (The SWOP web profile has 90%). Is there any reason FOGRA should be different?

4)Finally, Dan states that having areas of solid ink surrounding a picture can cause that ink to flow too heavily in the surrounding areas as well. I have a number of areas with a picture set in a solid black surround. How big of a problem can that be, and what kind of distance from the solid ink does it have an effect in? I'm wondering if I should lighten the K channel in those images, and if so, how much?

I greatly appreciate any advise from those with experience!

Regards,
Jeremy Stephenson
___________________________________________________________________________

Re: K in commercial printing
Posted by: "John Romano"
Thu Aug 14, 2008 2:23 pm (PDT)

If your printer is suggesting the Fogra Coated profile I would use it.

Or try maybe the LATEST Swop profiles and NOT custom CMYK.

http://www.gracol.org/resources/iccaccept.asp

Mixing K generations on a per image basis is a very bad idea. It can make press adjustments unpredictable....

Regards
John Romano
___________________________________________________________________________

Re: K in commercial printing
Posted by: Terence Wyse
Thu Aug 14, 2008 2:23 pm (PDT)

Hi Jeremy,

Comments inserted...

1)I am preparing a book of photos for printing in Singapore. My
printer has told me to use the FOGRA coated sheetfed profile.

...which is the "ISOcoated_v2_eci.icc" profile and based on the FOGRA 39L data set.
Make sure you have the latest versions available here...
http://www.eci.org/doku.php?id=en:downloads
...under the heading "ICC profiles from ECI". You'll want to download the "eci_offset_2008" package.

I would like to use Custom CMYK for different black
generation possibilities.

What is your reasoning for wanting to do different black generation?

Does anyone know what inks the FOGRA profile uses?

"ISO Coated" uses the FOGRA 39L data set. The inks in FOGRA 39 conform to ISO 12647-2 colorimetry for Paper Type 1 (gloss coated). They are not "SWOP" inks at all. Incidentally, the "G7" data set for commercial sheetfed coated (GRACoL2006_Coated1) uses the same inks.

Are they even among the options listed in the Custom CMYK ink options, or
will I have to create the inks myself in the Custom Ink options?

No, they are not in the Custom CMYK ink options. And I would not recommend trying to creating the inks yourself since solid ink colorimetry is only one aspect of a good press profile. The other major component is the tonal curve.

The only sure way you're going to get a valid profile is to download the FOGRA 39 data set, the same data set used to build the ISO Coated profiles, and build your own profile using this data set as a reference.

2)As all of Dan's recommendations in PP5E deal with SWOP inks and
conditions, I don't think I can rely on any of his settings for anything. I've already
discovered that the colour relationships are markedly different than the SWOP profile, but I'm particularly interested in learning more about the black generation.

If you're interested in printing to standards, you should simply use the profile specified by your printer in Singapore. You need to have a VERY GOOD REASON to go in and create custom black generation settings that are non-standard. The only compelling reason I can think of to do that is if you have a "B&W" image you want to print using process inks. In that isolated incident you may want to use a full-range black with very aggressive GCR settings to minimize color casts with these B&W-out-of-process images.

The FOGRA profile has a black limit of 100% (and total ink of 350%).
I notice that the black generation is a lot heavier overall than the SWOP defaults are. Dan
recommends a fairly light black ink for general work in North American (SWOP)
conditions. Are there any reasons why I should NOT follow that guideline in creating a Custom CMYK profile for the Singapore inks?

Your information is incorrect. The latest ISO Coated profile from ECI has a total ink of 330%. There is also a variation of this profile with 300% total ink ("ISOcoated_v2_300_eci.icc")

It so happens that the latest *official* GRACoL and SWOP profiles use very similar settings to the ISO/FOGRA profiles with the exception of total ink amount (GRACoL Coated1 uses 320%, SWOP Coated3 is 300%).

3)Also, Dan states that it is advisable to have a black ink limit of 85-90% if I remember
correctly. (The SWOP web profile has 90%). Is there any reason FOGRA
should be different?

There's no reason you can't use upwards of 95% black in a profile. All the latest ISO and GRACoL/SWOP profiles for coated papers use a 98-100% black limit. This is per recommendations from various standards bodies and working groups. They should know what they are doing I suspect. The fact is, todays high-end profiling applications like ProfileMaker, MonacoPROFILER, PrintOpen and others can make very high quality profiles with high black limits and very aggressive GCR.

4)Finally, Dan states that having areas of solid ink surrounding a picture can cause that
ink to flow too heavily in the surrounding areas as well. I have a number of areas with a
picture set in a solid black surround. How big of a problem can that be, and what kind of
distance from the solid ink does it have an effect in? I'm wondering if I should lighten the
K channel in those images, and if so, how much?

With a solid K surround, I would strongly recommend using a "rich black" for the black surround. There's various recipes for a rich black but 100k+40c is common or you could go with a 4-color recipe of, for example, 100k+40c+30m+20y. The 40c30m20y will insure a "neutral" balance that is biased slightly towards blue which is preferable to one that could go warm.

Using a rich black in this way insures that they won't have to increase the black density beyond a normal density value which will keep dot gain under control. Without the CMY undercolor to "beef up" the K density, the pressman may be forced to increase the K density and lose control of the K dot gain which will affect the nearby images.

I can't stress enough that you should stick with the printer's recommendation as far as standard profiles. If you feel you have a very good reason to use a different K generation, you could possibly ask the printer to create one for you from the same FOGRA data set used for the ISO?FOGRA Coated profile.

Hope this helps,
Terry Wyse

_____________________________
WyseConsul
Color Management Consulting
G7 Certified Expert
wyseconsul at mac dot com
704.843.0858
___________________________________________________________________________

Re: K in commercial printing
Posted by: Michael Jahn
Thu Aug 14, 2008 2:23 pm (PDT)

Hi Jeremy,

my opinion on your questions

-- (using my older mean professor voice!)

1. If they are asking you to create color separations that are designed for printing in a press condition that follows the Fogra specification, I can't imagine why on earth you would be considering doing any custom CMYK conversion - do you feel you are somehow have a better grasp than the color scientist than the folks at Fogra ?

If you can't find out WHICH Fogra profile to use for some reason, I would suggest you use; for your CMYK working space - Use the Fogra 39 (Coated FOGRA39 (ISO 12647-2:2004)) - or better (higher number, more recent data)

If you are wondering what RGB working sapce, I would personally go with Adobe 1998 for my RGB working space.

2. why concern yourself with SWOP - if they are asking for Fogra - SWOP is mainly for magazine publications here in the USA, where we are printing heatset offset on cheap paper.

- if this job is being printed in Singapore, and they are saying that they want Fogra, this printer is following the method and specification as outlined by FOGRA. They use different ink densities, may use different inks - in a very rough way, you can consider Fogra to the the Europe version of SWOP, but it is not really THAT simple (TVI vs G7 method of getting the press set up, yada yada)

3. see answer to 1.

4. see answer to 1.

5. see answer to 1 - hey, if Dan were a printer, and he was asking you to follow some "custom" method for preparing your image, then Dan would say "hey, here is my custom profile, please use it to generate your CMYK conversions so i can take advantage of <insert reason here>

But Dan is not doing the printing, some people in Signapore are, they print with their printing method, and they are simply asking you to use the Fogra profile to generate you CMYK image - they may indeed re-separate into some custom CMYK from their end, using their own custom profile (why know, they may reseparate to CMY with a teeny tiny black, or convert your CMYK to RGB and then to Hexachrome - YOU DONT KNOW - so, I would gently suggest that since you have no idea what they need, not to guess or some how say "but DAN told me DIFFERENT!" and do as your told.

and sit up staright. and clean your room.

LOL

So, do not overthink !!!

- the last thing you want to learn is that you did something that tied their hands and shoe laces and when you complain about the result they say "hey, we printed what you sent us, and your crappy CMYK did not following FOGRA".

Michael Jahn
___________________________________________________________________________

Re: K in commercial printing
Posted by: Michael Jahn
Thu Aug 14, 2008 3:51 pm (PDT)

Well said Terry - Gold star for you

@ John Romano - NO SOUP FOR YOU. <wink> sorry, I think no one in Singapore would like (or even ACCEPT) separations that required a different ink set - so do not think for even a nanosecond you can subsitute SWOP when FOGRA is requested -- please understand that this is not some teeny tiny difference, things like flesh tones and gray balance would be in an uncorrectable condition with miserable looking results.

--
Michael Jahn
Jahn & Associates
___________________________________________________________________________

Re: K in commercial printing
Posted by: Terence Wyse
Thu Aug 14, 2008 5:46 pm (PDT)

Thanks Mean Old Professor Jahn.

Now, can I go back and sit in the corner and sleep during the rest of your class?

;-)

Terry
___________________________________________________________________________

Re: K in commercial printing
Posted by: Dan Margulis
Thu Aug 14, 2008 9:54 pm (PDT)

Jeremy Stephenson writes,


1)I am preparing a book of photos for printing in Singapore. My
printer has told me to use the FOGRA coated sheetfed profile.

In the absence of independent verification that the printer actually prints to these standards, I would give little credence to this advice.

I would like to use Custom CMYK for different black generation possibilities. Does anyone know what inks the FOGRA profile uses? Are they
even among the options listed in the Custom CMYK ink options, or will I have to create
the inks myself in the Custom Ink options?

The question is premature unless you or some other competent party has actually assessed how well this printer performs in real life. There is no point worrying about FOGRA or any other standard when there is no assurance that the printer can hit it.

If you have no clue at all about what the printer's true capabilities are other than their unsupported word, creating a FOGRA-like profile in Custom CMYK seems like excessive work. Instead, I'd follow the directions on p. 325 for merging Custom CMYK black generation with a preexisting third-party profile, and hope for the best. Remember, though, unless you have clear evidence of unusual competence on the printer's part, you are better off leaving your final CMYK files slightly light, because if there is considerable variation during the pressrun lighter images look better than darker ones.

2)As all of Dan's recommendations in PP5E deal with SWOP inks and
conditions, I don't think I can rely on any of his settings for anything. I've already
discovered that the colour relationships are markedly different than the SWOP profile, but
I'm particularly interested in learning more about the black generation.

The FOGRA profile has a black limit of 100% (and total ink of 350%).
I notice that the black generation is a lot heavier overall than the SWOP defaults are.
Dan recommends a fairly light black ink for general work in North American (SWOP)
conditions. Are there any reasons why I should NOT follow that guideline in creating a Custom CMYK profile for the Singapore inks?

No, although there is less reason to worry about it given sheetfed rather than web printing. Sheetfed presses are easier to control and you are less likely to get a really bad color shift.

3)Also, Dan states that it is advisable to have a black ink limit of 85-90% if I remember
correctly. (The SWOP web profile has 90%). Is there any reason
FOGRA should be different?

No. A higher limit than this is excessive, and makes the file more difficult to correct, even if the printer is known to be competent. In a case like this, where the printer's competence is apparently unknown, going with a higher limit is a really poor idea.

4)Finally, Dan states that having areas of solid ink surrounding a picture can cause that
ink to flow too heavily in the surrounding areas as well. I have a number of areas with a
picture set in a solid black surround. How big of a problem can that be, and what kind of
distance from the solid ink does it have an effect in? I'm wondering if I should lighten the
K channel in those images, and if so, how much?

This can be quite a considerable problem, which is why such large, solid black areas are best avoided. (Adding other colors to the black mix won't help the gain in black density).

As a rule of thumb, if the black area is large, one can expect that densities will vary noticeably in a surrounding area of the same width. If would be one thing if you knew for a fact that only one picture was going to be affected. In that case, you could create a special sep for it: say one version separated with a UCR setting, and a second separation with None for black generation, then merge the two 75-25.

It might seem like that would solve your problem if, say, you had one picture in the middle of a page that was otherwise solid black. Unfortunately, the density variation will also affect the page that's underneath or above that one on the press form. Unless you know how the book is being imposed on press, you won't know what page that is, so you may get a nasty surprise in print when some seemingly unrelated image turns out muddy.

If there's no way of avoiding using large black design elements throughout the book, then probably the best approach would be to separate *everything* UCR.

If you are thinking, if this extra inking caused by the large black areas is such a problem, then how come the cover of PP5E, which uses such large black areas, came out OK, here's the story. The cover was designed in-house. When I saw it, I told Peachpit that no offset printer on the face of the planet could match the proof given this layout. Peachpit consulted the printer, who agreed with this assessment. Consequently, since Peachpit would not abandon the design, the cover was printed CMYKK--a 5-color job, with the black background provided by a plate that didn't share anything with the images.

Dan Margulis
___________________________________________________________________________
.
Re: K in commercial printing
Posted by: Michael Jahn
Fri Aug 15, 2008 7:12 pm (PDT)

So, I guess Dan brings up a good point.

Just because they say they are a 5 start hotel on the phone, you may find yourself in a hut without running water and no electricty.

Dan is convinced that since he cant verify, he is skeptical of a printer in Singapore and (perhaps) Skeptical of a printers claim of following a standard (Fogra, SWOP, or otherwise)

so - basically, he suggests that you starve the separations and put the printer into a position where the worst they can do is print too light, but make it nearly impossible to print to dark.

I can't fight a ghost.

That is, I can't argue with someone who says "I don't wear seat belts because my car might catch fire or go off the side of a bridge and go under water and then ..."

Hey, I was dropped on my head as a child, and fed lead paint. Maybe that is why I actually trust standards, and think Dans suggestion is crazy. But then, most of my experience has been limited to working with big iron companies who I can actually meet personally and review the press sheet with.

Perhaps Dan has experience in Singapore, and perhaps Dan never wears his seatbelt when he drives in Singapore because he knows better.

But it is against every fiber of my being to think that if some printer knew enough to mention, suggest - even ask - for Fogra, that I would not do that, and think I am more clever (because I have no way to prove it)

But I can reject the job if i do what they say, and it does not 'match' my Fogra simulation proof.

--
Michael Jahn
___________________________________________________________________________

Re: K in commercial printing
Posted by: "Jeremy Stephenson"
Fri Aug 15, 2008 7:12 pm (PDT)

Thanks to everyone for their responses.

There seems to be two camps here - Dan and everyone else! Everyone else says use the supplied profile, and advise against other black generation - Dan says pretty much the opposite.

At least you all agree against creating Custom CMYK inks. Dan's suggestion (on pg 325 of PP5E) is to create a different black generation using Custom CMYK, but retaining the color of the supplied profile by doing a color/luminosity blend of the two.

For your information, I have a wide variety of images in my book - colour with important shadow detail (lots of night shots), and B/W - and quite a lot of the B/W has important shadow detail. I also have the added issue mentioned before of 100%K (actually rich black in most places) being the background colour of some pages. I realise now that it makes it tough on the printing, but it's too late to redesign the whole book now.

Also, for your information, from what I can gather, I would trust this printer to be as consistent as possible. They have a lot of good reviews, and I have seen samples (printed books on the same paper stock) of their work which looks very good. I haven't seen the original files, but I tend to being more than less confident. The exact profile I was told to use was Coated Fogra 27.

Dan, John Romano states "Mixing K generations on a per image basis is a VERY (emphasis mine) bad idea. It can make press adjustments unpredictable...." Any thoughts on that? You seem to say in PP5E that it is a good idea to adjust black generation in different images differently in the same publication depending on what you want to protect against - either shadow detail loss, or colour variation (particularly in neutral or B/W images.

I am asking Stamford (the printer) if they can print a second K ink. That would make life a lot easier. The other option would be to use UCR like Dan suggested, though that jeopardises my black and white images.

Incidentally, in response to Terry Wise's comment, that I was incorrect saying that the total ink limit is 350%, I found that figure by converting 0,0,0 RGB into the FOGRA 27 profile, and the sum of the inks produced was 350%, not 330%, as you stated, Terry.

Thanks for all your input. I am awaiting more info from the printer to decide on my next step.

Jeremy Stephenson
___________________________________________________________________________
.
Re: K in commercial printing
Posted by: Terence Wyse
Fri Aug 15, 2008 7:12 pm (PDT)

On Aug 15, 2008, at 12:47 AM, Dan Margulis wrote:

In the absence of independent verification that the printer actually
prints to these standards, I would give little credence to this
advice.

Dan, don't you think you should at least take them at their word first before taking the default position that they are NOT printing to the specification that they suggested to the customer? In my experience, the fact that they even had a clue what *ICC profile* to recommend indicates that there's a better-than-even chance that they print at least close to that specification. There's a much better chance that this is true compared to a sheetfed printer here in the USA that says they print to "SWOP". That would be your first clue that they do NOT know what they are talking about (SWOP = web offset, what business does a *sheetfed* printer have printing to SWOP??).

The question is premature unless you or some other competent party has
actually assessed how well this printer performs in real life. There
is no point worrying about FOGRA or any other standard when there is
no assurance that the printer can hit it.

Which is why you provide a *proof* that can be verified to be within the specification that they claim to be printing to and have them match it (within tolerance). Honestly, with CERTIFIED inkjet proofing systems that are under $2K these days, there's no excuse for not being able to provide such a proof for folks serious about supplying CMYK images and printing to standards.

As for the rest of your comments, what your "audience" should realize is that more-and-more printers are using "ink optimization" software in prepress before sending the job out the pressroom. "Ink optimization" software is essentially the use of device link profiles during the plating process to *re-separate* the entire job with whatever amount of GCR/K generation that the printer deems fit for their pressroom. This is done both as an ink savings measure but also as a seamless method to print to certain specifications relatively accurately compared to the "old days" of simple plate/press curves. When the device link profile is built, they simply marry up their custom press profile with whatever "standard" profile they care to use.

Point is, all this wrangling and wringing-of-hands as to what exactly my K generation should be on a given image is pretty much for naught since the job may be getting re-separated on its way to the press anyway. This trend is only going to continue as more domestic print work goes overseas and the folks here in the states and other countries look for ways to cut costs. The irony is that this "cost- cutting measure" of using ink reduction software and device link profiles actually has the potential to IMPROVE printed color reproduction.

Regards,
Terry Wyse

_____________________________
WyseConsul
Color Management Consulting
G7 Certified Expert
wyseconsul at mac dot com
704.843.0858
___________________________________________________________________________

Re: K in commercial printing
Posted by: Terence Wyse
Sat Aug 16, 2008 10:14 am (PDT)

On Aug 15, 2008, at 4:10 AM, Jeremy Stephenson wrote:

Incidentally, in response to Terry Wise's comment, that I was
incorrect saying that the
total ink limit is 350%, I found that figure by converting 0,0,0 RGB
into the FOGRA 27
profile, and the sum of the inks produced was 350%, not 330%, as you
stated, Terry.

All I can say (and I said as much in my first post) is that the *current* ISO Coated profile is NOT based on FOGRA 27 but is based on FOGRA 39. FOGRA 39 superseded 27 around the time that the IDEAlliance published the GRACoL2006_Coated1 data set which happens to be also based (loosely) on FOGRA 39 with a tweak to have it conform with the G7 NPDC.

I personally believe John Romano's comments regarding mixing different black generations on the same form is right on.

I simply reject the notion that using UCR/GCR and K generation as a *creative* tool in preparing images is a good idea. As pressroom color management (via device links) becomes more pervasive, K generation, total ink limit, etc. will be properly taken out of the hands of the creative process and put back in the hands of prepress and the pressroom where it belongs.

Regards,
Terry Wyse
_____________________________
WyseConsul
Color Management Consulting
G7 Certified Expert
wyseconsul at mac dot com
704.843.0858
___________________________________________________________________________
.
Re: K in commercial printing
Posted by: Michael Jahn
Sat Aug 16, 2008 10:14 am (PDT)

@ Dr. Terry;

you wrote (ah, you "note" and you spote !)

"... all this wrangling and wringing-of-hands as to what exactly
my K generation should be on a given image is pretty much for naught
since the job may be getting re-separated on its way to the press
anyway."

Which - is fact. That is what many people do - and they save money. CGS, GMG, Alwan - all follow what Eric at Left Dakota made popular, the wonders of a device link profile - CMYK to CMYK. Now, mind you - this too has it color science problems....but i digress...

When I was director of the prepress division of a gravure facility (now quebcor, was shea communication) - when we were asked to send 9 track mag tape to other engravers, we would send the smallest black possible - as we knew that there was no Gravure ink standard, we all collectively agreed to exchange CMY with a teeny tiny black.

So, while I fully grasp and understand Dans "protect yourself!" position - obviously, like so many of us over 50 - created in an environment of working with morons who had no clue about something as simple as gray balance was - well - i feel dans pain, but like you, have decided to trust paypal, the gas I pump, my water bill and the concept that if someone has asked me to generate color separations that using a profile popular enough to actually be included in the Adobe Photoshop installer, well - i trust that profile.

Dan would have been doing a disservice to the forum if he simply said "do what you are told" he probably has more scars than we could count on his checkbook, his psyche and his body doing that.

Soon, when it is announced that converting into LAB space was a mistake in the first place, we will all know why we are in the mess we are in.

hoo-ray for I J K.

(do not bother trying to google I J K - you will only find the math part. but if you are itchin' for a headache;

handprint.com/HP/WCL/color2.html -- just add the www - and search for *Helmholtz-Kohlrausch effect*

Michael Jahn
___________________________________________________________________________

Re: K in commercial printing
Posted by: Dan Margulis
Sat Aug 16, 2008 10:14 am (PDT)

Jeremy Stephenson writes,

There seems to be two camps here - Dan and everyone else! Everyone else says use the
supplied profile, and advise against other black generation - Dan says pretty much the
opposite.

The question is not one of how to *print* but how to *prepare* a file to be printed. The print industry has no history of participation in this process and limited expertise in how to do it sucessfully.

It is closely analogous to *taking* a photograph vs. color correcting a photograph. Retouchers by and large don't understand the nuances of photography or what the various options are at the photography stage. They don't know what the photographer might have done to make their life easier. They are given a file to process, and they do so as best they can.

Similarly, the print industry's historic job is to take an existing CMYK file and put it on press. They have not been involved with how to prepare it, do not generally understand what the options were in preparing it, and do not understand how the preparation thereof might make their jobs easier. Consequently, just as retouchers are generally unqualified to tell photographers how to shoot, printers are generally unqualified to tell retouchers how to prepare files for printing, except in terms of stating an ink limit. The inability to give such advice in no way suggests that retouchers or printers are incompetent--they can do their own jobs well, but it's hard for anybody to go beyond their own expertise and specialty.

My suggestion would be that anyone attempting to lay down strictures on how photographs should be lit or exposed should, at a minimum, have a couple of years experience as a professional or near-professional photographer. Also, I would suggest that anyone attempting to lay down strictures as to how to use GCR or other alternate black settings should have at a minimum a couple of years experience as a professional retoucher who specializes in delivering CMYK files to a variety of printers, including cases where he knows little or nothing about the printer's capability.

Since the others who have responded do not, AFAIK, have such pertinent experience, the suggestion could be that what they consider "good enough" quality might not be considered so elsewhere.

Another way of extending the analogy: if retouchers nevertheless took it into their head to advise photographers on how to shoot, the result would not be worthless, there would probably be some sensible suggestions, just as some of what the print-oriented people here say is sensible.

However, it is very likely that photography guidelines prepared by retouchers would be most effective where the photographer is a fool and the retoucher an expert--and less effective elsewhere.

Similarly, the advice you are being offered here by others is the advice I would myself give if a) I knew you to be a fool; and b) I knew for a fact that the printer is very good. As I do not know either of these things, my advice is different.

For your information, I have a wide variety of images in my book - colour with important
shadow detail (lots of night shots), and B/W - and quite a lot of the B/W has important
shadow detail. I also have the added issue mentioned before of 100%K (actually rich black
in most places) being the background colour of some pages. I realise now that it makes
it tough on the printing, but it's too late to redesign the whole book now.

Images that have important shadow detail are more difficult to print with the excessively heavy black channel being advocated here. This is particularly so if the printer is not very good. If shadow detail is critical, you should not be using such a heavy black. Rather, the black should be full-range, full of detail.

Also, for your information, from what I can gather, I would trust this printer to be as
consistent as possible. They have a lot of good reviews, and I have seen samples (printed
books on the same paper stock) of their work which looks very good. I haven't seen the
original files, but I tend to being more than less confident.

That's an important piece of information, as it would incline you to be less conservative in preparing your files, less inclined to create the files defensively.

Dan, John Romano states "Mixing K generations on a per image basis is a VERY
(emphasis mine) bad idea. It can make press adjustments unpredictable...." Any thoughts on that?

On the contrary, it makes press er, "adjustments", *more* predictable. Given two apparently identical images one of which uses a heavier GCR, any "adjustment" to both will affect the darkness of the heavier GCR more and the color less. This is not just predictable--it's a certainty. Because successful CMYK file prep depends on anticipating these "adjustments", and because different images have different priorities, retouchers have been trying to control these "adjustments" by selective use of GCR ever since it became possible to do so.

As you may notice, I am not a big fan of "adjustments" on press. While occasionally they are necessary through no one's fault, normally they indicate either 1) the pressman is trying to correct what he sees as an obvious flaw in the client's work or 2) the pressman is trying to chase his own proof because the company's process control is not very good and/or because its color management is inadequate. Hopefully, #1 is not relevant in your case. If #2 is the cause of the "adjustments", all the more reason to be skeptical about the supplied profile.

There is not the beginning of a ghost of a scintilla of a technical reason to avoid mixing black generations on press. Every commercial printer receives different black generations from different clients. Every carefully-prepared large CMYK job likely has different black generations somewhere. Every job containing CMYK files that originate from different sources has different black generations throughout. The printer doesn't know this, because they don't need to know, and there's no reason to tell them--their job is to print what they are given, as best they can.

You seem to say in PP5E that it is a good idea to adjust black generation in different
images differently in the same publication depending on what you want to protect
against - either shadow detail loss, or colour variation (particularly in neutral or B/W images.

Well, more against an overall feeling of muddiness than shadow detail specifically, but that's a part of it. Yes. This has been standard in professional work for a long time. It has a large impact on print quality, although most commercial printers don't even know that the factor is present.

Dan Margulis
___________________________________________________________________________

Re: K in commercial printing
Posted by: Dan Margulis
Sat Aug 16, 2008 10:14 am (PDT)

Terry writes,

Dan, don't you think you should at least take them at their word first
before taking the default position that they are NOT printing to the
specification that they suggested to the customer?

I don't take them at the word, no, nobody has suggested that they are NOT printing to the specification, I suggest that we don't know whether they are or they aren't and we don't know how repeatable they are or how good their process control is. I suggest that, granted a lack of knowledge, we do not hamstring ourselves by unwarranted assumptions, and that we should *not* treat them as if we had the much more desirable situation where we knew for a fact that the printing was going to be real close to the profile and the process control was reliable.

In my experience,
the fact that they even had a clue what *ICC profile* to recommend
indicates that there's a better-than-even chance that they print at
least close to that specification.

In my experience, I'd say about a 5 percent chance that they print acceptably close to the spec, but your experience may vary and so may your idea about what constitutes "acceptably close". I would say there is a good chance that they print somewhere in the general neighborhood of what they are suggesting, so I would start with your suggested profile as a base. However, because I do not feel that the black generation will give the best opportunity to maximize quality with an unproven printer, I have to advise editing it, and since the Photoshop team refuses to put this essential feature into the program, I have to advise the OP to upgrade to Custom CMYK.

There's a much better chance that
this is true compared to a sheetfed printer here in the USA that says
they print to "SWOP". That would be your first clue that they do NOT
know what they are talking about (SWOP = web offset, what business
does a *sheetfed* printer have printing to SWOP??).

That's correct, it would be more analogous to a *web* printer claiming to print SWOP. Yes, a sheetfed printer claiming to print SWOP is an important indication of incompetence, which brings up an interesting point.

Put yourself in the shoes of a retoucher. You don't generally get to pick the printer, it's done on price or other factors by the client. Sometimes you have no information, sometimes a lot, and sometimes in between. In this hypothetical case, your information is that the sheetfed printer chosen to do the job claims to print SWOP.

Standard rules of the game apply: if the client is dissatisfied with the final job, YOU are to blame, regardless of what the printer did. Excuses about how the printer isn't printing to his standard, how a different printer should have been chosen, etc., etc., don't fly.

This is quite a real-world example, BTW. You have a clear indication that the printer is incompetent, but you have to work with him nonetheless and your success depends on his. What do you do?

My answer involves a lot of different things, but one for sure is, don't give him a heavy black plate.

Which is why you provide a *proof* that can be verified to be within
the specification that they claim to be printing to and have them
match it (within tolerance). Honestly, with CERTIFIED inkjet proofing
systems that are under $2K these days, there's no excuse for not being
able to provide such a proof for folks serious about supplying CMYK
images and printing to standards.

If there is a proof that all agree in advance is to be the guidance then the problem does not exist, regardless of who provided the proof. From the OP I would guess that there is no such proof. Whether there is an excuse for not having one is not particularly relevant, what's important is there is one.

As for the rest of your comments, what your "audience" should realize
is that more-and-more printers are using "ink optimization" software
in prepress before sending the job out the pressroom.

Yes. A strong argument for reminding them that they should *not* embed CMYK profiles when handing off the files, because of the possibility that they can be used as a hook by some thoughtless printer who may wish to reseparate them, and also that they include a statement in the job jacket prohibiting the printer from reseparating the files and indicating that doing so will result in rejection of the job.

Meanwhile, *your* audience should be reminded that they need to consult with their lawyers before invoking such a workflow. Those lawyers are likely going to tell them that they need to have a specific release from each client authorizing them to reseparate, and not just the absence of a prohibition. Otherwise, if they reseparate the work of a job whose quality depends on an effective black, they are likely to get themselves not just a large nonchargeable remake but also a bill for consequential damages.

Point is, all this wrangling and wringing-of-hands as to what exactly
my K generation should be on a given image is pretty much for naught
since the job may be getting re-separated on its way to the press
anyway. This trend is only going to continue as more domestic print
work goes overseas and the folks here in the states and other
countries look for ways to cut costs. The irony is that this "cost-
cutting measure" of using ink reduction software and device link
profiles actually has the potential to IMPROVE printed color
reproduction.

The potential is there, yes, but there's also a huge potential for damage. Screen grabs, comic strips, and other files with thin black lines depend upon one kind of black. Alternate settings are often used to avoid trapping issues. 4/c grayscale images depend on a kind of black that isn't suitable for other CMYK. Shots of silver jewelry, drop shadows, and other objects of mandatory neutrality require a heavier black, while clothing catalog work where the ability to adjust color on press is essential wants a lighter one.

If the printer takes it upon himself to reseparate, and as a result the thin black lines show up as a fuzzy CMYK rather than a crisp 100K, and a white line that indicates poor trapping appears, and the jewelry gets green because there was a tiny inking imbalance on press that would never have been noticeable in the client-supplied file, and the drop shadows turn purple because there isn't enough black in them, and the clothes can't be changed enough on press because there's too much black in areas where the client-supplied file had none, then you will understand that the clients is absolutely entitled not just to reject the job, but to ask the printer to pay for any delay in reprinting it.

Dan Margulis
___________________________________________________________________________

Re: K in commercial printing
Posted by: "Mike Russell"
Sat Aug 16, 2008 10:14 am (PDT)

From: "Michael Jahn"
...
But it is against every fiber of my being to think that if some
printer knew enough to mention, suggest - even ask - for Fogra, that I
would not do that, and think I am more clever (because I have no way
to prove it)

The burden of proof is in the other direction. Having someone at the printer's mention the name of a profile says very little about the printer's ability to match that profile.

Maybe the printer is Annie Oakley, and can hit the pip of the ace of spades at 100 yards. More likely, he is not, in which case Dan's strategy of making the target larger is a more conservative strategy. Annie Oakley will prevail in either case. A lesser marksman is much more likely to hit the target in the second case.

If the printer is above average, he will be able to make any reasonable job look good. If the printer's abilities are more modest, then Dan's precautions would help to ensure that the job will be acceptable.

Mike Russell - www.curvemeister.com
___________________________________________________________________________

Re: K in commercial printing
Posted by: Terence Wyse
Sun Aug 17, 2008 7:38 am (PDT)

On Aug 16, 2008, at 1:05 PM, Dan Margulis wrote:

Meanwhile, *your* audience should be reminded that they need to
consult with their lawyers before invoking such a workflow. Those lawyers are likely
going to tell them that they need to have a specific release from each client authorizing
them to reseparate, and not just the absence of a prohibition. Otherwise, if they reseparate
the work of a job whose quality depends on an effective black, they are likely to get
themselves not just a large nonchargeable remake but also a bill for consequential damages.

So you think the prepress/printer's job is to print exactly the CMYK values supplied in the original file? Ever hear of plate curves (of course you have)? They "alter" the CMYK values before they go to press. The on-plate values are no longer what was described or specified by the original content. Would this constitute some sort of legal violation of the intent of the content provider? If so, you'd probably have to put about 80% of the printers in jail because a large majority of them utilize plate curves to either help them print to a specification or at least some notion of target print condition. Plate curves were almost universally used with the advent of CTP to help folks achieve TVI that was similar to the previous film-to-plate workflow. As folks have gotten used to the "cleaner" look of linear CTP, there use sort of fell away after a while but now with the advent of the "G7" methodology and the push for global print standards in general, plate/press curves are once again being used extensively in prepress.

Device link profiles to prepare a set of plates for a specific press condition is simply a natural evolution from the common usage of plate curves. They both constitute a form of "color management" for the pressroom. Some may say device links are a different case since they (can) alter GCR/K generation etc. but you can't get around the fact that both methodologies are still a form of color management and that they both alter the intended CMYK values to better suit a target print condition.

I would say that the printer's ONLY job is print the job in such a way that it meets the customer's expectation. This expectation should be set based on a known print condition, "standard" if you will, and a contract proof that reflects that standard. How the printer chooses to meet that expectation with his particular equipment is up to them and not for the content provider to determine, anymore than it would for the printer to tell the designer how to design or the photographer how to photograph.

Regards,
Terry Wyse
___________________________________________________________________________
.
Re: K in commercial printing
Posted by: J Walton
Sun Aug 17, 2008 9:40 am (PDT)

On Sat, Aug 16, 2008 at 10:56 AM, Terence Wyse wrote:

I would say that the printer's ONLY job is print the job in such a way
that it meets the customer's expectation.

That makes sense as a general rule of business, but what if the client is expecting that they don't monkey with their files?

This expectation should be
set based on a known print condition, "standard" if you will, and a
contract proof that reflects that standard.

But the contract proof won't show misregistration, or accidental press "adjustments." Dan brought up a few pretty good examples of times when a non-standard Black is essential to quality work. The cartoon example WILL NOT look right, or like the contract proof, without respecting the separation. The 4C Black, or quadtone, example likewise WILL NOT look right without a heavier black. I agree that this practice of automatically re-separating is common, and when dealing with neophytes probably improves the final product.

I suppose one question I would pose is - what does the typical device link setup do if a CMYK file does not have an embedded profile? Will it ignore the file or assume a source that is inaccurate?

How the printer chooses to
meet that expectation with his particular equipment is up to them and
not for the content provider to determine, anymore than it would for
the printer to tell the designer how to design or the photographer how
to photograph.

I disagree with this point. I am a retoucher, but I tell photographers how to photograph all the time *when they screw things up.* "No, dude, you can't move the camera when you are taking product shots. Do it again." "You aren't giving me anything to work with using this lighting setup. Do it again." Of course, as a retoucher I am down the line from the photographer, which would be like the printer telling designer how to design.

Having the retoucher tell the printer how to print would be like...

The art buyer telling the photographer how to photograph
The photographer telling the retoucher how to approach the retouching
The retoucher telling the designer how best to lay things out

...and all of those things happen every day. It should be expected that the person handing the job down the line will likewise pass on some instructions on how best to do the next step. The difference is that the printer is often some humongous organization with 20 people between you and the pressman. But the bottom line is that the client has a right to have their instructions respected, and if anybody along the line intentionally ignores them they run the risk of not getting paid.

J
___________________________________________________________________________

Re: K in commercial printing
Posted by: Michael Jahn
Sun Aug 17, 2008 3:07 pm (PDT)

@ Dr. Wyse - I HEART YOU.

related to Dans 'lock up da bums' position saying it is a crime to re-separate !

Well, in Dans world, my only outfit is jumpsuit orange - so, I am guilty as charged -- and un-remorseful.

Hi, My name as Michael and I am a alco-re-separator.

I - unlike great unwashed who belive in Santa Claus, the Easter Bunny and 50 cyan MEANs 50% tone on a printing plate, I will comfortably and proudly state for the record that if you want what you WANT to create a 50% Cyan patch in an image or in some desktop design too - and you WANT to print that out on your <insert your device here> - well, guess effin what, you need to take that 50% cyan set of pixels or object color fill call in you PDF - and convert it into your local <insert your device here> required colorspace so it can SIMULATE what 50% Cyan is SUPPOSED to look like.

if there any out there - even if you only have 4 fireing brain cells...please do at LEAST listen to this small next part;

THIS is why all STANDARD PDF/X files REQUIRE an output intent !

Otherwise, my Acrobat application can't tell my monitor how to display a SIMULATION of 50% cyan.

Otherwise, my RIP can convert that PDF "whatever colorspace" into a SIMULATION of 50% cyan

and guess what !

If I am printing that 50% cyan on my Gravure press (non-SWOP inks, non SWOP densities) - using type 4 inks, the ink that looks like it is Cyan-ish, well, it is NOT.

BTW - there is neither one standard "cyan" ink, (call us Sun Chemical on the phone as ask for a can of SWOP cyan ink, I double dog dare ya!)

sunchemical.com/ASSETS/PREV/FILE/na/publication/product/7_PrintEasyApp.pdf--yuesp, just add that http://and the www and enjuy !

-- no more that their one standard 'white' paper (as ink is translucent, well, that tends to 'color' what you might get even when you create a solid PATH of 100% cyan) - and (ahem) - since there is really no 'standard' way to set the water and ink ratio, then (ahem) that would mean that pressure and ink "density" would not be something that is simples as press of a big shinny red SWOP button...

sorry, there is no such thing as the tooth fair either (that was one of your parents SIMULATING the tooth fairy)

Ah - If this was that simple !

- we could make a SINGLE simple set of plates, and send them everywhere and hand them on every printing press, and say "hey, run your ink densities at XYZ, BUT IT IS NOT.

from page 7 of the SWOP guidelines;

Gray balance was traditionally defined as the CMY percentages needed to match the color of a 50% black ink tint, or the color of paper, but these definitions are too vague for today's ICC workflows. To avoid these ambiguities G7 defines an *arbitrary table* of CMY percentage 'triplets' based on the *generic 50c, 40m, 40y ratio*, and pre-defined a* and b* values for each triplet *taking into account the paper color.

*Which, btw if you might notice, if you CHNAGE THE PAPER COLOR, well, you would need to ADJUST that CMY ratio to simulate the intent or 50c 40m 40y .*

*I will finish by saying that this would ABSOLUTELY require me to resparate what every CMYK that walks into my door, using the output intent.

ie - if you say to me;

"hey mike, i don't CARE what device you are using (Xeorx iGen, for example) - please SIMULATE Fogra 39 so i can see exactly what i will get in singapore iwhen they do what they say they will do - well, i can.

Without that, I do not have any chance to meet the expectation of 'how to make that sheet look like some person in Toad Suck, Arkansas "intended'

Yes, there really is such a place - I actually worked with a woman from there - Barbara something. Worked with Mark Napier, tech QC guy all Walmart...

relevant real life story - Barb was who we presented our gravure press proofs for the following weeks Sunday Walmart color newspaper inserts. We would fly in every week whith our sheets to in Walmarts corporate offices in Bentonville, AR and mark them up - as they printed at 10 different printers, we would all meet at the Holiday inn in Springville (near Fayettville). When i say "we" I mean reps from RRD, Quebor, World Color, Reniger, Transcon - you see, Walmart had to print at 10 different locations to save money on FREIGHT - it was far to expensive to print in one single location and SHIP to all the newspaaper nation wide.

Now, In case you are going "gee, i had no idea that you did show press proofs in Gravure!" - well, yes, you do. Even though we digitally engraved that cylinder with our HelioKlischograph, hung all 8 cylindars on the press and fired it up, and even though they images were pulled from an archive ...

-- ( i amiagine you are scratching you head at this point - i imagine, saying "if you printed that image before, why would you need to modify it?)

Well - that is because there was that occasional (PUN intended here) "ghosting' issue, where we had an image on one page with a very dark red pile of towels 'inline' with another page that had some silverware of white china, so we would mark up the press proofs and actually either chemically etch or gently fill in one the other cylinders (in the case where the red was 'staved' - that would be the cyan cylinder)

So, without the ability to 'change' the separations, would would all print 40 million copies only to argue over that weeks invoice.

So, before anyone tells me the way it SHOULD be for your photograhers, until you stand press side and reach for the ink key settings, sorry, please sit quiety in that near the d50 viewing booth and keep your hands in your pockets..

I think we have have a clue as to what it takes to meet a buyers requirement, otherwise they would not pay us every week (there are some pretty big invoices!)

I do not tell you all to use 4 stop photography when lighting a scene (which is what you end up with BTW) - but I will indeed gently suggest that perhaps you might consider the idea of using a more "print like' (yes, icky flattened rGB) -- Adobe RGB -- if you intend to not be disappointed when you see your image on a press sheet.

And, just use SWOP to separate, weather we actually tell you (or choose to LIE to you since you are telling us NOT to and say okay) - we will always re-separate, because we have to to get you to pay for that invoice.

--
Michael Jahn
___________________________________________________________________________

Re: K in commercial printing
Posted by: Terence Wyse
Sun Aug 17, 2008 3:07 pm (PDT)

Hi John,

On Aug 17, 2008, at 11:13 AM, J Walton wrote:

That makes sense as a general rule of business, but what if the client
is expecting that they don't monkey with their files?

Files get "monkeyed with" or they do not print correctly. Examples: The moment a job gets imported into a workflow system, it will get altered including..
* All elements either get rasterized or interpolated to suit the output device (platesetter). So images all get resampled to something other than their original resolution.
* Either manual or auto-trapping occurs. This alters CMYK values.
* Like I mentioned previously, plate curves are likely to get applied, again, altering original CMYK values.

There's numerous ways customer files get altered, all in the name of making it "printable".

But the contract proof won't show misregistration, or accidental press
"adjustments."

Exactly. The proof shouldn't show misregistration unless that's what you expect to occur on press. If the contract proof is in-register, it's the pressman's job to match that. There's certain things that are simply a given...registration, no mechanical problems such as slur, etc. and a host of others.

Dan brought up a few pretty good examples of times when
a non-standard Black is essential to quality work. The cartoon example
WILL NOT look right, or like the contract proof, without respecting
the separation.

How so?

The 4C Black, or quadtone, example likewise WILL NOT
look right without a heavier black.

What do you mean by "not look right"? If the press is sufficiently under control from a gray balance standpoint, it wouldn't make any difference what GCR level was used in the image as long things like total ink amount, etc., was done correctly.

I can take an image and separate it using at least 5 different levels of GCR/K generation and they will all be visually and colorimetrically identical on a proof. And given GOOD PROCESS CONTROL on the press, they will print identically. But good process control isn't always a reality so we use GCR to improve the stability of the press. GCR separation techniques is a *pressroom * thing, not a creative tool.

I agree that this practice of
automatically re-separating is common, and when dealing with neophytes
probably improves the final product.

I suppose one question I would pose is - what does the typical device
link setup do if a CMYK file does not have an embedded profile? Will
it ignore the file or assume a source that is inaccurate?

Depends. You can have it any way you want. If the customer who supplied the images is color management-savvy and is embedding profiles, then possibly the device link workflow component should respect the embedded profiles, especially if the customer has provided a proof that reflects the embedded profile. In a case where profiles are NOT embedded, most CM workflow products will assign an assumed source profile before converting to a destination profile (no other choice really). But if I were to guess how most device link workflows are set up today, I would say most are set to discard embedded profiles and assign an assumed source. I don't think this is necessarily desirable, I think it says more about the level of education that's needed in our industry. I think as device link workflow products come into the mainstream, we'll see this level of knowledge/education increase rapidly.

I disagree with this point. I am a retoucher, but I tell photographers
how to photograph all the time *when they screw things up.* "No, dude,
you can't move the camera when you are taking product shots. Do it
again." "You aren't giving me anything to work with using this
lighting setup. Do it again." Of course, as a retoucher I am down the
line from the photographer, which would be like the printer telling
designer how to design.

I wasn't talking about retouchers, I was mainly talking about folks downstream telling the folks upstream how to do there jobs. If you're a retoucher and you're "midstream" and in a position to tell either designers or photographers how to do their job, go for it.

...and all of those things happen every day. It should be expected
that the person handing the job down the line will likewise pass on
some instructions on how best to do the next step. The difference is
that the printer is often some humongous organization with 20 people
between you and the pressman. But the bottom line is that the client
has a right to have their instructions respected, and if anybody along
the line intentionally ignores them they run the risk of not getting paid.

I agree that instructions should be followed of course. But telling a print shop how best to prepare plates for their press is not within a customer's domain in my opinion anymore than specifying what trap width they should use and other aspects that are printer/press- specific. In the case of job re-separation via device links, if the customer explicity states that they cannot alter the CMYK values in any way, then they shouldn't be surprised if the printer comes back with a surcharge due to the increased ink usage and cost that will occur on the job.

Regards,
Terry
___________________________________________________________________________

Re: K in commercial printing
Posted by: "Romano, John"
Sun Aug 17, 2008 10:15 pm (PDT)

I agree that instructions should be followed of course. But telling a
print shop how best to prepare plates for their press is not within a
customer's domain in my opinion anymore than specifying what trap
width they should use and other aspects that are printer/press-
specific. In the case of job re-separation via device links, if the
customer explicity states that they cannot alter the CMYK values in
any way, then they shouldn't be surprised if the printer comes back
with a surcharge due to the increased ink usage and cost that will
occur on the job.

Hi Terry

Well your exactly right, if anyone expects us to not touch their files its their Dime on press All Makeovers are AAs and would be Charged to the customer...

I can also tell you that EVERY single file that comes through our prepress GETS reseparated unless its TAGGED with the GRACoL2006_Coated1 profile.

Regards

John Romano
___________________________________________________________________________
.
Re: K in commercial printing
Posted by: J Walton
Sun Aug 17, 2008 10:15 pm (PDT)

On Sun, Aug 17, 2008 at 10:28 AM, Terence Wyse wrote:

Hi John,

I'll make you a deal. I'll agree to call you Terence if you will call me J. My brother's name is Jon, but he doesn't spell it with an "H."

:-)

Files get "monkeyed with" or they do not print correctly. Examples:
The moment a job gets imported into a workflow system, it will get
altered including..
* All elements either get rasterized or interpolated to suit the
output device (platesetter). So images all get resampled to something
other than their original resolution.

No, bad example. In the case of RIP'ing a file, that is something that is implied in the process of output. You *cannot* output an image without rasterizing it, but you can certainly output an image without changing the profile. If, as you put it later, the blacks are reseparated in order to save on ink costs, and you really want to use the RIP process as a comparison... then this would be like going into Illustrator and decreasing the object resolution because it goes through the RIP faster that way. Or res'ing down the client's high res files because you can save on storage costs. Those things are not a necessary part of the process, but they can save money.

* Either manual or auto-trapping occurs. This alters CMYK values.

This alters the CMYK values IN AN AREA THE SIZE OF A TRAP. Do you really believe this stuff?

* Like I mentioned previously, plate curves are likely to get applied,
again, altering original CMYK values.

This is the best example, but even this one is flawed in that it does not address the issues Dan brought up. Yes, the CMYK values are altered *in some pressrooms that use plate curves.* The entire structure of the black channel is not altered by press curves. If I choose to use a heavier black ink on a particular image a plate curve is not going to make it a skeleton black.

There's numerous ways customer files get altered, all in the name of
making it "printable".

I suppose every time a file is copied there are little ones and zeros that change as well, but that does not equal intentionally changing a file where such a change is not desired.

Exactly. The proof shouldn't show misregistration unless that's what
you expect to occur on press.

If you don't expect a measure of misregistration you're in for a disappointment. Let's use the cartoon example. You could use a skeleton black with lots of undercolor and the contract proof will look fine. But when it hits a newspaper it WILL NOT look fine. And you can talk about modern presses all you want, I've never seen a newsprint press reliably print in register. Now, you could say that I chose an extreme printing condition to illustrate the point, but that would be wrong. I didn't choose the example of a cartoon - Dan did a few days ago. It's a valid example.

Dan brought up a few pretty good examples of times when
a non-standard Black is essential to quality work. The cartoon example
WILL NOT look right, or like the contract proof, without respecting
the separation.

How so?

The paper moves around when traveling through the press, so it's not going to be in the exact same place when going through each ink unit. You know a lot about printing, I *know* you know that.

What do you mean by "not look right"? If the press is sufficiently
under control from a gray balance standpoint, it wouldn't make any
difference what GCR level was used in the image as long things like
total ink amount, etc., was done correctly.
I can take an image and separate it using at least 5 different levels
of GCR/K generation and they will all be visually and colorimetrically
identical on a proof. And given GOOD PROCESS CONTROL on the press,
they will print identically.

What you say is true, and is only humorous because of your next statement...

But good process control isn't always a
reality so we use GCR to improve the stability of the press.

Wow.

GCR separation techniques is a *pressroom * thing, not a creative tool.

From what I am hearing from you and the fellow in the orange jumpsuit, there's a lot of printers who pretty much automatically convert everything on the RIP without really looking at it too much. So if I spend hours retouching a single file, and the printer is going to spent, say, less than one second thinking about it... I'm gonna say that it would be better if *I* decide how the black generation goes. Just like *I* decide where the highlight should be, and *I* decide how much sharpening to apply.

Depends. You can have it any way you want. If the customer who
supplied the images is color management-savvy and is embedding
profiles, then possibly the device link workflow component should
respect the embedded profiles, especially if the customer has provided
a proof that reflects the embedded profile.

If a customer is color-management savvy, and has intentionally separated files for a particular reason, why would you change what they've done?

In a case where profiles
are NOT embedded, most CM workflow products will assign an assumed
source profile before converting to a destination profile (no other
choice really).

No, there is another choice. And that choice is to NOT reseparate their files automatically if there is no embedded profile.

But if I were to guess how most device link workflows
are set up today, I would say most are set to discard embedded
profiles and assign an assumed source. I don't think this is
necessarily desirable, I think it says more about the level of
education that's needed in our industry.

This is not just "perhaps not desirable," this flies in the face of what color management was supposed to accomplish. We were supposed to be able to move files back and forth and communicate output intent. We were supposed to accurately characterize devices so that intelligent conversions could take place. This is ignoring the output intent so that stupid conversions can take place.

I think as device link
workflow products come into the mainstream, we'll see this level of
knowledge/education increase rapidly.

This is where I guess I have to start sounding like Dan. We've been hearing that same line for the last 10 years or more, and I for one am *shocked* that things have regressed so far that embedded profiles are being discarded. The level of knowledge is not increasing if that is what is really going on in the industry. I've been away from the printing end of things for a few years - but even still it's hard to believe that, a) a printer would do that, and, b) a color management specialist would quasi-defend it.

I wasn't talking about retouchers, I was mainly talking about folks
downstream telling the folks upstream how to do there jobs. If you're
a retoucher and you're "midstream" and in a position to tell either
designers or photographers how to do their job, go for it.

So it's OK for someone "midstream" to tell the photographer what they need to do, but not OK for the same person to tell the printer what they need to do? The whole stream metaphor is now confusing to me.

I agree that instructions should be followed of course.

Good.

But telling a
print shop how best to prepare plates for their press is not within a
customer's domain in my opinion anymore than specifying what trap
width they should use and other aspects that are printer/press-specific.

No, they client is not telling the printer how to prepare plates for their press. They are telling the printer how to prepare the paper that eventually goes in the boxes. If the printer can automatically convert the file and get away with it, because their process control is such that they can close their eyes and hit "convert", then more power to them. What happens inside the press room is, IMO, none of the client's business. If the printer wants to print yellow first, great. If they want to print magenta first, I'm not going to tell them not to.

But if what they do inside the press room adversely affects the end product, then we've got a problem. By telling them that I have created a special separation on a particular image, I am telling them that I have put some extra time into making sure that image prints well. If they can do something to make it print better then I'm all for it. If they auto-convert, or assume I'm printing SWOP when my image says otherwise, and it looks worse... that's not good.

In the case of job re-separation via device links, if the
customer explicity states that they cannot alter the CMYK values in
any way, then they shouldn't be surprised if the printer comes back
with a surcharge due to the increased ink usage and cost that will
occur on the job.

That's where we get down to communication, where these debates usually end up anyway. If I am up-front about what I'm doing, then they need to honor their estimate and follow my instructions. If they are up-front about what they are doing and I don't like it, then I go somewhere else. But no, giving me an estimate and then charging me more because I said not to change my files is probably not going to go over too well.

Regards,

J Walton
___________________________________________________________________________

Re: K in commercial printing
Posted by: Dan Margulis
Sun Aug 17, 2008 10:15 pm (PDT)

Terry writes,

So you think the prepress/printer's job is to print exactly the CMYK
values supplied in the original file? Ever hear of plate curves (of
course you have)?

So you think that the best method of sound reproduction is the Victrola? Ever hear of CDs? (of course you have)

They "alter" the CMYK values before they go to
press. The on-plate values are no longer what was described or
specified by the original content. Would this constitute some sort of
legal violation of the intent of the content provider? If so, you'd
probably have to put about 80% of the printers in jail because a large
majority of them utilize plate curves to either help them print to a
specification or at least some notion of target print condition.

Under the legal principle of "Illegitimum non carborundum" (Latin for "no harm, no foul") nobody can recover damages unless there *are* damages (or the law permits otherwise, which it does not here). It would be easy for the printer to prove that not taking these steps would have made the job look worse. Also, the client's permission can be inferred from the fact that no reasonable client would ever deny permission for this.

That's very different from the printer unilaterally reseparating the file. There, damage is likely and can readily be proven, even to a judge and jury who are laypeople. It would be easy for a plaintiff to show that other printers, given the same files, could have achieved the desired result. It would be clear from examination of not just color correction but color management literature that authorities have long understood the need for maintenance of client-supplied black generation. Consequently, it would be easy for the plaintiff to demonstrate that he would never have agreed to permit the printer to reseparate, had he been asked.

We are only talking, here, about a case where the printer acted unilaterally without consulting the client (because the client can obviously give the printer permission to reseparate) and where inadequate printing resulted (because if the job turned out fine, nobody cares about the methodology.)

If the printer's unilateral reseparation of the files caused damage on press, it goes without saying that the printer is going to have to rerun the job properly at no charge. That's what happens when a job is misprinted, but ordinarily that's the limit of the printer's liability, since everyone knows that accidents occasionally happen on press.

When a printer willfully goes beyond the agreement with the client, and takes steps that he has no authority to take, thus damaging the final result, then it's a different story. I can think offhand of at least one job, a clothing catalog, that if the printer had reseparated it would have caused close to a million dollars' worth of lost sales (not to mention a six-figure printing bill) because the catalog, full of seasonal merchandise, could not have been released as printed. The defect would not have been apparent until the job hit press. If I were the catalog house and this had happened to me, I would have attempted to recoup all of this damage from the printer, on grounds of beach of contract.

If I were the lawyer representing the plaintiff, I would expect little difficulty in proving the specification of breach of contract. I could show ample literature indicating the proper way to separate such files and showing the consequences of separating them incorrectly. I would point out that the practice has never even been heard of until recently. I would give the files to another printer to demonstrate that the desired result could have been achieved if the black generation had been proper. I would, as indicated above, introduce not just my own writings but those of several color-management consultants indicating that the maintenance of black generation is critical and that high-end users would not accept random reconversions.

I would call experts on printing to testify that no printer, however incompetent, could possibly believe that a thin line printed as four halftones could look as good as one printed in solid black, or that a CMY grayscale image could look as neutral as one printed in black only. From which, it would be obvious that the plaintiff would never have granted permission for the reseparation to take place, and the printer had gone beyond the contractual mandate to do so.

Being a nice guy, if a bad lawyer, I would let it drop there, but some of my litigator friends like to add additional, imaginative grounds for complaint that are costly to research and to defend, like:

1) Plaintiff is a professional retoucher whose livelihood depends upon convincing others that he can get good-looking results from a variety of commercial printers. That fine lines are to be separated into black only where possible is extremely well known, as at least half a dozen different authors have offered this recommendation, usually with a graphic or two illustrating what happens if they aren't. There is at least one readily available case where a "professional" was publicly ridiculed at some length for having failed to take this elementary precaution. The defendant printer, by its actions, made it appear that the plaintiff himself had committed this error by submitting a file that many amateurs would know is defective. The printer committed this allegation into written form, with knowledge of its falsity, with knowledge that it might harm plaintiff professionally as well as holding him up to contempt, ridicule, and obloquy.

2) A copy of the client-supplied file in which there has been slight alteration to channels is substantially similar to the original. One in which the channels have been completely scrambled is (as can be proven by reference to the large body of Photoshop literature on channel structure and its ramifications) a different, derivative work. As much of what plaintiff supplied is copyrighted, the question may be raised by what authority the printer created this derivative work, which was then used to damage plaintiff. The law provides statutory damages of up to $30,000 per occurrence, or $150,000 per occurrence if the court rules that the copying was willful. As many jobs contain images copyrighted by several different individuals, there exists the possibility of multiple parties filing claims.

For these reasons, I would repeat the suggestion that any printer thinking about implementing such a workflow should consult his lawyer first. Almost any competent lawyer, IMHO, would suggest (particularly if somebody shows him this message) that perhaps it would be for the best if the printer got the client's permission before reseparating anything.

Incidentally, It should also be pointed out that litigators look not just for the obvious defendant, but for anybody who might have assisted them. For example, if a color management consultant had, for a fee, convinced the defendant printer to take actions that would breach its contract with plaintiff (such as by reseparating files without permission), that consultant might himself find himself a defendant, as it is readily demonstrated that color consultants should know the adverse consequences of reseparation. So, if I were a color management consultant trying to fob this workflow off on a printer, I would make sure I had a release indicating that I had warned them of the potential consequences.

Device link profiles to prepare a set of plates for a specific press
condition is simply a natural evolution from the common usage of plate
curves. They both constitute a form of "color management" for the
pressroom. Some may say device links are a different case since they
(can) alter GCR/K generation etc. but you can't get around the fact
that both methodologies are still a form of color management.

Similarly, drinking a half-liter of black rum and taking the car out for a nice spin is simply a natural evolution from enjoying a small snifter after dinner. They both constitute a form of self-medication. Some may say that the presence of a vehicle makes it a different case, but you can't get around the fact that both methodologies can be pleasurable in certain circumstances and not in others.

I would say that the printer's ONLY job is print the job in such a way
that it meets the customer's expectation.

Agreed. And when the customer provides a quarter-point rule separated to 0c0m0y100k, the customer's expectation is a crisp, solid-looking rule--no obvious halftoning, no misregistration. If the viewer can detect the slightest variation from this, with or without a loupe, the customer's expectations are not being met. Any "industry standards" with respect to what degree of misregistration is permitted are not relevant, since the possible variation in registration in the client-supplied file is zero, and the possible visible halftoning is also zero.

Similarly, if the client supplies a 4/c B/W in which the entire light half of the image is printed black only, then the client's expectation is that the printed result will be a perfect neutral to the extent that the black ink itself is neutral. By "perfect" I mean NO variation whatsoever from this neutrality at any density level, as measurable either by spectrophotometer or naked eye. Again, industry "tolerances" are irrelevant. The client-supplied file guarantees that the variation is zero. If the printer chooses to reseparate into CMY, the tolerance must still be zero, or the client's expectations are not met.

There are no printers on this planet who can meet either of these expectations, if they reseparate the file.

Dan Margulis
___________________________________________________________________________

Re: K in commercial printing
Posted by: J Walton
Sun Aug 17, 2008 10:52 pm (PDT)

On Sun, Aug 17, 2008 at 4:09 PM, Romano, John wrote:

I can also tell you that EVERY single file that comes through our prepress GETS reseparated unless its TAGGED with the GRACoL2006_Coated1 profile.

That's not exactly what I get from reading your website. An excerpt...

"At Acme, we understand the critical nature of your print communications. We partner with you to understand what you are striving to communicate, the critical nature that timing plays in your success and then passionately pursue this goal with relentless attention to satisfying your needs."

That line kinda makes it seem like if I am trying to communicate that I only want black in my image you'll understand and respect that. And if I need you to leave my image alone you will passionately and relentlessly satisfy that need. I wonder how the home page would look if you worked in "EVERY single file that comes through our prepress GETS reseparated unless its TAGGED with the GRACoL2006_Coated1 profile." Something tells me sales would go down...

For how long has this been a standard practice? I worked with Eric at Left Dakota a few years ago and he wasn't proposing anything so extreme as automatic reseparation. He was all about device link profiles, of course, but not this.

J Walton
___________________________________________________________________________

Re: K in commercial printing
Posted by: Stephen Marsh
Mon Aug 18, 2008 1:34 am (PDT)

At this point in the topic, I would really appreciate somebody doing a conversion for the list using a DLP and to save the result as a Photoshop readable TIFF or other readable image format. I can supply the source image if required. I think that it would be good to see a conversion that does not alter the black plate and one that does alter black, both to the same target/destination.

As a DLP would probably be out of the direct experience of many list members, I think that the above would be helpful. Does anybody have any good links that explain Device Link Profiles?

If I understand the DLP process (never used one myself), they are mostly found in RIPs, Photoshop does not support them (perhaps with a plug?). This "links" together two profiles of the same colour mode (CMYK + CMYK) and bypasses the PCS allowing direct conversion between colour numbers without requiring reseparation. Is this correct? The DLP remaps the source Cyan values into the destination Cyan values (and similar for MY and or K)?

I thought the major use of a DLP was to avoid altering the K plate (or am I getting confused with the old Imation CFM?). If one is going to reseparate all four channels using a DLP - what does this achieve over a regular ICC profile to profile conversion, where one can select the source and destination descriptions? Is it just that it bypasses the PCS (XYZ or LAB) and rendering intents? Is it time/workflow? Is this a faster process at the RIP than a standard P2P conversion?

And finally, if this is a backend process, I presume that the DLP can be applied to raster only, vector only and or both elements? Can it or the RIP be told to reseparate raster K but not vector K data, for example, or both or neither?

Sincerely,

Stephen Marsh
___________________________________________________________________________

Re: K in commercial printing
Posted by: Michael Jahn
Mon Aug 18, 2008 2:33 am (PDT)

@ Stephen

Many vendors - GMG, Alwan Color, CGS Publishing Technologies - all provide DLP support - this can PDF in PDF out and this can be done online, over the net, via FTP, watched hot folders - or of course, in a PDF workflow system like Apogee or Prinergy - or in a RIP.

It remains to be seen if Acrobat 9 can now support conversion, but my understanding is that it was supposed to and that is Adobes intent.

Everyone has run into errors and hue shifts and saturation death without DLP, which is due to the fact that LAB is not a very good conversion space after all.

Couple that with the fact that i can make a much simpler and faster "dog in-sausage out" DLP, I can convert A to B and B to A.

One of the REASONS that many printing companies are VERY interested of late is that you can re-separate and use less ink. Many color management companies come in, do an audit, pitch the President and CFO how much money they will save a year, then put togther a test, prove it uses more black and less CMY, then invoice them for what the market will bear...

If you are a Newspaper or magazine, this is more or less a no brainier, and the money if HUGE (savings).

If you consider these 'custom' separations make it harder to drift, then one can also add "save more sheets!" to that Value proposition.

on the down side, there may be pastey flesh tones and with no wera to drif, if you are out, you may need to replate, as ink key settings might not be enough.

I have simplified this as best I can here, and will say these are hard to test without going to press - which is expensive - even when you own the press.

Michael Jahn
___________________________________________________________________________

Re: K in commercial printing
Posted by: Terence Wyse
Mon Aug 18, 2008 8:52 am (PDT)

On Aug 18, 2008, at 4:26 AM, Stephen Marsh wrote:

If I understand the DLP process (never used one myself), they are
mostly found in RIPs, Photoshop does not support them (perhaps with a
plug?).

Yes, there are plug-ins available to support device links.

This "links" together two profiles of the same colour mode
(CMYK + CMYK) and bypasses the PCS allowing direct conversion between
colour numbers without requiring reseparation. Is this correct?

The most COMMON usage is conversions using the same color mode, mostly CMYK, but you can create device links for RGB to CMYK and other color modes, including multi-channel (CMYK+) color modes. This allows, for example, RGB conversions where R=G=B to be converted directly to the K channel.

The DLP remaps the source Cyan values into the destination Cyan values
(and similar for MY and or K)?

I thought the major use of a DLP was to avoid altering the K plate (or
am I getting confused with the old Imation CFM?). If one is going to
reseparate all four channels using a DLP - what does this achieve over
a regular ICC profile to profile conversion, where one can select the
source and destination descriptions?

The main advantage of a DLP is that you can selectively protect channels and even combinations of channels. You can also protect only 100% pure solids but allow the tints of pure channels to get "color- managed". This is not necessarily a *universal( feature of device link profiles in general, only of certain device link products. Device link profile products can range from <$500 to around $7,000 while device link *workflow* products (color servers) can range from a couple of thousand to over $30,000. Naturally, they are not all created equal.

So, yes, you can protect the K channel if you like but that generally means only PURE K such as K-only drop shadows and such. If the K is in a normal mix of CMYK, then it generally allows it to be re-separated.

There's even one product that I can think of that allows you to protect specific recipes of colors. You can create a list of CMYK recipes and either have them passed through with no change or you can specify the exact recipe for the "out-going" conversion. Pretty cool and powerful. In the example somebody gave of "cartoon" color where a limited number of CMYK recipes are being used, you could literally spec each in-coming CMYK value and alter the out-going CMYK values to whatever you wanted. And it doesn't matter if its raster or vector art.

Is it just that it bypasses the
PCS (XYZ or LAB) and rendering intents?

Yes, but this bypassing of the PCS is a big deal. Since you're "hardwiring" two profiles together, the gamut volume of each profile is known which is not the case where a PCS is involved (the PCS is used during the DLP building process but discarded after that). And, yes, it hard-wires the rendering intent as well. One advantage here is that the gamut-mapping/compression can be adjusted in a way that is unlike a standard rendering intent and can be a VERY powerful tool for dealing with out-of-gamut colors. Mind you, this is not true of ALL device links, just certain products.

Is it time/workflow? Is this
a faster process at the RIP than a standard P2P conversion?

Mainly a workflow thing. I can imagine a DLP is a faster conversion but I've never personally tested this. I'll try to remember to do this next time I'm at a customer that uses links.

And finally, if this is a backend process, I presume that the DLP can
be applied to raster only, vector only and or both elements? Can it
or the RIP be told to reseparate raster K but not vector K data, for
example, or both or neither?

Depends on the workflow but, yes, that is certainly possible. Typically, if the file is converted PRIOR to entering the workflow system (Rampage, Prinergy, Nexus, etc.), you can have these options depending on the device link color server product you're using. If the conversion takes place WITHIN the workflow system and *just* prior to making plates for the press, then quite often the distinction between raster and vector is gone (it's all raster at that point). But this also depends on the workflow system being used. Typical "ROOM" workflows rasterize fairly early on while "NORM" workflows will maintain an internal PDF that is not "rendered" until plates are made. Both methods have their pros/cons but I wouldn't get into that here.

Regards,
Terry Wyse
___________________________________________________________________________
.
Re: K in commercial printing
Posted by: "Romano, John"
Mon Aug 18, 2008 8:52 am (PDT)

Jon

I guess you can take whatever you want from our website....

Botom line is most of our customers use our profiles and if they don't we are expected to make the files look good at loose color stages.

Been in practice for many years.....SOP in our shop.

We know the value of having everything in the same color space before going to press. Anyone who claims this to be not relevant doesn't spend much time in a pressroom.

Clients sign off on our proofs, as long as we match them thats all that matters.

So what is the difference if we separate at the plate with a DVL ? Still changing your CMYK ?

Regards

John Romano
___________________________________________________________________________

Re: K in commercial printing
Posted by:remaleydan
Mon Aug 18, 2008 8:52 am (PDT)

Hi all, (a GCR supporter here) Taga some years ago did some work with GCR and noted that colors, (out of gamut) colors) that were separated with GCR had a closer delta E than images without GCR settings. The Delta's difference was reduced by 50% or more. In theory, the "hue shift" caused by y-m-c combinations will be eliminated by using a high black channel content. PIA/GATF studies claim that the number one reason for reprinting jobs wasn't WRONG color, or BAD color, it WAS INCONSISTENT color!!! True the dot gain has to be measured and controlled but that's for all colors - not just black. If black's printed wrong the image is too light or too dark. If y-m-c is printed wrong - we've loas gray balance and the hue. By the way, I no longer work for GATF so I'm happy to answer any questions anyone may have - here's my personal cell number 412.889.7643.
All the best - Color that is . . . . . .
Dan Remaley
___________________________________________________________________________

Re: K in commercial printing
Posted by: Terence Wyse
Mon Aug 18, 2008 8:52 am (PDT)

I'll respond to this last part since my eyes glazed over at all the
lawyer-speak.

Moving on...

On Aug 17, 2008, at 9:28 PM, Dan Margulis wrote:

Agreed. And when the customer provides a quarter-point rule
separated to 0c0m0y100k, the customer's expectation is a crisp, solid-looking rule--no
obvious halftoning, no misregistration. If the viewer can detect the slightest variation
from this, with or without a loupe, the customer's expectations are not being met. Any "industry  standards" withrespect to what degree of misregistration is permitted are not
relevant, since the possiblevariation in registration in the client-supplied file is zero, and
the possible visible halftoning is also zero.

Similarly, if the client supplies a 4/c B/W in which the entire light half of the image is
printed black only, then the client's expectation is that the printed result will be a perfect
neutral to the extent that the black ink itself is neutral. By "perfect" I mean NO variation
whatsoever from this neutrality at any density level, as measurable
either by spectrophotometer or naked eye. Again, industry "tolerances" are
irrelevant. The client-supplied file guarantees that the variation is zero. If the printer
chooses to reseparate intoCMY, the tolerance must still be zero, or the client's expectations
are not met.

There are no printers on this planet who can meet either of these
expectations, if they reseparate the file.

The scenarios you describe are not at all what a typical device link profile would do by someone that knows how to build a device link correctly. In probably 99% of cases where a device link is used for a *press* conversion as opposed to an inkjet device, pure channels are protected. So anywhere where pure C, M, Y, or K exist, they would typically get passed through the device link profile in tact with the possibly exception of an adjustment for dot gain, similar to what a plate/press curve would do. The same can be done with secondary colors like C+M, M+Y, C+Y, etc. but this is probably less typical than the first example since its generally *desirable* that, for example, "blues" get corrected for hue errors (too magenta or too cyan depending on wet trap characteristics of the press).

Device links are like the Swiss Army Knife of profiles. They can be created to do what a typical ICC conversion would do (completely re- separate) or they can be set up to do a minimal conversion, maybe only what a typical plate curve would do or maybe only an adjustment for total ink. This last thing is VERY typical of how many newspapers would utilize device links, to take typical "SWOP" separations and have them limited to, say, 240% total ink.

Regards,
Terry
___________________________________________________________________________

Re: K in commercial printing
Posted by: Terence Wyse
Mon Aug 18, 2008 6:23 pm (PDT)

On Aug 18, 2008, at 11:45 AM, Dan Remaley wrote:

True the dot gain has to be measured and controlled but that's for all
colors - not just black. If black's printed wrong
the image is too light or too dark. If y-m-c is printed wrong -
we've loas gray balance and the hue.

What an excellent point. The idea of "protecting" the separation by using an unnecessarily light amount of GCR/K seems to assume that the effects of a an out-of-control K are more disastrous than out-of-control CMY. The truth is quite the opposite as Mr. Remaley points out. So in the interest of "protecting" the image we've forced the pressman to have to control 3 ink/units more precisely than controlling one ink more precisely. I think we'd prefer that ALL inks are controlled equally well but I think most folks would agree that having an image print slightly too dark/light is preferable to one that prints with gray balance and colors out of whack.

Thanks Dan (Remaley) for pointing that out...again! Sometimes we forget the basics.

-Terry Wyse
___________________________________________________________________________

Re: K in commercial printing
Posted by: Dan Margulis
Mon Aug 18, 2008 6:23 pm (PDT)

John Romano writes,

Jon
I guess you can take whatever you want from our website....

J Walton has already asked once in the last 24 hours to be addressed by his correct name.
I hope not to see any further coincidental errors along this line.

Been in practice for many years.....SOP in our shop.
We know the value of having everything in the same color space before going to
press.
Anyone who claims this to be not relevant doesn't spend much time in a
pressroom.

Generally, prepress people have seen more *different* pressrooms than professional printers have. They consequently have a more rounded view of consequences of certain procedures.

For example, the case of how to construct drop shadows within CMYK images often comes up. Every experienced prepress person suggests that they should be built with much more black than would be the case in any normal method of separation, because they know that it makes the jobmore likely to print reliably.

For example, one list member wrote in 2004, "If you want to use the existing shadow I would use the channel mix method, removing the CMY and putting the shape into the k . Most times we go with a max of 20 to 25% k for Shadows.... Black only shadows are easy to control on press, When shifting color around on press the shadows stay neutral."

By doing this, the final CMYK shadow is of course not black only, since there is CMY underneath the K-only shadow. It's just that there's far more black, and far less CMY, than in similar areas that are not drop shadows.

As the list has gotten larger, we have had some confusion with similar names. For example, the person giving the above advice was coincidentally also named John Romano. Obviously it is not the same person, because the John Romano I am currently responding to states that for several years his company's practice has been to reseparate incoming CMYK images, rendering any such manipulations by the user futile. Presumably *this* John Romano thinks the other John Romano hasn't spent much time in a pressroom.

 I wish that the other John Romano were still here, because he would probably add, "but beware: certain printers nowadays will not honor your CMYK build, and your shadow will come out just the same as if you had never bothered to fix it at all. If you do build a shadow this way, and the printer reseparates it out and a color cast results, I think you should withhold payment until the printer reruns the job properly."

I can also tell you that EVERY single file that comes through our prepress GETS
reseparated unless its TAGGED with the GRACoL2006_Coated1 profile.

That is the type of simple statement that is needed more often. Once the client knows the printer's intent, and vice versa, everybody's life is made a lot easier.

Now that we know Acme Printing's intent, those interested in quality have an easy solution. Prepare the file properly, and then Edit: Assign Profile>GRACOL etc. Acme will never know the difference and will probably print the job properly in spite of themselves.

Warning, though. Be sure that this file goes ONLY to Acme, and nobody else, because the embedded profile is now wrong, and you don't want some other clueless printer to take it seriously.

Which brings up the question: if Acme is converting these files to a new profile, how does it know FROM what profile to convert, if the incoming CMYK file doesn't have one? (This question was put to Terry Wyse earlier, but he did not answer).

For many years we have been hearing such profile-free files described as "meaningless mystery meat" that cannot be deciphered without making unwarranted assumptions as to what the file preparer's intent was. I am pretty sure that the words "brain-dead", "utterly indefensible" and "a crap-shoot" have been used to describe the practice of guessing at what the missing incoming profile was supposed to be. These words of condemnation came from color management consultants, not from me, although I agree with the sentiments. And yet, if Acme is reseparating "EVERY single file", apparently it is reseparating those without profiles, which means that it must be arbitrarily assigning one, thus guessing at hue and luminosity values as well as black generation.

For around a decade I've been suggesting that people not embed CMYK profiles, because there have been many reports of accidental conversions (of course, I did not imagine that a printer could possibly do it on purpose). I made this recommendation because I thought that the lack of an incoming profile would be an foolproof method of assuring that a subsequent reseparation would not take place.

I hope, John, that you'll be able to say that Acme's practices with respect to CMYK files without profiles do not represent a validation of the old saw: "Anyone who thinks that a method is foolproof simply hasn't encountered sufficiently talented fools."

Dan Margulis
___________________________________________________________________________

Re: K in commercial printing
Posted by: Dan Margulis
Tue Aug 19, 2008 7:54 am (PDT)

Terry writes,

The scenarios you describe are not at all what a typical device link
profile would do by someone that knows how to build a device link correctly.

They were the specific scenarios presented by J Walton, and to which you incorrectly responded that good press process control could compensate for the lack of a proper separation. It was not immediately apparent, from said response, that you considered that the problem was irrelevant because you would in fact separate them the same way as a sane person's would.

In probably 99% of cases where a device link is used for a
*press* conversion as opposed to an inkjet device, pure channels are
protected. So anywhere where pure C, M, Y, or K exist, they would
typically get passed through the device link profile in tact with the
possibly exception of an adjustment for dot gain, similar to what a
plate/press curve would do.

This is better than a poke in the eye with a sharp stick, certainly. However, the remainder of the examples I gave you, which are just as compelling as reasons to avoid gratuitous reseparation, do NOT have pure values. I may be misunderstanding what you are saying and for that reason would suggest that Stephen Marsh's request for an example of before-and-after black plates be posted. If you are saying that you wish to retain the general character of the channel structure, fine, but I am not understanding you to say that. If anyone is saying that they wish to regenerate a new black channel that may be heavier or lighter than what the client intended, then I say that they are committed to third-rate color while demonstrating a complete disinterest in satisfying their client.

Another of your posts apparently misunderstands a further example. Another person had suggested constructing drop shadows in black only, and you responded that this would not be a problem for your DLP as pure blacks will be passed on as pure blacks. Unfortunately, this is not how drop shadows work. The shadow does not exist in a vacuum; it surprints what's underneath it. So the RIP is not getting pure black, it is getting (say) 20c15m10y40k, which is a much heavier black than any sensible algorithm would generate. As I understand your explanation, the DVL would treat this color as the same as (approximately) 50c40m30y0k and would reseparate both to the same value. Thus, the client's carefully thought-out method of making it impossible for color to vary on press is replaced by a system where he is at the mercy of the pressman, in a company so incompetent that it would actually allow such a conversion to occur.

The same can be done with secondary colors
like C+M, M+Y, C+Y, etc. but this is probably less typical than the
first example since its generally *desirable* that, for example,
"blues" get corrected for hue errors (too magenta or too cyan
depending on wet trap characteristics of the press).

Great. Now you're also trying to outguess what the client intended for hue.

In a reply to John Romano, I incorrectly stated that you had been asked whether you were going to assign a profile to an incoming CMYK file that lacked one, and that you had not answered. I now see that, yes, you did answer. You said you were going to take what your colleagues have been dismissing for years as "meaningless mystery meat", and decide that you *do* know its meaning, take a guess as to what the client intended, and base a subsequent conversion on that guess. So, while I cannot believe that you actually said this, the archived file indicates that you *did* say it, so I apologize for suggesting you had not answered.

Granted the above, I can do no better than to quote J Walton:

"This is not just 'perhaps not desirable,' this flies in the face of what color management was supposed to accomplish. We were supposed to be able to move files back and forth and communicate output intent. We were supposed to accurately characterize devices so that intelligent conversions could take place. This is ignoring the output intent so that stupid conversions can take place."

He's right. What this workflow appears to do, unless I misunderstand it is:

A) Given that the client, who may well be more expert than the printer, has given an emphatic command not to convert the file by a deliberate decision not to give you the source profile, you decide that the client wants you to convert the file.

B) Granted that there is no source profile, you take a wild guess as to what it might be, and then actually act on it, substituting your own judgment not just about channel structure but about color, for the client's.

C) You are willing to override client decisions that are well known to enhance printability by selective use of GCR, methods that often give 100% chance of success, and substitute a file that gives the pressman every possibility of screwing up.

OK. I've been in this business for 35 years or so, and I cannot with a straight face say that this is the stupidest workflow I've ever encountered, although it's right up there. What is difficult to comprehend is how somebody else can take this stinking morass and apply to it--with a straight face, yet--the term "color management". I give J the last word:

"This is where I guess I have to start sounding like Dan. We've been hearing that same line for the last 10 years or more, and I for one am*shocked* that things have regressed so far that embedded profiles are being discarded. The level of knowledge is not increasing if that is what is really going on in the industry. I've been away from the printing end of things for a few years - but even still it's hard to believe that, a) a printer would do that, and, b) a color management specialist would quasi-defend it."

Dan Margulis
___________________________________________________________________________

Re: K in commercial printing
Posted by: Dan Margulis
Tue Aug 19, 2008 7:54 am (PDT)

Terry writes,

What an excellent point. The idea of "protecting" the separation by
using an unnecessarily light amount of GCR/K seems to assume that the
effects of a an out-of-control K are more disastrous than out-of-control CMY.

Quite so. That is indeed the underlying assumption, an eminently correct one, although there are many other practical reasons that are perhaps more significant. This accounts for why skeleton blacks have been the standard at prepress houses from time immemorial, whether or not printers understand why they are a good idea.

So in the interest of "protecting" the image we've forced the
pressman to have to control 3 ink/units more precisely than
controlling one ink more precisely.

This is roughly equivalent to stating that a person with $35 in his pocket has more money than one with $100, because he has three bills while the other guy has just one. If the black channel is more than a skeleton, variation in it is more potent than the three CMY channels put together. This is why, as has been emphasized on this list many times, heavier GCR is workable ONLY at a highly disciplined, and quality-oriented printing firm.

Even in a vacuum, assuming that all errors are randomly generated, the difficulties caused by a slight mistake in a heavy black can only be equalled if there is a *huge* mistake in one of the CMY, or else simultaneous errors in two or more CMY channels. The odds of a problem with heavy GCR are therefore greater from the outset, but the odds get much bigger because of other, less obvious factors.

I think we'd prefer that ALL inks
are controlled equally well but I think most folks would agree that
having an image print slightly too dark/light is preferable to one
that prints with gray balance and colors out of whack.

The final phrase has been known not to be true for more than 100 years.

A heavier-GCR file does give better gray balance in the midtone. It does *not* prevent visible color shifts, because the client perceives color shift chiefly in pure colors, where black is not present regardless of separation method.

Example: easy picture, blue sky, green grass, white house. Assuming light GCR, the incompetent printer lays down the magenta plate too heavily, resulting in a purple cast. The sky and the house become purple and the grass loses saturation. Assuming heavy GCR and the same mistake in magenta, the resulting file is no better. The sky and the house are just as purple as before, because they have no black to control them. The grass should be marginally less desaturated, but either one will look unacceptable.

Make the house gray rather than white, and then the heavy-GCR version will look better, as the presence of black, which would be lacking in the light-GCR version, would compensate for the mistake in magenta. So, again: heavier GCR preserves midtone gray balance. It does *not* guard against color shifts. Most mistakes in the CMY wind up just as bad one way or the other. Mistakes in the black are *not* just as bad one way or another, they are much worse given a heavy-GCR file; in light-GCR files the CMY *do* ameliorate black mistakes to a considerable extent, but in heavy-GCR one the black does not do as good a job of ameliorating CMY mistakes.

Dan Remaley's statement that mistakes in the black make the image either lighter or darker needs to be amended to take account of black's ability to shut off color. We should say that, if there is a mistake, the choices are--in theory--image too light/colors too clean or image too dark/colors too muddy.

The unstated assumption is that an incompetent printer is just as likely to screw up any channel as any other, and just as likely to print too dark as too light.

In the real world, that's not how it works. Incompetent printers are FAR more likely-maybe five times as likely, in my experience--to screw up the black. They key in on the type, not the image, or on graphic elements such as the black backgrounds that the OP is using.

I don't know how many times I've gotten back plugged shadows (in light-GCR files, mind you) with the pressman's explanation, "We print PUNCHY here." Incompetent pressmen don't measure densities, and they don't pay much attention to images, because most of their clients supply garbage so they're used to the look. If the skeleton-black file doesn't have important shadow detail, then the WPPH pressman may actually make it look better. If it does, we'll get a second-rate result, which is a lot better than the fifth-rate result that would occur if he had been given a heavy-GCR version.

Even the most doltish pressmen can see when type is too gray, and they won't tolerate it. So, the possibility of black too light/colors too clean can be ignored in the real world. If there is a mistake, the black will be printed too dark. (AFAIK, there's no such pattern to the practices of incompetent printers with respect to CMY).

Therefore, the restated version of Dan Remaley's proposition is this. Of course, we prefer the job to be printed correctly, but if it is not, do we prefer to risk a change in midtone gray balance, or the considerably greater possibility of image too dark/colors too muddy?

Having researched the question, showing typical press variation, with various juries, I can give a real-world answer. It depends on the image. For typical press errors, the preference is for bad gray balance over too dark/too muddy by about 8 or 10 to 1--and that's *before* we consider that black errors are more likely than CMY. And it's also before we consider that incompetent pressman are much better able to adjust light-GCR images than heavy-GCR ones.

Since, however, there *are* images where we would prefer too dark to bad gray balance--bridal gowns, silver jewelry, and the like--the sensible practitioner uses heavy GCR in them.

Summary: for a disciplined, quality-oriented printer, there are advantages and disadvantages to making heavier GCR a standard. It is definitely not a no-brainer, either way. Suppose, however, we are constructing a profile for general use at printers who *claim* to meet a certain standard but are not known to be disciplined or quality- oriented. In that case, the considerations are:

1) If we are lucky and the printer turns out to be well above average, it won't matter whether we use light or heavy GCR. The results will likely be visually indistinguishable.

2) If the printer is not well above average we are vastly more likely to get a satisfactory result with a lighter black.

This decision on what style of black generation to put in a "standard" profile IS a no-brainer, for anyone with a modicum of common sense plus a rudimentary knowledge of CMYK file prep. Unfortunately, the standard profiles being fobbed off on us now are generally the products of color management consultants and printers, which groups are not known for possessing either. Consequently, we get these heavy-black "standard" profiles that are an excellent way of making sure that a bad printer will get the bad result that his skills merit. aAd then we get a lot of people saying that color management doesn't work, and then X-Rite goes down the toilet, and then color management consultants start crying the blues on the ColorSync list about what an unjust world we live in and how it is all my fault for pointing out that the average user would usually get more reliable results with profiles made by, er, other methods, as opposed to the cat food coming out in the guise of profiles from standards organizations.

Not wishing to end on a negative note, I give the last word to Ogden Rood, from 1876. Explaning why light GCR is preferable, he wrote: "We forgive, then, a partial denial of the truths of colour more easily than those of light and shade, which probably is a result of the nature of the optical education of the race. For the human race, thus far, light and shade has been the all-important element in the recognition of external objects; colour has played only a subordinate part, and has been rather a source of pleasure than of positive utility."

Dan Margulis
___________________________________________________________________________

Re: K in commercial printing
Posted by: "dacolorman"
Tue Aug 19, 2008 7:54 am (PDT)

Last Friday as I saw this subject develop all I could say to myself was "holy @#$%, here we go again". Now I come back to work on Tuesday and dreaded what I would find on the color theory group. But, I was not disappointed, there was plenty of @#$% to sift through. What can you do? You have to sift through it just to find those "jewels" of knowledge that are actually useful.

I have worked for and with many printers, and used many workflows (yada, yada, yada). Here is the bottom line as I have found so far.

If the printer claims to run to a certain standard, fine. Have him send you that profile and use it to control your proofs, but I wouldn't use it to separate my photos. Keep within the total ink limit (and the suggested K limit), but feel free to use a variety of custom photoshop settings to do the conversions. And, YES, by all means DO use a custom setting that allows you to steepen the K curve to bring out detail.

If a printer requires your files to be tagged, tag 'em with the supplied profile. But don't feel guilty for not using it to convert your files.

And what about plate curves? They are great. No, your 50% will not remain 50% by time it hits the plate - so what. The goal of the printer is to match your/their proof. If they have done their job correctly, those curves aid them in doing just that.

You can't stop a printer from re-converting your files - but the steep black curve you originally used to bring out detail in the shadows still did it's job, and now their job is to print that detail without introducing a color-shift (they probably will, but that is their problem).

Last time I sent a job to Hong Kong I was given a profile and all the total ink details I could have ever needed. Did I use their profile when converting - No, just for proofing. Did the job print ok? Yes. Did it match my proofs? Yes. Did they re-convert my photos? Dunno-don't care, the end result matched the proofs. I tagged my files with their profile and everyone was happy.

The above may be viewed as @#$% by some, but just maybe there is a "jewel" in there somewhere. So, as I personally keep sifting through the comments on "K in commercial printing" I am sure so find more "jewels" mixed in with the @#$%. But that's printing.

Andy Adams
___________________________________________________________________________

Re: K in commercial printing
Posted by: "John Romano"
Tue Aug 19, 2008 7:54 am (PDT)

On 8/18/08 9:16 PM, "Dan Margulis" wrote:

By doing this, the final CMYK shadow is of course not black only, since there
is CMY underneath the K-only shadow. It's just that there's far more black, and far
less CMY, than in similar areas that are not drop shadows.

Shouldn?t be confusing there is only one John Romano !

Also I never said I trashed black only shadows, what your quoting me on was how to BUILD them.

If your making the profile change you can simply put back the Black only shadow without any cmy. Depending on where your making the change, with a DVL or in Photoshop....only takes an extra step. I?m certainly not going to be worried about 25% K only shadow, so that will not be changed.

Our customers can get our profiles just by asking, not an issue....also something we have been doing for Years !

So if a file comes in without a profile....

Simple Just assign something that makes it look visually good on your calibrated and profiled monitor and Convert to your profile.

I probably wouldn?t be making suggestions on not to embed profiles on the premise that it will go to press untouched...You know....never Assume !

With all of the tools that are available EVERYONE is converting incoming files either in Photoshop or better yet with Color servers like Alwan color, GMG and Oris.....and you can set them up to convert from the embedded profile or if nothing is embedded to assume one and convert.

I would say its better to identify what is incoming so you know the conversion will be done correctly.

Regards

John Romano
___________________________________________________________________________

Re: K in commercial printing
Posted by: Terence Wyse
Tue Aug 19, 2008 7:54 am (PDT)

On Aug 18, 2008, at 9:16 PM, Dan Margulis wrote:

Generally, prepress people have seen more *different* pressrooms
than professional printers have. They consequently have a more rounded view of
consequences of certain procedures.

For example, the case of how to construct drop shadows within CMYK
images often comes up. Every experienced prepress person suggests that they should be
built with much more black than would be the case in any normal
method of separation, because they know that it makes the jobmore
likely to print reliably.

For example, one list member wrote in 2004, "If you want to use the existing shadow I
would use the channel mix method, removing the CMY and putting the shape into the k .
Most times we go with a max of 20 to 25% k for Shadows....
Black only shadows are easy to control on press, When shifting color
around on press the shadows stay neutral."

By doing this, the final CMYK shadow is of course not black only, since there is CMY
underneath the K-only shadow. It's just that there's far more black,
and far less CMY, thanin similar areas that are not drop shadows.

Dan, let me calm your fears. Any prepress dept. worth their platesetter is going to use device links in such a way that not only will pure K be unaffected by usually even pure CMY will be protected as well (by "pure" I mean any place where C, M and Y are uncontaminated by other inks). Things like K-only drop shadows, pure C, M and Y, and 100% solids are generally left alone. Not guaranteed mind you, but a good practitioner of device link profiles would be aware of these special cases and build their device link profiles accordingly.

As the list has gotten larger, we have had some confusion with similar names.
For example, the person giving the above advice was coincidentally also named
John Romano. Obviously it is not the same person, because the John Romano I am
currently responding to states that for several years his company's practice has been to
reseparate incoming CMYK images, rendering any such manipulations by
the user futile. Presumably *this* John Romano thinks the other John
Romano hasn't spent much time in a pressroom.

I wish that the other John Romano were still here, because he would
probably add, "but beware: certain printers nowadays will not honor your CMYK build,
and your shadow will come out just the same as if you had never bothered to fix it at
all. If you do build a shadow this way, and the printer reseparates it out and a color cast
results, I think you should withhold payment until the printer reruns the job properly."

I can also tell you that EVERY single file that comes through our
prepress GETS reseparated unless its TAGGED
with the GRACoL2006_Coated1 profile.

That is the type of simple statement that is needed more often. Once
the client knows the printer's intent, and vice versa, everybody's life is made a lot
easier.

Now that we know Acme Printing's intent, those interested in quality have an easy
solution. Prepare the file properly, and then Edit: Assign
Profile>GRACOL etc. Acme will never know the difference and will probably print
the job properly in spite of themselves.

Warning, though. Be sure that this file goes ONLY to Acme, and nobody else,
because the embedded profile is now wrong, and you don't want some other
clueless printer to take it seriously.

Which brings up the question: if Acme is converting these files to a
new profile, how does it know FROM what profile to convert, if the incoming CMYK file
doesn't have one? (This question was put to Terry Wyse earlier, but he did not answer).

Once again, easy answer. In such a workflow, the images would be proofed using something like GRACoL Coated1, SWOP Coated3, ISO Coated or whatever "standard" the shop has set internally. Whatever the proof is supposed to represent, that would be the assumed source profile for any DVL conversions downstream of proofing. Customer signs off on the GRACoL/SWOP/ISO/whatever proof and that is what gets delivered from the press regardless of whether DVLs or plate curves are used downstream. Color appearance is maintained which should be the goal between any off-press proof and the actual printed job.

For many years we have been hearing such profile-free files described as
"meaningless mystery meat" that cannot be deciphered without making
unwarranted assumptions as towhat the file preparer's intent was. I am pretty
sure that the words "brain-dead", "utterly indefensible" and "a crap-shoot" have been used to describe the practice of guessing at what the missing incoming profile was supposed
to be. These words of condemnation came from color management consultants,
not from me, although I agree with the sentiments. And yet, if Acme is reseparating
"EVERY single file", apparently it is reseparating those without profiles, which means
that it must be arbitrarily assigning one,thus guessing at hue and luminosity values
as well as black generation.

I guess I'm confused now. I think what you're saying is that it's PREFERABLE to supply tagged CMYK. That would certainly make a (good) prepress dept's job easier since now they KNOW what the intended color appearance should be. No more crap shot. In the case of untagged CMYK, the internal proofing standard more than likely becomes the assumed source since that is effectively what happens when sending untagged CMYK to a color-managed proofing RIP. The only exception might be the case where the client supplies a guide proof. Even in this case, the first question I would be asking is "what source profile was used to generate the supplied proof?" If there's an answer, then that's the new assumed source profile for the images. If there's no answer, then the customer's going to incur a lot of expensive AA's since the image will need to be color-corrected by the prepress operator using THEIR proofing system as the reference (wouldn't THAT be a step backwards to the bad-old days of Matchprints and Cromalins?).

For around a decade I've been suggesting that people not embed CMYK
profiles, becausethere have been many reports of accidental conversions
(of course, I did not imagine that a printer could possibly do it on purpose).
I made this recommendation because I thought that the lack of an incoming profile
would be an foolproof method of assuring that asubsequent reseparation would
not take place.

I hope, John, that you'll be able to say that Acme's practices with
respect to CMYK files without profiles do not represent a validation of the old saw:
"Anyone who thinks that a method is foolproof simply hasn't encountered sufficiently talented fools."

Well, maybe you're finding the first glimmer of savvy prepress depts. like John's that are color management aware and are taking the steps to provide color proofs and printed jobs that represent the customer's intended color. I'm not saying that everyone supplying images to a prepress/print shop should start embedding profiles in CMYK images, I'm saying it pays to communicate with your print provider and find out what their SOP is for supplied images, tagged or untagged. You correctly point at that if you find a case where they're converting, then by all means tag your image with the same source profile they are using or assuming in their workflow AS LONG AS you've either soft-proofed or hard-proofed your images using this assigned profile and it still represents your intent.

Regards,
Terry Wyse
___________________________________________________________________________
.
Re: K in commercial printing
Posted by: Dan Margulis
Tue Aug 19, 2008 10:05 am (PDT)

Dan Remaley writes,

PIA/GATF studies claim that the number one reason for reprinting jobs
wasn't WRONG color, or BAD color, it WAS INCONSISTENT color!!!

Those studies sound spot on to me. And, for each job actually rejected, there are probably 50 that *could* be rejected for inconsistency but aren't because the client doesn't have time to reprint them, or doesn't care about quality, or doesn't realize that the printing is unacceptable, or suspects it but doesn't want to fight with the printer.

The point of several of the posts to this thread is that we do not live in a perfect world. We have to prepare files for printing understanding that we may be dealt a poor or inconsistent printer. Yet we have to make the best of it, because if the bad printer prints badly, we get blamed for it. Consequently, we have to try to ensure that the bad printer prints reasonably well, which is an art form in itself. Making the assumption that any given printer is a good one is not helpful.

By the way, I no longer work for GATF so I'm happy to answer any questions
anyone may have - here's my personal cell number 412.889.7643.
All the best - Color that is . . . . . .

I'm sure I speak for everyone on the list in wishing you well in future endeavors and in hoping that you will continue to join in our discussions.

Dan Margulis
___________________________________________________________________________

Re: K in commercial printing
Posted by: "Rick Gordon"
Tue Aug 19, 2008 7:21 pm (PDT)

Could someone here get more specific about what a plate curve actually is, in comparison to a device-link profile? It would seem to me that ultimately it could be very non-destructive to original intent, if the following is true:

* The shape of the black curve might be squeezed or stretched in some way, but it's essential shape maintains its integrity. This would still supply an increase of black proportional to the increase of black that was originally intended in a black-only multiplied drop shadow, and it would also proportionately mirror different levels of GCR.

* The curves for the other three plates are adjusted for gray balance in relation to the K, based on a enough data to accomplish that task smoothly.

This would seem to be similar to the process of creating a profile, but maintaining the essential integrity of the K plate, which sounds much like what I understand a device link profile to be able to do.

The missing link here is how modifications in TIL would be implemented. But it would seem to be likely that such an approach would avoid many of Dan's concerns.

This post is not addressing the different matter of the printer changing the GCR, which sounds much more offensive to me.

Rick Gordon
___________________________________________________

RICK GORDON
EMERALD VALLEY GRAPHICS AND CONSULTING
___________________________________________________

WWW: http://www.shelterpub.com
___________________________________________________________________________ .

Re: K in commercial printing
Posted by: "dacolorman"
Wed Aug 20, 2008 6:52 am (PDT)

Plate curves (as I am accustomed to):
Each press typically has it own particular curves to help that press (and the others) maintain whatever the printer deems as their "standard".

For a gloss/dull/matte coated stock the printer knows (at least it is nice to think he
knows) what "ideal" densities to run to and what his dot gains should be (@50% C&M: 20%
gain or a dot value of 70%, Y: 18% gain or a dot value of 68%, K: 22% gain or a dot value of
72%). An initial press sheet is ran with plate curves that are linear (25=25, 50=50, etc).
From those sheets it is determined how much to change those linear curves to ones that
will help the press achieve the gains mentioned above (C&M: 20%, Y: 18%, K: 22%). So,
back on the press they go, using the new plate curves - they run to the "correct" densities,
and then measure the press sheets again. If they are still not at their "ideal" gains more
tweaking will be done until the plate curves give the press the desired results. The same
procedure is done for uncoated stock (running to different densities and gains of course).

The catch (and there are many) in this is having a press room that maintains strict control
of their presses/procedures. If there are any "cowboys" in the pressroom who don't take
too kindly to following procedures the plate curve will be meaningless.

Andy Adams
___________________________________________________________________________

Re: K in commercial printing
Posted by: George Machen
Wed Aug 20, 2008 10:21 am (PDT)

Nobody has yet mentioned that in Photoshop, every time one converts to CMYK - whether via some profile or via Custom CMYK - an estimated dot gain already gets built-in to the image, e.g., possibly 22% for web offset or around 17% for sheet-fed. A lot more for newspapers.

So those downstream plate curves are double-counting, and would render the image too light than intended, right?

To avoid such double-counting, if one separated in Photoshop with dot gain set to zero, one then often would encounter deleterious artifacts such as scum dots from making subsequent color corrections, so not separating with pre-estimated dot gain isn't an option.

- George Machen
___________________________________________________________________________ .

Re: K in commercial printing
Posted by: Henry Davis
Wed Aug 20, 2008 10:21 am (PDT)

On Aug 19, 2008, at 7:19 AM, John Romano wrote:

Simple Just assign something that makes it look visually good on
your calibrated and profiled monitor and Convert to your profile.

Once again, as has been the case in this discussion for many years now, there is an assumption that the proof will always be accepted. This is not always the case. There are plenty times when a designer/creative believes that adjustments are necessary.

When a conversion takes place downstream, it is very, very difficult for upstream folks to predict how his adjustments will go. I'm not talking press curves here - there is a relative, understandable, acceptable and predictable logic behind press curves that can be accounted for when making adjustments. But, other than the prediction given by a monitor display, there is no way that I know of to knowledgeably noodle a color that will ultimately be converted downstream.

There are those who say that the proof is the last word. At this point in this discussion, these same folks are also implying that the monitor is the last word. Which is it? Will you take me or your other customer's at their word that their monitors are trustworthy?

When a proof suggests the need, how is it that I am saving all of the trouble and expense of continuous proofing when my adjustments are - because of conversion - mere guesses? Is it that I am "forced" to be satisfied with the first proof - end of story? Sounds sort of authoritarian, eh? And just to touch on the ethical portion of the discussion: who should pay for the additional proofs when downstream conversions make it less possible for upstreamers to adjust colors accurately?

Henry Davis
___________________________________________________________________________

Re: K in commercial printing
Posted by: Terence Wyse
Wed Aug 20, 2008 12:51 pm (PDT)

On Aug 20, 2008, at 12:48 PM, George Machen wrote:

Nobody has yet mentioned that in
Photoshop, every time one converts to
CMYK - whether via some profile or via
Custom CMYK - an estimated dot gain
already gets built-in to the image,
e.g., possibly 22% for web offset or
around 17% for sheet-fed. A lot more
for newspapers.

So those downstream plate curves are
double-counting, and would render the
image too light than intended, right?

Not how it works George.

First off, don't get "dot gain" and "TVI" confused with actually ADDING 20% to the 50% dot. It's a combination of physical and *optical* dot gain, with most of it being optical gain.I should note that the "correct" term these days for this effect is "TVI" "Tone Value Increase" since this better describes what's happening. The older term "dot gain" implies a PHYSICAL spreading of the dots on press which is not the whole story. Sometimes the physical gain is only a couple of % while the lion's share of the "gain" is optical. The typical pressroom densitometer can't really distinguish between the two so it reports the total dot area it's measuring as ratio of solid ink density to the tint density via, typically, the Murray- Davies dot area formula. If you want to find the actual PHYSICAL gain, you have to use a video capture device such as a video plate reader.

Plate curves are added to give a press the intended or "target" dot gain. Typically with CTP, dot gain runs about 4-7% less than what was "standard" with film-to-plate imaging. To get back to the typical 18-22% dot gain, a "bump" curve of , say, 4% was added to the CTP plate to get back to this 18-22% dot gain target.

Terry Wyse
___________________________________________________________________________

Re: K in commercial printing
Posted by: Henry Davis
Wed Aug 20, 2008 12:51 pm (PDT)

On Aug 20, 2008, at 12:48 PM, George Machen wrote:

Nobody has yet mentioned that in
Photoshop, every time one converts to
CMYK - whether via some profile or via
Custom CMYK - an estimated dot gain
already gets built-in to the image,
e.g., possibly 22% for web offset or
around 17% for sheet-fed. A lot more
for newspapers.

So those downstream plate curves are
double-counting, and would render the
image too light than intended, right?

To avoid such double-counting, if one
separated in Photoshop with dot gain
set to zero, one then often would
encounter deleterious artifacts such as
scum dots from making subsequent color
corrections, so not separating with
pre-estimated dot gain isn't an option.

Nope. Without a press curve (linear printing), current and legacy separations wouldn't print correctly at all. When the file you deliver has gain accounted for, not having the gain on press is not a
good outcome.

It is possible to print linear or near linear, but both the file and the printing must be linear. This yields some very high quality results when done properly. However, this practice is not real-world.

Henry Davis
___________________________________________________________________________

K in commercial printing
Posted by: "Paul Foerts"
Thu Aug 21, 2008 6:56 am (PDT)

Hi Terry,

Let me correct you...

The terminology "dot gain" was/is used in a "film-workflow" with the dot (%) on the film as reference. "Dot gain" includes mechanical and optical "gains".

Mechanical: ink spread and "doubling" on multicolor presses. Optical: light scattering/absorbtion by the ink layers.

The terminology "Tone Value Increase" (TVI) was redefined by the introduction of computer-to-plate workflows. As there was no film value/reference available, plate values became the reference for calculating TVI. TVI includes "dotgain" + the difference (+ or -) between the "film-based" and the "C-T-P" workflow.

As it is not trivial to measure tone values on plates (different instruments may calculate different values), too much fiddling with plate curves may kill your "reference" and measuring TVI may result in meaningless values.

Plate curves may be useful for "pressroom balancing". Plate curves should however only be used as a last resort. Process control can only live with trusted references within every stage of the process.

Mixing tone values and densities with ICC definitions is not the best way to save color management credibility.

Paul
___________________________________________________________________________

Re: K in commercial printing
Posted by: Henry Davis
Thu Aug 21, 2008 9:12 am (PDT)

I am hopeful that this term will be hashed out on the list. I have taken to the convenience of equating TVI/Dot Gain in much the same way as DPI and PPI - knowing that they aren't the same thing. However, my assumption has been that TVI as discussed by professionals on this and other lists, is to be equated with dot gain in a real sense.

Having read pieces written with respect to TVI being measured from the G7 test form, I have assumed that what they were measuring was Dot Gain on a printed sheet, and calling it TVI. So, I am confused.

The "TVI problem" has been described as a problem of densitometry vs. colorimetric measurement for grey balance with the caveat that both could be used successfully for that purpose. Gee, I hope so.

Henry Davis
___________________________________________________________________________

Re: K in commercial printing
Posted by: Terence Wyse
Thu Aug 21, 2008 9:12 am (PDT)

Sorry Paul, but as far as I know the terms "dot gain" and "TVI" can be used interchangeably. "TVI" is the more current term since it's a more correct label for what's actually going on ("tonality" is increasing but "dots" aren't necessarily gaining).

If you could point me to reference that states otherwise, I'll gladly look into it but I've never heard of a distinction made between the use of the term TVI and it's exclusive use for CtP.

More comments inserted...

On Aug 21, 2008, at 9:28 AM, Paul Foerts wrote:

Let me correct you...

The terminology "dot gain" was/is used in a "film-workflow" with the dot (%)
on the film as reference. "Dot gain" includes mechanical and optical "gains".
Mechanical: ink spread and "doubling" on multicolor presses.
Optical: light scattering/absorbtion by the ink layers.

The terminology "Tone Value Increase" (TVI) was redefined by the
introduction of computer-to-plate workflows. As there was no film
value/reference available, plate values became the reference for calculating
TVI. TVI includes "dotgain" + the difference (+ or -) between the
"film-based" and the "C-T-P" workflow.

The "reference" as you call it is still the values in the original data/file whether you're talking film-based or CtP workflows. We linearize/calibrate film AND plates based on the values that were in the original file and then recorded to film/plate.

As it is not trivial to measure tone values on plates (different instruments
may calculate different values), too much fiddling with plate curves may
kill your "reference" and measuring TVI may result in meaningless values.

Sorry, I don't follow where "fiddling" with plate curves "kills" the reference.

Of course you need to use the right instrument to measure the plate (a video-capture based device such as a CCdot, ICPlate and many others).

Plate curves may be useful for "pressroom balancing".
Plate curves should however only be used as a last resort.
Process control can only live with trusted references within every
stage of the process.

Unless your press magically hits the intended specification or target in it's native state, either plate curves or device link profiles are *required*, and are not a last resort.

Mixing tone values and densities with ICC definitions is not the best way to
save color management credibility.

I honestly don't know what you mean by that statement. Different terminology for different processes.

Regards,
Terry Wyse
___________________________________________________________________________

Re: K in commercial printing
Posted by: Michael Jahn
Thu Aug 21, 2008 11:54 am (PDT)

@ Paul,

I think it was George Leyda, then retired from 3M (before it was Imation) and GATF consultant who may have coined the TVI, then later regretted this as he discovered that in some printing conditions it was actually TVD (Tone Value Decreace) - and tried very hard to change this term to TVD (Tone Value Difference) or TVV (Tone Value Variance)

This term !*** i think*** ! was coined far before CTP or CTP systems.

George was trying to come up with a term so the 3M Matchprint sales people could explain how "optical gain" worked in the actual physical Matchprint proofs - since you needed to 'simulate' how the film will actually print, you needed to compensate for the dot gain curve applied to the film

Since the film had the curve "built into" it - where a 50% tint was actually less than 50%-- using a combination of exposure settings and the colorant(s) and the lamination process and the stock (i never liked Pub Stock !) - well, you get to 'see' what you may actually 'get' on a press sheet

the Matchprint systems was popular for making analog proofs from film - we really did not like Matchprint much, since we were rotogravure and used now SWOP Type 4 inks, we used Cromalin.

Never mind that we never used the film for anything but making proofs (we digitally engrave the rotogravure cylinders using the helio from 9 track Mag Tapes) - again, long before CTP systems

In any event, two other small points;

1. if you set up Photoshop properly it will 'ad back' the dat gain and "simulate' what you are GOING to get, not the actual pixel values
2. If you go to the info pallet options, you can set this up so you can get both the actual pixel values AND the proof simulation values
3. This whole process is broken and approach should be tossed out the window anyway.

<!***! - this is my recollection, but do not count on me, i have sometimerz disease - sometimes i rememberz and sometimes i forgetz.

--
Michael Jahn
___________________________________________________________________________

Dotgain (was K in commercial printing)
Posted by: "Paul Foerts"
Fri Aug 22, 2008 10:29 am (PDT)

Hi Michael,

I wrote:
The terminology "Tone Value Increase" (TVI) was REDEFINED by the introduction of computer-to-plate workflows.

Dotgain is "by definition" the difference % between printed "dots" and "dots" on film where the film value is the reference. See SWOP booklets. Example: printed tone value 70% - film value 50% = dotgain of 20% or a tone value increase of 20% (this is no typo).

In a filmless workflom however, the "dotgain" terminology is misused completely as there is no "film" reference.

Tone value increase/decrease, as more general/basic terminology can be used with whatever reference you like. You may choose film, plate or whatever reference to compare with. For example a tone value increase between film and print, plate and print, system and inkjet print... of X% at 20%, 50% or whatever tone value. Without this reference, mentioning just "TVI of 15%" is risky business. Do you remember "The Seybold Report" mentioning "zero" dotgain for the "digital" Indigo press in their first preview?

In conventional platemaking (offset printing) you get bigger dots on the plate with negative working plates (US workflow) or you get smaller dots than those on the reference film strip with positive working plates (EU workflow). This is documented by GATF, FOGRA and System Brunner (starting in the seventies). Dot containing control strips were introduced to complement the continuous tone strips for exposure control.

The film values became the reference for the digital prepress systems. So 50 % system value would result in 50 % film value.

By the introduction of computer to plate, the film reference wast lost, so the plate had to be the reference...

Hence the notion "linear plates" where 50 % system value would result in 50% plate value.

There has been some discussion about the fact that CTP should be adjusted to the conventional plate values. Plate curves were proposed by FOGRA and GATF but printers did not follow this route... The CTP systems were calibrated to act linear on delivery. The 3 % difference in the midtones was no big deal... to most.

This "linear" thing was able to erase the 6-7 % difference at 50% tone value between the American and European/Asian printers. (Neg. vs. pos. working conventional plates).

Sadly, the "dotgain" thing survived. Even inkjet users echo "dotgain" all over the place...

TVI redefined?
If all the parties involved would take care of using precise references we would no longer have to repeat the same boring? sermon.

Reading "history" books about the (r)evolution of things in offset prepress would not hurt the "newbees". When "old" and "new" get mixed a "purple haze" is probably unavoidable...

Paul
___________________________________________________________________________

Dotgain (was K in commercial printing)
Posted by: "David Creamer"
Sat Aug 23, 2008 6:22 pm (PDT)

Way back when, I was taught that there were two types of dot gain: One for dot enlargement on film, and One for the ink spreading when absorbed by paper.

Of course, different industries and different regions/countries have different terms, as I've heard the paper dot gain called "ink spread" too.

Is "dot gain" a legitimate term describing ink spreading when the ink is absorbed by paper? If so, it is still a valid term when discussing printing.

Dave Creamer
I.D.E.A.S.
___________________________________________________________________________

Re: K in commercial printing/history of Matchprint
Posted by:Dan Remaley
Sat Aug 23, 2008 6:51 pm (PDT)

Hi Michael and all, a little 'clean up' here. . . for history's sake. . .
The reason TVI was used, is that CTP people tried to paint dot gain as a BAD thing - not so though.

If I gave you a file or tint that was 50% and told you not to increase it - how would you print it? There is always some gain.

CTP sales people did a great dis-service to the industry. They commented you could run higher densities 'because' your CTP - REALLY? I don't believe the press can tell the difference between aCTP plate and a Film based plate! It's the ink and paper that determines how much ink you can print on a sheet of paper - not the plate. Try printing 1.40 density on a uncoated sheet. The second statement is that you print 'sharper' because your CTP - not true - it's because the plates were produced linear.

Film was always produced "linear", a 50% in file became 50% in film. The difference was in plating. If a film based plate was made correctly 6-8 microns the 50% film based plate measured around 54%. When I started at GATF some 12 years ago, they burned their plates a 24 microns, making the plates read 60% on a film based plate. Not to good! I could follow CREO around all year and correct for their 'linear' strategy. You need 4 separate and distinct curves for Y-M-C-K. On to Matchprint, I loved Matchprint! It was consistant, repeatable, and was in register, the colorants were very close to ink on press. If the Matchprint had 22% midtone gain, you could print 22% on press and match it!
 The downfall of Matchprint was that the 1/4 gain was 22% also! This created a mismatch for the pressman (1/4 tones only gain around 16% and therefore printed "lighter") this was actually good for most images because the pressman would get accolades for making the press "print better than the proof"!. The lighter 1/4 tones printed, created a "better" contrast for the image. The downside is that 1/4 dependent pictures could never be matched (like jewelry). Arrival of DuPont Waterproof, it's gain was closer to the press throughout the tone scale - but it never did register well. Photoshop is still using a legacy, film based curve when you convert RGB>Lab>CMYK. The G7 method uses color metric data and adjusts the 1/4 and 3/4 tone curves (lighter). No magic here. Photoshop, which nearly everyone uses is based on legacy 20% (SWOP) midtone curves.

Dan Remaley
___________________________________________________________________________

Re: K in commercial printing/history of Matchprint
Posted by: Michael Jahn
Sun Aug 24, 2008 5:01 am (PDT)

@ Dr. Dan Remaley,

i LOVED this part where Dan wrote;

" Photoshop is still using a legacy, film based curve when you convert
RGB>Lab>CMYK. The G7 method uses color metric
data and adjusts the 1/4 and 3/4 tone curves (lighter). No magic here.
Photoshop, which nearly everyone uses is based
on legacy 20% (SWOP) midtone curves."

indeed. Photoshop is broken - no surprise everyone is frustrated.

LAB fails to compensate for the the helmholtz-kohlrausch effect.

ICC does all it wizardry in LAB, so everyone has chased their tail, then Eric Magnenson (sp?) -- Left Dakota -- finally said "enough" and took the baby step away from all that and began doing Device link profiles - but this approach is all a big rats nest.

meh.

--
Michael Jahn
___________________________________________________________________________

Re: K in commercial printing/history of Matchprint
Posted by: Terence Wyse
Sun Aug 24, 2008 9:14 am (PDT)

Hey Dan (Remaley), didn't think you was going to get off without SOME comment did you? I'm going to liberally snip and take things out of context. That's OK isn't it?

:-)

CTP - REALLY? I don't believe the press can tell the difference
between a CTP plate and a Film based plate! It's the ink
and paper that determines how much ink you can print on a sheet of
paper -not the plate. Try printing 1.40 density on a
uncoated sheet.

I agree with this (solid ink is solid ink regardless of the type of plate) except for the fact that inks have evolved as well so perhaps nowadays you CAN print with a slightly higher density at the SAME ink film thickness. Even the international/ISO standards dictate L*a*b* values for solids that generally result in densities about +.10 higher than legacy SWOP densities.
 
The second statement is that you print 'sharper' because your CTP - not
true - it's because the plates were produced linear.

As they should be, at least as a starting (process control standard) point.

Film was always produced "linear", a 50% in file became 50% in film. The
difference was in plating. If a film based plate
was made correctly 6-8 microns the 50% film based plate measured
around 54%. When I started at GATF some 12 years
ago, they burned their plates a 24 microns, making the plates read
60% on a film based plate. Not to good!
I could follow CREO around all year and correct for their 'linear'
strategy. You need 4 separate and distinct curves for Y-M-C-K.

If films were produced "linear", then why shouldn't plates be linear? That's just good process control. Sure, you're going to lose that extra 4-7% that you had when exposing plates from film but eliminating that "system" dot gain because of the extra generation is a good thing, no? With linear CTP, we're now seeing the true "native" behavior of the printing press as it was meant to be without the system dot gain upstream of the press. All it means is we need to adjust our separations going forward, just like they did when going from letterpress 4-color process to offset..or to gravure. In the meantime, we can simply use device link profiles to convert legacy separations to the new "linear" CTP. Just because we have all these legacy separations out there doesn't mean it's a good idea to cling to old working habits caused by some of the compromises inherit in the system, especially since it's quite simple to take the old seps and adjust them for the "new" way of printing. CTP has been in the mainstream now for at least 10-15 years so I think it's time we stop trying to make the CTP systems behave like the film systems of yore since those film systems have all but disappeared. Of course, I'm an advocate for something different than simply linear CTP.

On to Matchprint, I loved Matchprint! It was consistant, repeatable,
and was in register, the colorants were very close to
ink on press. If the Matchprint had 22% midtone gain, you could
print 22% on press and match it!
The downfall of Matchprint was that the 1/4 gain was 22% also! This
created a mismatch for the pressman (1/4 tones only
gain around 16% and therefore printed "lighter") this was actually
good for most images because the pressman would get
accolades for making the press "print better than the proof"!. The
lighter 1/4 tones printed, created a "better" contrast
for the image. The downside is that 1/4 dependent pictures could
never be matched (like jewelry). Arrival of DuPont
Waterproof, it's gain was closer to the press throughout the tone
scale -but it never did register well.

Which is why (I believe) that, generally speaking, color-managed inkjet proofing systems and to a lesser extent digital halftone-dot proofing systems are superior to the old film-based systems. Using color-management, we can have an inexpensive inkjet system emulate a press much closer than the legacy analog systems for perhaps 1/10th the cost.

Back in the day, how did you do an *uncoated* Matchprint? I know how we did it with Fuji ColorArt and Dupont WaterProof...we laminated the proof to a sheet of uncoated stock(!), believing that this was all we could do. Of course, the dot gain never changed and the solid "ink" density was the same as coated, we just laminated to a different background is all. Of course, with a color-managed inkjet system, it's as simple as using an uncoated press profile or dataset to get a very close match to an uncoated press sheet. Not perfect mind you, but damn close.

Photoshop is still using a legacy, film based curve when you convert
RGB>Lab>CMYK. The G7 method uses color metric
data and adjusts the 1/4 and 3/4 tone curves (lighter). No magic here.
Photoshop, which nearly everyone uses is based
on legacy 20% (SWOP) midtone curves.

This is where you need to do some catching up Dan (sorry). Photoshop is no more *based on* legacy 20% dot gain than my Dodge is *based on* driving conditions in Detroit Michigan. Photoshop's separation settings are based on whatever I happen to choose as my default CMYK working space profile. In other words, it's based on whatever I want it to be based on.

Now, if you're talking about what happens when I open up Photoshop's Custom CMYK engine and choose the SWOP (Coated) ink set, I think it was *you* who pointed out that the default midtone gain for Cyan was 24% and not 20%. In fact, if you look at the other inks (Magenta=20%, Yellow=20%, Black=20%) ONLY the Magenta is correct based on legacy SWOP values of C=20%, M=20%, Y=18% and K=22%. You'll find a similar discrepancy with ALL the ink sets in Photoshop (CS2 is what I'm checking it with so ) where the Cyan ink is about 3-4% above Magenta and all the other inks have the SAME midtone dot gain as Magenta. This makes Photoshop's Custom CMYK, if not outright wrong, at least VERY suspect in my book, to say nothing of the fact that the separation settings (Medium GCR, 100% K limit, 300% total ink limit) NEVER CHANGE when going from coated to uncoated to even newsprint. But of course, this same Custom CMYK is touted as being every bit the equal of high- end profiling packages. Nothing could be further from the truth. It's "broken" as Michael Jahn is fond of saying.

BTW, I missed seeing you last year at the GATF conference (I wasn't able to attend). Hope you're doing well.

Regards,
Terry Wyse
___________________________________________________________________________ .

Dotgain (was K in commercial printing)
Posted by: "Paul Foerts"
Sun Aug 24, 2008 8:14 pm (PDT)

Hi Dave,

Let's fresh up your memory a little :-) ...
Dot gain has a mechanical (process based) and an optical (substrate based) component. These components are not easily individually measurable. Densitometers/spectrodensitometers are used to capture both in one go. "Dot" (cfr. the "Dottie") meters (and other transmission densitometers) measure screened dots on film. (This is probably history to most)

For calculating dot gain in offset printing you need reference "film tone values" (film target) and the resulting "printed tone values". Adding a note about the workflow (negative/positive working plates) and screening specifics would provide you with enough information to be sure about the numbers. This is so by "definition" (not my invention/imagination), and accepted worldwide (a long time ago). Without honouring this convention offset dot gain numbers would have lost/have lost their meaning. Comparing "dot gain" numbers these days is comparing apples and oranges. So forget "dot gain" for communication.

If you would like to describe ink spreading, well... Are you talking about offset, rotogravure, screenprinting, inkjet printing flexo or some other process where ink is involved? For all these types of printing in combination with different substrates, specific terminology is used.

So my question: why would you use "dot gain" to characterise ink absorbtion or is it perhaps "light" absorbtion and ink spread you had in mind?

Paul
___________________________________________________________________________ .

commercial printing/history of Matchprint
Posted by: Dan Remaley
Sun Aug 24, 2008 10:39 pm (PDT)

From: Terence Wyse:

Hey Dan [Remaley], didn't think you was going to get off without SOME
comment did you? I'm going to liberally snip and take things out of
context. That's OK isn't it?

:Sure that's what I'm here for. . . .

I agree with this (solid ink is solid ink regardless of the type of
plate) except for the fact that inks have evolved as well so perhaps
nowadays you CAN print with a slightly higher density at the SAME ink
film thickness. Even the international/ISO standards dictate L*a*b*
values for solids that generally result in densities about +.10 higher
than legacy SWOP densities.

Yeah, and is this a "good" thing? Most people (not you of course) can't see a .10 difference!

And the major press manufacturer's aren't too thrilled either! Y-C-K don't matter much but Magenta look out! It has a tendency to "take off" and become troublesome when the density or gain change.

Why do you guys get excited about density, or Lab? Felix Brunner states "80% of color problems are caused by dots, only 20% are caused by density or solid ink colors! He's pretty sharp, and has been saying it for years.
 
Check out his website <systmbrunner.ch>. Don't let this hurt you, but he also says that Lab is not the way to control color on a printing press.

As they should be, at least as a starting (process control standard) point.

Yeah good for a reference - not for printing to.

If films were produced "linear", then why shouldn't plates be linear?
That's just good process control. Sure, you're going to lose that
extra 4-7% that you had when exposing plates from film but eliminating
that "system" dot gain because of the extra generation is a good
thing, no?

It doesn't matter what the gain is - if it's compensated for, along the process, it doesn't matter.
Plus the Pantone book, for years were printed to the same numbers based on film.

With linear CTP, we're now seeing the true "native"
behavior of the printing press as it was meant to be without the
system dot gain upstream of the press. All it means is we need to
adjust our separations going forward.

So now I'm going to adjust my separations for 'linear' and the guy down the street adjusts his for something else? I don't know about that one. I'm thrilled that no one touches Photoshop and just uses the default - what if their were 3 different programs to select from. Yes I know, color management. . .so what's the percentage of printers using color management? I know the percentage using process control - less than 10%!

Which is why (I believe) that, generally speaking, color-managed
inkjet proofing systems and to a lesser extent digital halftone-dot
proofing systems are superior to the old film-based systems. Using
color-management, we can have an inexpensive inkjet system emulate a
press much closer than the legacy analog systems for perhaps 1/10th
the cost.

The most critical printers (packaging) are hard pressed to give up their "dot proof".

The inkjets are certainly way above where they have been - but since we're "high tech", why not soft proof?

Back in the day, how did you do an *uncoated* Matchprint? I know how
we did it with Fuji ColorArt and Dupont WaterProof...we laminated the
proof to a sheet of uncoated stock(!), believing that this was all we
could do. Of course, the dot gain never changed and the solid "ink"
density was the same as coated, we just laminated to a different
background is all. Of course, with a color-managed inkjet system, it's
as simple as using an uncoated press profile or dataset to get a very
close match to an uncoated press sheet. Not perfect mind you, but damn
close.

Certainly agreed,

This is where you need to do some catching up Dan (sorry). Photoshop
is no more *based on* legacy 20% dot gain than my Dodge is *based on*
driving conditions in Detroit Michigan. Photoshop's separation
settings are based on whatever I happen to choose as my default CMYK
working space profile. In other words, it's based on whatever I want
it to be based on.

And most people (not you) 80/20 rule use the default! No?

Now, if you're talking about what happens when I open up Photoshop's
Custom CMYK engine and choose the SWOP (Coated) ink set, I think it
was *you* who pointed out that the default midtone gain for Cyan was
24% and not 20%. In fact, if you look at the other inks (Magenta=20%,
Yellow=20%, Black=20%) ONLY the Magenta is correct based on legacy
SWOP values of C=20%, M=20%, Y=18% and K=22%. You'll find a similar
discrepancy with ALL the ink sets in Photoshop (CS2 is what I'm
checking it with so ) where the Cyan ink is about 3-4% above Magenta
and all the other inks have the SAME midtone dot gain as Magenta. This
makes Photoshop's Custom CMYK, if not outright wrong, at least VERY
suspect in my book, to say nothing of the fact that the separation
settings (Medium GCR, 100% K limit, 300% total ink limit) NEVER CHANGE
when going from coated to uncoated to even newsprint. But of course,
this same Custom CMYK is touted as being every bit the equal of high-
end profiling packages. Nothing could be further from the truth. It's
"broken" as Michael Jahn is fond of saying.

Again it's consistency, if I know the legacy values are the same I can adjust.

I don't know anyone that separates for "uncoated" - the majority of the time they don't even know what paper it's going on! So I make plate curves to adjust the uncoated to the standard 20% gain - just like coated.

You used to run a high end scanner, what did you separate to? No 2 printers were alike, nor was their gain or gray balance. The "standard" years ago was based on midtone (50%) gain. None of the European presses have midtone patches, funny huh? How did they control the color, to the standard? - They didn't!

(KBA/Manroland/Heid.) Their bars have 20/40/60/80 from the film positive days. Komori got it right - GATF made the color bars for them!

Brunner's got it right and the rest of us are still trying. . . . .

Dan

BTW, I missed seeing you last year at the GATF conference (I wasn't
able to attend). Hope you're doing well.

Always great to debate. . . .
___________________________________________________________________________

Re: K in commercial printing/history of Matchprint
Posted by: Henry Davis
Mon Aug 25, 2008 12:43 pm (PDT)

On Aug 24, 2008, at 11:10 AM, Terence Wyse wrote:


Just because we have all these
legacy separations out there doesn't mean it's a good idea to cling to
old working habits caused by some of the compromises inherit in the
system, especially since it's quite simple to take the old seps and
adjust them for the "new" way of printing. CTP has been in the
mainstream now for at least 10-15 years so I think it's time we stop
trying to make the CTP systems behave like the film systems of yore
since those film systems have all but disappeared. Of course, I'm an
advocate for something different than simply linear CTP.

From the time I discovered the possibility of printing near linear, I have posted my testimony for the improved quality found when both the separation method and printing are coordinated for near linear results. But the problem with making this SOP is legacy files, and matching specific colors. Linear approaches just don't work in these cases.

Spot colors can be problematic anyhow, but if you change the gain you change the whole game - for Pantones and custom corporate colors. Whether using spot ink or simulating with process, changing gain is a barrel of fun.

Oooh, and what about the cases where several images on the same form that are separated differently. This comes up all of the time.

But I hear you, it would be nice to move the gain game towards the present capabilities of printing, but it will take a massive and coordinated super human effort - and this is almost laughable considering the history.

Henry Davis
___________________________________________________________________________ .

Re: K in commercial printing/history of Matchprint
Posted by: Terence Wyse
Mon Aug 25, 2008 1:44 pm (PDT)

Hi Henry,

On Aug 25, 2008, at 11:51 AM, Henry wrote:

From the time I discovered the possibility of printing near linear,
I have posted my testimony for the improved quality found when both
the separation method and printing are coordinated for near linear
results. But the problem with making this SOP is legacy files, and
matching specific colors. Linear approaches just don't work in these
cases.

I'm assuming you mean using linear plates and not that the printing (on paper) was actually linear. I've had folks try to do that (zero dot gain/TVI on paper) and it, uh, doesn't really work. "Linear printing" in terms of no physical dot spread will still "gain" (optically) in the 8-12% range. Try to remove the optical gain component to make it truly linear and you'll have a handful! Main problem with this is that it really mucks with the colorimetric distribution of the inks in L*a*b* space. Turns out the more-or-less "natural" gain of offset printing in the 12-20% range is colorimetrically correct. In other words, Gain is Good!

But getting back to what you were talking about, the "linear" approach could easily work with legacy files if A) the original target printing condition is known (via embedded profile or via assignment of a profile taken from a MatchPrint or whatever) or B) they were stored in a device independent space such as L*a*b* (not likely, I know).

Spot colors can be problematic anyhow, but if you change the gain you
change the whole game - for Pantones and custom corporate colors.
Whether using spot ink or simulating with process, changing gain is a
barrel of fun.

Spot colors are-what-they-are, colorimetrically speaking. As long as the ink in the can or the colorimetric definition hasn't changed, spot colors would be uneffected by dot gain...assuming we're talking about solid spot colors and not tints of spots or process simulations. Throw spot tints in the mix (there's NO definition on how those should print!) or process simulations and you've got a barrel of monkeys for sure. But, again, if you knew the target print condition of the process simulation or you had a tint of the spot color that you could measure with a spectro, you should be able to get through, assuming none of those monkeys gets away from you. :-)

Oooh, and what about the cases where several images on the same form
that are separated differently. This comes up all of the time.

That sort of gets back to a comment John Romano was trying to make before getting shouted down. This is where inserting device links into the workflow really helps "normalize" these kinds of issues. There's even workflow software out there that can INDIVIDUALLY ANALYSE separations on a page and "dynamically" adjust the separations for the best possible result.

But I hear you, it would be nice to move the gain game towards the
present capabilities of printing, but it will take a massive and
coordinated super human effort - and this is almost laughable
considering the history.

Thanks Henry, I guess I'm just tired of hearing the old saw about "legacy seps" when it's been over 10-15 years now that we've made the switch to CTP. Had we started even 5-10 years to make the switch, all those legacy images would probably be flushed out of the system by now! Glad to hear you're one of the progressive ones and recognize some of the benefits of casting off the old ways and taking a new approach.

Regards,
Terry Wyse
___________________________________________________________________________

Re: K in commercial printing/history of Matchprint
Posted by: Henry Davis
Mon Aug 25, 2008 8:53 pm (PDT)

On Aug 25, 2008, at 4:24 PM, Terence Wyse wrote:

I'm assuming you mean using linear plates and not that the printing
(on paper) was actually linear. I've had folks try to do that (zero
dot gain/TVI on paper) and it, uh, doesn't really work. "Linear
printing" in terms of no physical dot spread will still
"gain" (optically) in the 8-12% range. Try to remove the optical gain
component to make it truly linear and you'll have a handful! Main
problem with this is that it really mucks with the colorimetric
distribution of the inks in L*a*b* space. Turns out the more-or-less
"natural" gain of offset printing in the 12-20% range is
colorimetrically correct. In other words, Gain is Good!

Thanks for the post Terry. When I wrote "near linear" it was with reference to the measured dot gain on the press sheet. What I meant by "near linear" is the least gain possible - both physical and optical. The differences in traditional ink and those that HP Indigo uses was the initial cause for my interest in linear - a decade or so ago when the first Indigo arrived on the scene. I'll take your word for it that zero optical gain is problematic - it seems that there are snags on every front. I suspect that the colorimetric distribution of inks with regard to Lab was of keen interest for the Indigo developers. I understand that Indigo "ink" and traditional ink is and apple/orange comparison. But, just as you have pointed out, advances in ink are not out of the question, and so it could be that an ink will be developed that has completely different characteristics than are currently available. The Indigo was an example of this.

But getting back to what you were talking about, the "linear" approach
could easily work with legacy files if A) the original target printing
condition is known (via embedded profile or via assignment of a
profile taken from a MatchPrint or whatever) or B) they were stored in
a device independent space such as L*a*b* (not likely, I know).

These are big ifs when it comes to todays printing scene. It would be nice, though, if everything came to shop ready to print. It would also be highly irregular.

Spot colors are-what-they-are, colorimetrically speaking. As long as
the ink in the can or the colorimetric definition hasn't changed, spot
colors would be uneffected by dot gain...assuming we're talking about
solid spot colors and not tints of spots or process simulations.

Throw spot tints in the mix (there's NO definition on how those should
print!) or process simulations and you've got a barrel of monkeys for
sure. But, again, if you knew the target print condition of the
process simulation or you had a tint of the spot color that you could
measure with a spectro, you should be able to get through, assuming
none of those monkeys gets away from you. :-)

Yes sir, the solids are what they are and as such are more well behaved. But I was under the assumption that tints can't be reconciled in a colorimetric manner. I wasn't aware that various tints of the same solid could be measured individually and each given a cmyk build based on that measurement - and that it would actually work. I had always figured that 1) colorimetric matching was not a roll that profiling was designed for and 2) the measured hue differences in each tint would be problematic.

If there is a way to measure a desired color in such a way as that measurement returns the cmyk values for that color, I would be very keen on learning how to do that.

That sort of gets back to a comment John Romano was trying to make
before getting shouted down. This is where inserting device links into
the workflow really helps "normalize" these kinds of issues. There's
even workflow software out there that can INDIVIDUALLY ANALYSE
separations on a page and "dynamically" adjust the separations for the
best possible result.

These tools may be the framework for a shop's SOP, but they won't eliminate guesswork and assumptions. Sure, these tools are helpful, and perhaps necessary in some cases. But until every aspect of printing operates under one grand unified standard of practices, it will still involve guesswork.

"Normalize" is the operative word here, and while it is helpful for a certain percentage of jobs, there are nonetheless cases where solutions are not so close at hand. I am not a nay-sayer here - good tools are good. I believe that the low-level shout-down was more or less because the same ground has been plowed a number of times. The same old problems don't just go away, and despite some good tools, there are still problems that require more effort. The big red help button hasn't been invented yet - at least not the one that you can always get away with using.

Thanks Henry, I guess I'm just tired of hearing the old saw about
"legacy seps" when it's been over 10-15 years now that we've made the
switch to CTP. Had we started even 5-10 years to make the switch, all
those legacy images would probably be flushed out of the system by
now! Glad to hear you're one of the progressive ones and recognize
some of the benefits of casting off the old ways and taking a new
approach.

Yep, it is tiring, but if there were a perfect solution it would go away without a negative word to be heard. I am progressive, but not to a fault. Most print buyers, whether they are small-time or large corporate accounts, could care less how things happen in the print shop. It doesn't matter how or what they submit for files, it just isn't their concern how many resources are spent to make their jobs print like they want them to look. Very few print buyers are even the slightest bit interested in changing anything that is done on their end - even if it can be demonstrated to be to their advantage. And, it isn't always a good practice to deliver ultimatums. One's insistence that a client submit files that conform may result in there being one less problems to solve, or perhaps one less client. Legacy files aren't going away any quicker than are the attitudes of creatives and print buyers changing.

Someone ought to write some print shop lyrics set to the tune "Imagine".

Henry Davis
___________________________________________________________________________

Dotgain (was K in commercial printing)
Posted by: "David Creamer"
Mon Aug 25, 2008 12:44 pm (PDT)

Let's fresh up your memory a little :-) ...
Dot gain has a mechanical (process based) and an optical (substrate based)
component. .... (This is probably history to most)

Thanks for the refresher, but I'm familiar with this already. Unfortunately, that wasn't my question, and I didn't feel it was necessary to waste the electrons for something I wasn't asking about.

If you would like to describe ink spreading, well...
Are you talking about offset, rotogravure, screenprinting, inkjet printing
flexo or some other process where ink is involved?
For all these types of printing in combination with different substrates,
specific terminology is used.

Offset (SWOP) for magazine printing (don't remember the actual paper specs).

So my question: why would you use "dot gain" to characterise ink absorbtion
or is it perhaps "light" absorbtion and ink spread you had in mind?

As I said, if you read my post, you would have noted that it was a term that I was *taught*. I have not problem correcting my terminology if necessary, but I was asking if dot gain on paper is an incorrect term. As ink is absorbed, does the ink "dot" not enlarge or gain size? Or is it incorrect in any usage for describing in ink spread?

Dave Creamer
I.D.E.A.S.
___________________________________________________________________________

Re: K in commercial printing/history of Matchprint
Posted by: Dan Margulis
Tue Aug 26, 2008 8:27 pm (PDT)

Terry Wyse writes,

Just because we have all these
legacy separations out there doesn't mean it's a good idea to cling to
old working habits caused by some of the compromises inherit in the
system, especially since it's quite simple to take the old seps and
adjust them for the "new" way of printing. CTP has been in the
mainstream now for at least 10-15 years so I think it's time we stop
trying to make the CTP systems behave like the film systems of yore
since those film systems have all but disappeared.

The concept is fine but it's easier said than implemented. "Been in the mainstream" is not the same as being so dominant that the competition "has all but disappeared." The first is, as you say, 10 years at least. The second development is much more recent. That means there was a period where there was a serious inconvenience if CTP did not appear to the client as if it were a film-based system. Consequently, the development of CTP was "it must behave as film-to-plate did." Now that we don't have to worry about film-to-plate very much, we could theoretically make the changeover and get somewhat better quality, but it would be a difficult sell with this much history behind doing it the other way.

Photoshop is no more *based on* legacy 20% dot gain than my Dodge is *based on*
driving conditions in Detroit Michigan. Photoshop's separation
settings are based on whatever I happen to choose as my default CMYK
working space profile. In other words, it's based on whatever I want
it to be based on.

That's right. And if you make a good choice then you will get good-looking, predictable results, and if you make a stupid choice you'll get ugly-looking results that nobody will like. Some years ago the likes of Newt Gingrich and Bob Barr made the point that it is extremely reprehensible and disgusting for politicians to indulge in extramarital dalliances--provided that the politician is a Democrat, as otherwise boys will be boys. While the quoted material above is correct as far as it goes, what happens next seems to suggest the Gingrich-Barr approach.

As I read it, you are suggesting that if a person loads a stupidly made third-party profile and get a stupid result, well, boys will be boys, but if the same person loads a stupidly made Custom CMYK profile and gets the same poor result it proves that the whole system is broken and anybody who uses is it a troglodyte.

Now, if you're talking about what happens when I open up Photoshop's
Custom CMYK engine and choose the SWOP (Coated) ink set, I think it
was *you* who pointed out that the default midtone gain for Cyan was
24% and not 20%. In fact, if you look at the other inks (Magenta=20%,
Yellow=20%, Black=20%) ONLY the Magenta is correct based on legacy
SWOP values of C=20%, M=20%, Y=18% and K=22%.

No, it was *me* who pointed this out in PP1E (1994) and every edition since. Every graphics program I worked with has its share of stupid defaults that knowledgeable users change. Simultaneous with the last major change in Custom CMYK (1998, Photoshop 5), there came a major change in RGB handling, to wit, all incoming tagged RGB files would immediately, and without warning, be converted to sRGB. Personally, I think that any change in colorspace wihout a warning is a stupid and quality-damaging action, and accordingly I condemned this unrevealed RGB-to-RGB conversion at the time, and now. In the interest of logical consistency, I condemn (as even more damaging, which they are) the unrevealed CMYK-to-CMYK conversions being advocated in various quarters in the last few days on this list.

For similar logical consistency, if you are to suggest that Custom CMYK is to be avoided because of its silly defaults, you must either 1) agree that converting all incoming RGB files to sRGB without warning is a good thing; or 2) caution Photoshop users that RGB is to be avoided, and all work should be done in LAB or CMYK.

This makes Photoshop's Custom CMYK, if not outright wrong, at least VERY
suspect in my book, to say nothing of the fact that the separation
settings (Medium GCR, 100% K limit, 300% total ink limit) NEVER CHANGE
when going from coated to uncoated to even newsprint.

Examination of the Adobe-supplied sheetfed uncoated v2 profile vs. the web uncoated v2 profile reveals that they are one and the same. These two profiles (or one profile, depending upon how you look at it) are, AFAIK, basically machine-generated. Again, for logical consistency, it would be helpful to know whether 1) you agree that sheetfed uncoated and web uncoated printing conditions are exactly the same and that one profile suffices for both; or 2) machine-generated profiles are, if not outright wrong, at least VERY suspect in your book.

But of course,
this same Custom CMYK is touted as being every bit the equal of high-
end profiling packages.

Good grief. Speaking of logical consistency, if you must attribute a ludicrously extreme position to me, you have to decide which of the two contradictory ludicrous extremes to use. It is OK to accuse me of being far to the right of Dick Cheney or far to the left of Ted Kennedy, but it is difficult for any one to be both simultaneously.

Since 1998, I have pointed out what is now obvious, that high-end users would never accept a system of noneditable profiles, and that if Adobe was serious about color management it had to provide *at least* as much functionality as Custom CMYK has had since 1992 (revised 1998) in handling third-party profiles. Unlike some, I do not consider it overly burdensome for programmers to produce something at least equivalent to what was available 15 years ago, and I do not consider it unreasonable for users to request that a professional-level program have at least as much mission-critical functionality as it did in 1992. I have therefore repeatedly used words like "rudimentary" and "bare-bones" to describe the Photoshop features that would be needed to make third-party profiling acceptable to high-end retouchers, and I have further said that a competent programmer would be able to code such features in two days.

Less than a year ago, this list was deluged with comments to the effect that my suggestion for "rudimentary, bare-bones" editing actually meant something "every bit the equal of high-end profile packaging" and that I was therefore demanding that the Photoshop team devote months of programming time to match all features of a product that, we were constantly reminded, costs $2,500 separately.

But now, it seems from the above that I was actually saying that Custom CMYK was "every bit the equal of high-end profiling packages" from the get-go. So, you need to clarify *which* preposterous extreme you claim is my view. When I say that the Photoshop team needs to devote a couple of day's of one person's time to implement some rudimentary capability, does this mean that 1) the Photoshop team needs to devote tens of millions of dollars' worth of time to get what I want; or that 2) They already have what I want so that they don't need to do anything at all?

The *true* translation of what I am saying is, I believe, known to every member of this list except the color management consultants. It is that, given a choice of how to drive to work, I prefer a 1992 Ford Escort that runs to a 2009 Lamborghini that has no engine. Put any kind of engine into the Lamborghini, original equipment or not, and my preference will change. Bellicose rhetoric to the effect that engineless cars are the wave of the future, that it's all a matter of educating the driver, that we've finally turned the corner in their adoption, and that they use less gasoline does not impress.

I remind you, that the principal losers to the Photoshop team's intransigence are not skilled retouchers, but color management consultants, whose insistence on the take-it-or-leave-it profiling approach dooms them to the latter, as ten years of experience has brutally shown. All it would take to solve the OP's problem is this thread is the ability to generate a proper black from within the profile that the printer recommends. With that, there will be no problem getting those serious about quality to come aboard. Without it, there is no option but Custom CMYK.

To judge by the recent ColorSync list obituary for X-Rite, there is considerable bitterness amongst consultants about, shall we say, the limited appreciation the market shows for their efforts, and particularly about influential people who claim to be able to get better quality without it. While this may have been a defensible position a decade ago, after the intervening experience it takes more than an ordinary amount of stubbornness not to give in and give the experts what they want.

If the business climate is really that dire in the color management community, perhaps you could get some of your colleagues together and hold a bake sale, donating the revenues to the Photoshop team to hire a high school student or two to help them part-time if coding a simplified profile editor is too difficult for existing personnel. That way, instead of trying to ram a quality downgrade down people's throats, you all could sit back and relax as users and Photoshop commentators made the migration that would finally make sense from a quality POV. And you wouldn't have to go to all the mental effort of conjuring up nonsensical views of opponents in order to condemn them.

Dan Margulis
___________________________________________________________________________ .

Dotgain (was K in commercial printing)
Posted by: "Paul Foerts"
Tue Aug 26, 2008 8:27 pm (PDT)

Dave,

Sorry for the wasted atoms...
As english is not my native tongue and I did not know you were familiar with printing curves, tone value increase etc., I was unsure about the intention of your question.

"Dot gain" or "dot spread" is terminology from the "film" era. This terminology is no longer used in the latest SWOP specifications booklet.

Concerning: Offset (SWOP) for magazine printing. If you would use a loupe, a microscope or a video based magnification device, you could examine the printed image and try to measure the surface of the printed "spots". You could compare this information with the surface of the same "spots/dots" on the printing plate and come to the conclusion that you have a surface gain or loss... Maybe you can observe minor or heavy distortion caused by doubling or some other mechanical error.

Ink (=pigments, resins, oils...) penetrates more or less into the substrate, depending on the properties of the coated surface. As printed inkfilms are thinner than 1,5 micons, chances of "spreading" are extremely limited.

Conclusion: ink spread is a neglectible factor in "surface grow".

A better way to evaluate printed halftones is by using a densitometer. This way, tone values and differences (plate ref.) can be measured easily. Midtones will show the highest tone value increase. Finer screens will show more tone value increase than coarser ones. Line screens show the lowest and irregular screen shapes the highest TVI. Very fine AM screens and stochastic (FM) screens have their own characteriscics.

Presses with deficiencies are at the origin of extreme "gains" caused by "doubling".

Thicker inkfilms/darker colors can absorb more reflecting/scattering light from the substrate than thinner inkfilms/lighter colors resulting in more or less TVI.

Paul Foerts
___________________________________________________________________________

Re: K in commercial printing/history of Matchprint
Posted by: "dacolorman"
Wed Aug 27, 2008 6:41 am (PDT)

Years ago I was all about keeping the final result at the press to 12%-16% (since that is what the CTP allowed me to do). But, all was not well for a number of reasons: supplied seps based on 20%+ gains, correct densities on press still resulted in a sheet that was "starved" for color, etc. So, I chose not to reinvent the wheel, but to work with the wheel I was given (by others much smarter than myself).

I could have (finally) gotten everything to work, but for what? For less gain? That only amounted to a "savings" of 5%-10%. That may sound like a lot, but if everything is and has been built to work (separations, plates, press, etc.) with higher legacy gains the results still "looked" good (because everyone is compensating for the well know "legacy" gains).

Once I went back to the "standard" gains (similar to film), all the planets lined up. Now, I am not a "brainiack" like Dan Margulis or Dan Ramaley, but as time passes, I wonder if the previous system with it's higher gain was more about "the best achievable result" than "the best we can do with film".

Andy Adams
___________________________________________________________________________

Re: K in commercial printing/history of Matchprint
Posted by: Terence Wyse
Wed Aug 27, 2008 12:23 pm (PDT)

(deep breath Terry...Dan's just being cantankourous as usual)

OK, here we go...

On Aug 26, 2008, at 10:29 PM, Dan Margulis wrote:

The concept is fine but it's easier said than implemented. "Been in
the mainstream" is not the same as being so dominant that the
competition "has all but disappeared." The first is, as you say, 10
years at least. The second development is much more recent. That
means there was a period where there was a serious inconvenience if
CTP did not appear to the client as if it were a film-based system.
Consequently, the development of CTP was "it must behave as film-to-
plate did." Now that we don't have to worry about film-to-plate very
much, we could theoretically make the changeover and get somewhat
better quality, but it would be a difficult sell with this much
history behind doing it the other way.

If, as you say, that the disappearance of film-based platemaking is a relatively recent event, then perhaps now's the appropriate time to start making this switch so we won't be talking 10 years from now about trying to accommodate legacy separations when the technology they were based on are but a glimmer.

As I read it, you are suggesting that if a person loads a stupidly
made third-party profile and get a stupid result, well, boys will be
boys, but if the same person loads a stupidly made Custom CMYK
profile and gets the same poor result it proves that the whole system
is broken and anybody who uses is it a troglodyte.

Luddite maybe, but not troglodyte. :-)

OK, here's a fact:

Take either Monaco PROFILER or ProfileMaker, two "mainstream" profiling applications, load in a set of newspaper/newsprint characterization data ("ink set" effectively) and select either "newspaper/newsprint" from the separation settings preset menu and what you'll get is a profile totally usable for producing "newspaper/ newsprint" separations. GCR will be appropriate, total ink will be correct etc. Remember, I haven't touched a thing from a "customization" standpoint, I'm only selecting what's already built in to the application.

Try doing the same in Photoshop using Custom CMYK. You will NOT get an appropriate profile/separation unless you KNOW something about the target printing condition/specification. Even then, you will have to customize the settings in Photoshop's Custom CMYK engine to get an
appropriate profile.

More to the point (and I think this is the CRITICAL bit), it's not transparent to the user what the various ink sets are based on. "Newspaper", is that SNAP, IFRA, what, and what version of SNAP or IFRA would it be referenced to? We have no idea without, at the very least, interrogating the L*a*b* values in the ink set and seeing if they correlate to any current or legacy standards. Maybe that information is somewhere but it's not available in the UI. And even MORE to the point, AFAIK none of the built-in ink sets are based on CURRENT printing specifications embodied in ISO 12647 and/or any of the current FOGRA and GRACoL/SWOP data sets. So, even if you KNEW what the appropriate separation settings should be, you're still unlikely to build a profile that's usable relevant to today's standard printing specifications. To do this, Photoshop would need the ability to import a characterization data set (MACHINE-generated, no doubt!) supplied (freely) by FOGRA, ISO, IDEAlliance or other standards and/or specification bodies.

No, it was *me* who pointed this out in PP1E (1994) and every edition
since. Every graphics program I worked with has its share of stupid
defaults that knowledgeable users change. Simultaneous with the last
major change in Custom CMYK (1998, Photoshop 5), there came a major
change in RGB handling, to wit, all incoming tagged RGB files would
immediately, and without warning, be converted to sRGB. Personally, I
think that any change in colorspace wihout a warning is a stupid and
quality-damaging action, and accordingly I condemned this unrevealed
RGB-to-RGB conversion at the time, and now. In the interest of
logical consistency, I condemn (as even more damaging, which they
are) the unrevealed CMYK-to-CMYK conversions being advocated in
various quarters in the last few days on this list.

Now you're dredging up the same old saw about PS5. We've moved beyond that. Sure, the IMPLEMENTATION of color management in PS5 was flawed but was more-or-less fixed in PS6. I suppose if any of your students are still using PS5, then, yea, they're in trouble.

As far as CMTK-to-CMYK conversions via device link profiles being damaging, fact is, I don't believe you've even worked with any of these applications. You've worked with Left Dakota Link-o-Later, Alwan LinkProfiler and CMYK Optimizer? If not, HOW WOULD YOU KNOW if these conversions are "damaging"? If you'd actually worked with any of them, you'd realize that the control and options you have are so broad that it would be impossible to make any kind of general statement like this.

For similar logical consistency, if you are to suggest that Custom
CMYK is to be avoided because of its silly defaults, you must either
1) agree that converting all incoming RGB files to sRGB without
warning is a good thing; or 2) caution Photoshop users that RGB is to
be avoided, and all work should be done in LAB or CMYK.

"Logical consistency"? I'm reading this and don't even know what point you're trying to make. Sounds like another rehash of PS5 antics.

Examination of the Adobe-supplied sheetfed uncoated v2 profile vs.
the web uncoated v2 profile reveals that they are one and the same.
These two profiles (or one profile, depending upon how you look at
it) are, AFAIK, basically machine-generated. Again, for logical
consistency, it would be helpful to know whether 1) you agree that
sheetfed uncoated and web uncoated printing conditions are exactly
the same and that one profile suffices for both; or 2) machine-
generated profiles are, if not outright wrong, at least VERY suspect
in your book.

What is this "machine-generated" thing you're speaking of? If the "machine" is a computer, yea, that's how profiles are made today. Pretty tough otherwise.

If by "machine-generated", you mean a profile created from measurements made with a spectrophotomer, then duh, yea, that's how data sets are created from which to build profiles from. Even the L*a*b* values used in Photoshop's Custom CMYK must have came from one of these "machines", it's just not clear what the printing conditions were that generated these measurements.

As far as the two profiles you mention, yea, I've checked them out before and they are simply inappropriate profiles and appear to be a duplicate of each other. What does this tell YOU exactly, that the "machine" is at fault for measuring two identical press sheets or proofs but then this data was used to build two identical profiles for two quite different printing conditions? I guess I would fault the person "behind the wheel" for either being too lazy or too stupid to know not to use the same set of characterization data to describe two different sets of printing conditions. How in God's name does this implicate "machine-generated" (your term) profiles? And what, for crying out loud, IS a machine-generated profile and how does it differ from either other profile?

As to the rest of your comments below, I simply wish you would state unequivocally whether you feel Photoshop's Custom CMYK "profile creation tool" (my words) is up to the same quality standard as dedicated profiling applications? If it is, how is that, given some of the deficiencies that have been pointed out by you, myself and others? If it is not up to the standards of dedicated profiling apps, would it not make sense to at least alert your readers/listeners by way of at least a footnote in your publications? One thing is clear, at least to me, is that you cannot generate a profile based on CURRENT North American and International standards and printing specifications using what's available in Photoshop today, period. And if you WERE able to somehow pull off this magic act, I guarantee that it would be considerably more difficult than simply using any of the current dedicated profiling applications.

What I'm NOT saying is that any Photoshop user worth their retouching salt needs to go out and drop 2 grand on a profiling application, although if they were to do that, I strongly believe they would benefit by it. What I'm saying is that a Photoshop user could be perfectly happy using a standard suite of freely available profiles that are based on current international printing specifications but if you need more than that, you need to consider your options.

Regards,
Terry Wyse
___________________________________________________________________________

Re: K in commercial printing/history of Matchprint
Posted by: Henry Davis
Wed Aug 27, 2008 12:23 pm (PDT)

On Aug 27, 2008, at 9:36 AM, dacolorman wrote:

Years ago I was all about keeping the final result at the press to
12%-16% (since that is what the CTP allowed me to do). But, all was not well for a number
of reasons: supplied seps based on 20%+ gains, correct densities on press still resulted
in a sheet that was "starved" for color, etc. So, I chose not to reinvent the wheel,
but to work with the wheel I was given (by others much smarter than myself).

I could have (finally) gotten everything to work, but for what? For
less gain? That only amounted to a "savings" of 5%-10%. That may sound like a lot, but
if everything is and has been built to work (separations, plates, press, etc.) with higher
legacy gains the results still "looked" good (because everyone is compensating for the well
know "legacy" gains).

Once I went back to the "standard" gains (similar to film), all the
planets lined up. Now, I am not a "brainiack" like Dan Margulis or Dan Ramaley, but as time
passes, I wonder if the previous system with it's higher gain was more about "the best
achievable result" than "the best we can do with film".

I appreciate what you have related in this post. A great deal of thought and effort has been directed at the processes involving ink and paper and color etc. during the evolution of printing. Some very sharp minds have been at it for ages. That traditional dot gain is what it is, is not an accident - it was by design. I am glad to hear that your solution involves, um, less struggle - if only for now. Dread the day when, in a single instant, the specs for the old and the new become even further distanced. Oh, but I almost forgot - at that instant there will also be a few new tools that you will buy and train on, and a new convoluted workflow to adopt, and a new customer relations educational effort etc.

Whenever there is criticism leveled at someone for being "against change", it might be wise to uncover the true motivations behind the agents of change rather than joining with the band in criticism. This is especially the case when the "if it isn't broken . . " axiom is a good argument. In this thread, even the leading proponents of change admit that "dot gain is good". So, shouldn't it become a matter of providing sufficiently reasonable evidence that a change from traditional gain *ought* to occur? And where was this discourse when the very method by which separations are made was changed? When a change is claimed to be "necessary", is it too much to ask why, and expect a reasonable answer? "Because we can" isn't a good answer.

The same goes for the condemnation of the Custom CMYK tool. That there is no profile editing tool in Photoshop is a question that involves mystery motivations, and they are suspicious at the least. New "rules" were imposed for separations - and the motivations behind it should have been made clear to everyone in the game. There was no such clarity or coordination. Has Adobe acted responsibly with regard to its imposition of new "rules"? One needs only to ask this question to every professional Photoshop user and every print shop for the industry perspective. These new methods and specs were imposed in a disjointed manner and became a cause for problems at each stage in the print process.

Now, I'm not saying that the new tools are not good, even if some list members will take it that way. I am not bashing Adobe or the color management community. But for the hoi-polloi, well we just get jerked this way and that from one year to the next - while things weren't awfully, terribly broken to begin with. Sometimes this all reminds me of the old saw about allowing doctors to invent diseases.

Henry Davis
___________________________________________________________________________
.
Re: K in commercial printing/history of Matchprint
Posted by: Dan Margulis
Wed Aug 27, 2008 2:26 pm (PDT)

Henry Davis writes,

Whenever there is criticism leveled at someone for being "against
change", it might be wise to uncover the true motivations behind the
agents of change rather than joining with the band in criticism.
This is especially the case when the "if it isn't broken . . " axiom
is a good argument.

In the last 36 hours I happen to have gotten two off-list messages that bear on this thread (but have nothing to do with one another). As one is pretty much a rephrasing of what Henry is saying, I post it here, and the other, which is from a professional photographer having difficulty with CMYK, will go elsewhere.

Anyhow, the writer of the following is a prepress manager in a large printing facility. It is interesting in case there is any doubt as to the awareness of employees that conversions to heavier GCR is harmful to quality.

Dan Margulis

**********************
From:
Subject: CMYK to CMYK (ink savings) conversions
Date: August 27, 2008 12:30:44 PM EDT
To: Dan Margulis

You wrote and I wholeheartedly agree with:

"In the interest of logical consistency, I condemn (as even more damaging,
which they are) the unrevealed CMYK-to-CMYK conversions being advocated in
various quarters in the last few days on this list."

I am getting this very thing wadded up and jammed in places we can't discuss
in polite company. But we're prepress! It is all about the dollar.
Quality has been replaced with "good enough" and "maybe the client won't notice."

Ink and paper are in run away inflation with no end in sight. The bean
counters hear a wonderful presentation from a nice smelling salesman (who
has NEVER made a separation yet alone a good separation in his life) that
states that their software will save up to 25% on the ink bill, the product
will look the same, and nothing else matters.

People in my shoes realize that this is just buying a level spot in the
inflation trail and maybe increasing our competitive edge IF we can get away
with it.

I am afraid that us troglodytes are throwing stones at the storm. It just
may be time to walk away.

Best,
___________________________________________________________________________
.
Re: K in commercial printing/history of Matchprint
Posted by: Henry Davis
Wed Aug 27, 2008 2:26 pm (PDT)

On Aug 27, 2008, at 12:45 PM, Terence Wyse wrote:

<Snip>
In your reply to Dan you ask the question:

As to the rest of your comments below, I simply wish you would state
unequivocally whether you feel Photoshop's Custom CMYK "profile
creation tool" (my words) is up to the same quality standard as
dedicated profiling applications? If it is, how is that, given some of
the deficiencies that have been pointed out by you, myself and others?
If it is not up to the standards of dedicated profiling apps, would it
not make sense to at least alert your readers/listeners by way of at
least a footnote in your publications? One thing is clear, at least to
me, is that you cannot generate a profile based on CURRENT North
American and International standards and printing specifications using
what's available in Photoshop today, period. And if you WERE able to
somehow pull off this magic act, I guarantee that it would be
considerably more difficult than simply using any of the current
dedicated profiling applications.

Prior to the introduction, the Custom CMYK tool was the only such tool that was offered in the program. Even at that time, the tool was less than what one would want, but it was at least at tool that was in keeping within the times that it existed.

Since the introduction and imposition of ICC profiles, the game has radically changed and the program's tool has not kept apace. Enough years have passed for the program's management team to at least offer the users a reasonable explanation.

So, one answer to your question is - that is not a proper question. Custom CMYK was not designed to play according to the new rules. A proper question is this:

Since the game has changed, would a profile editor be a better solution than Custom CMYK?

As far as I know, no one has suggested that Custom CMYK is even in the same league as dedicated profiling programs. But, it is the only tool that the program provides, and there is the suspicion that it will be eliminated. This would leave nothing at all for the user. If one has enough dedicated profiling packages that one can edit any profile, regardless of the profile's originator, then perhaps the concerns voiced in this thread will not seem that, um, concerning.

Henry Davis
___________________________________________________________________________

Re: K in commercial printing/history of Matchprint
Posted by: Terence Wyse
Wed Aug 27, 2008 3:42 pm (PDT)

On Aug 27, 2008, at 4:47 PM, Dan Margulis wrote:

Anyhow, the writer of the following is a prepress manager in a large
printing facility. It is
interesting in case there is any doubt as to the awareness of
employees that conversions
to heavier GCR is harmful to quality.

I feel his pain, just like I felt pain 15 years ago when I saw the drum scanning job I loved gradually get replaced by cheap flat bed scanners.

Notice that his complaint is not really about any real change in QUALITY or the supposed "damage" these conversions might do but is more about losing one more bit of his "craft" that he calls prepress. That's understandable but there is nothing in there about the FACT that these conversions IF DONE PROPERLY cause such damage. I guess we'd have to ressurect something akin to the 8 vs. 16-bit debate (NOBODY wants that!) and see if converting to/from the same profile via device link causes damage.

I can tell you this, I could relay stories about printers that have implemented this technology and, not only did they save ink, but in fact they're printing quality IMPROVED and they're now much better at matching their proofs that have been set to match an industry standard specification (GRACoL2006 Coated1).

The sales idiot that quoted "25% ink savings" is a dumb @ss in my opinion. You should be presenting this technology based on the color quality improvement and that ink savings (potentially) is only the icing on the cake and shouldn't be the primary goal. But I can see that the reasons would be different for different markets...commercial sheetfed, you could make the case for color quality improvement alone....for web offset publication, the driving force would likely be ink savings and the "icing" would be improved color matching.
Regards,
Terry Wyse
___________________________________________________________________________

Re: K in commercial printing/history of Matchprint
Posted by: Henry Davis
Wed Aug 27, 2008 5:20 pm (PDT)

On Aug 27, 2008, at 5:59 PM, Terence Wyse wrote:

I feel his pain, just like I felt pain 15 years ago when I saw the
drum scanning job I loved gradually get replaced by cheap flat bed
scanners.

Notice that his complaint is not really about any real change in
QUALITY or the supposed "damage" these conversions might do but is
more about losing one more bit of his "craft" that he calls prepress.
That's understandable but there is nothing in there about the FACT
that these conversions IF DONE PROPERLY cause such damage. I guess
we'd have to ressurect something akin to the 8 vs. 16-bit debate
(NOBODY wants that!) and see if converting to/from the same profile
via device link causes damage.

Surely your scanning and prep skills can beat those of the secretary's on a cheap flatbed! You can probably also nudge a profile around for a bit of improvement.

Push-button color, yep, it's here. Every cheap machine and every canned profile is push-button color. Not that the goal is bad, or that the color is necessarily bad. But a black box or proprietary approach that shuts off the primary user's ability is like handing someone a camera and telling them that it will only show them black and white, and that some other entity will ultimately be responsible for the color - no worries.

Kurt Vonnegut intended to ask this question if, after death, he met his creator: "what were people for, anyway?", or something along those lines. Perhaps one day color management consultants will go the way of the buggy-whip makers - once that magic color button has been developed and sufficiently imposed. It's not sour grapes for me personally, but I do think that there could be more flexibility in the process.

Imagine all the colors, coming out like you planned. I can't wait.

Henry Davis
___________________________________________________________________________

Re: K in commercial printing/history of Matchprint
Posted by: Dan Remaley
Wed Aug 27, 2008 7:46 pm (PDT)

Anyhow, the writer of the following is a prepress manager in a large
printing facility. It is interesting in case there is any doubt as to the
awareness of employees that conversions to heavier GCR is harmful to quality.

Dan Margulis

What? I have some doubt. "A printing manager in a large print facility" - means nothing!!!!!
I have personally been at a midwest print facility (brand name), with 8 sheetfed presses and 6 web presses - all full size. They were printing 'linear' plates on ALL these presses, no curves, no control, no nothing. In this case size doesn't matter! I'll also bet a color consultant, GATF - or otherwise, has ever crossed their these guys door.
Dan Remaley/former GATF

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

Quality has been replaced with "good enough" and "maybe the client won't
notice."

Bla, bla, bla - give me some numbers man, and I'll tell YOU how you print. The main press manufacturer's (from Europe), KBA ManRoland, Heidelberg can't even measure midtone gain, where the control is needed. The reason Komori got it right is that GATF made their color bars. No brag-just fact! Please, don't take my word for it, challange yourself and read articles in TAGA and system Brunner. Then for asmall fee I'll comne out and make it happen. :-)

Dan Remaley/former GATF
___________________________________________________________________________
.
Re: K in commercial printing/history of Matchprint
Posted by: "Louis Dery"
Wed Aug 27, 2008 7:47 pm (PDT)

On Aug 27, 2008, at 5:59 PM, Terence Wyse wrote:

The sales idiot that quoted "25% ink savings" is a dumb @ss in my
opinion. You should be presenting this technology based on the color
quality improvement and that ink savings (potentially) is only the
icing on the cake and shouldn't be the primary goal. But I can see
that the reasons would be different for different markets...commercial
sheetfed, you could make the case for color quality improvement
alone....for web offset publication, the driving force would likely be
ink savings and the "icing" would be improved color matching.

Hi Terry,

I agree with you about the ink saving. It depend of the market.

About 25% for web offset, I can tell that it is been verified by some of our clients using our device link software. Not only with artificial calculation; they took the weigh of each ink thanker, before and after the press runs. About quality, no difference BUT it depend which software is used to make GCR black generation. They are not all using the same way to calculate GCR.

Regards,

Louis Dery
TGLC inc.
___________________________________________________________________________

Re: K in commercial printing
Posted by: Dan Margulis
Wed Aug 27, 2008 3:42 pm (PDT)

Rick Gordon writes,

This post is not addressing the different matter of the printer changing the GCR, which sounds much more offensive to me.

It's a *lot* more offensive. The following note may be sobering to those who are not CMYK experts, but have to produce CMYK work, and who may be thinking of following some of the advice from the color management quarters given here.

The writer is a professional photographer of some repute, who has attended a lecture that I gave. I am not in a position to comment on his competence or whether he does, in fact, have a calibrated monitor as claimed. I suspect that this must be one of his first forays into the CMYK jungle, because otherwise he would know that the following explanation is very incomplete and a whole gang of other questions need to be answered.

His description of what's wrong with the final result is clear enough. We do not have enough information to know whose fault it is; the fact that three different printers are involved makes one skeptical about what was being supplied. The phrase "no detail in the blacks", however is the John Hancock of a printer who is laying down black too heavily. No separation methods account for this result. It is not at all too big of a coincidence for all three printers to be making the same error. Printing black too heavily is much more frequent than any other channel, as I pointed out in an earlier post.

The rest is unknown and maybe I will get some details after I reply offline. We don't know, for example, whether the photographer gave the printer a CMYK file, or wimped out and gave RGB because he believed the pap he read somewhere that printers are the experts at conversions and know way more about it than photographers do. (Based on the timorousness of the post, I suspect the latter.) And I repeat that we do not know whether the files themselves were any good because I can't vouch for the photographer's skill. But I will make the following observation:

If the photographer supplied RGB files and the printer separated using one of these heavy-black atrocities being hyped here (or worse yet, if he submitted CMYK and the printer reseparated it without telling him), that, in combination with the excessive black inking that is strongly suggested above, would produce the effect being complained of, even if the photographer's original files were good. If the CMYK images were produced with a skeleton black such as Custom CMYK's Light GCR or UCR, there would still be loss in the deep black because of the excessive inking, however the overall muddiness and darkness would not be there.

Food for thought for those getting ready to put their toe in to the CMYK water. Chances are this photographer is going to suddenly see the merits of Custom CMYK.

Dan Margulis

*************
From:
Date: August 26, 2008 10:22:11 AM EDT
To: Dan Margulis
Subject: question

Hi Dan
I am writing to you in hopes that you can answer this for me, because its become quite urgent.

I have had 3 jobs back to back printed, all on different paper, different magazines and from different press houses, and they have all come out the exact same way:

Extremely dark, and extremely high on contrasts. Nothing breathes, the skins are patchy and dark and there is no detail in the blacks.

My settings are as follows and my screen is calibrated:
SETTINGS : north American general purpose 2
RGB: Srgb IEC61966-2.1
CMYK: us web coated (SWOP) v2
GRAY: dot gain 20%
SPOT: dot gain 20%

In color management: they are all in "preserve embedded profiles"

I'm Hoping that asking you to shed some light on this matter is not too much to ask

Best regards,
___________________________________________________________________________
.
Re: K in commercial printing
Posted by: "Jim Bean"
Wed Aug 27, 2008 5:20 pm (PDT)

I have had 3 jobs back to back printed, all on different paper, different magazines and
from different press houses, and they have all come out the exact same way:

being among those referred to as a professional photographer, I have to say this person was at least consistent in his output, three different prints/papers.. identical results.. too bad they were not better results.. regardless of the rgb/cmyk submission... many of my associates live/breathe and sometimes die by what they think they are seeing on their monitors.. info palletes and a quick trip to a curves window is simply not in their workflow-they many times work in poor lighting environment on a quality 'calibrated'laptop'... the frequency of a pro shooter understanding elementary aspects of cmyk file prep is unfortunately low.

jim bean
___________________________________________________________________________

Re: K in commercial printing
Posted by: Terence Wyse
Wed Aug 27, 2008 7:47 pm (PDT)

I suppose all this (below) is meant to sound pretty scarey (and really has nothing to do with re-purposing separations via device links) but to me it's quite simple:

* Did he check with those printers to make sure that "USWeb(Coated)SWOP_v2" did indeed represent their printing conditions or did he just ASSUME that using "North American General Purpose 2" defaults would be correct?

* Did he provide a (certified) SWOP proof to the printer showing his intended color-match? Alternatively, did a request a contract proof from the printer that could be assumed to match what they could achieve on press?

If none of this is true, then he rolled the dice....and lost.

I can't comment on the state of his monitor calibration only to say that if he's relying on this as his ONLY verification of his separations are correct or not and wanted to save the cost of proofs provided by the printer, well, all I can say is having a "hard" proofing system of his own would look like pretty cheap insurance right now.

Regards,
Terry
___________________________________________________________________________

Re: K in commercial printing
Posted by: "Mike Russell"
Thu Aug 28, 2008 7:17 am (PDT)

From: "Terence Wyse"
...
If none of this is true, then he rolled the dice....and lost.

Hence the point of the skeleton black. The dice can be loaded so that the probability of boxcars or snakeeyes is close to zero.
___________________________________________________________________________

Re: K in commercial printing/history of Matchprint
Posted by: Terence Wyse
Thu Aug 28, 2008 7:17 am (PDT)

Glad you could join the discussion Louis!

Tell us, what is YOUR take on the notion that these device link conversions are inherently damaging and that, as some have suggested, that it's tantamount to a breach of contract to re-separate a client's supplied CMYK images. Loved to hear what you have to say and I'm sure others would too from the perspective of a vendor providing these solutions.

Take it away...

Terry Wyse
___________________________________________________________________________

Re: K in commercial printing/history of Matchprint
Posted by: "Louis Dery"
Thu Aug 28, 2008 9:34 am (PDT)

Hi Terry,

Here is what I think:

About devicelink conversions damage or breach of contract because of reseparation of supplied client files: what is "changing client’s file?
— Resepartion (CMYK to CMYK)
— Custom correction curves to the CTP
— Fixing high ink coverage with whatever tools

I think that all of the above are changing client’s file, BUT it is in order to get it match on press, avoiding "Photoshoping" with the printing press.The printing press is made to PRINT and not to constantly move the standard press conditions (densities, etc.) form job to job! What about those who ask color correction on press even if the press match the contract proof? Lack of communication!

The whole idea is communication!
Communication between client and prepress/printing
Communication between proofing (prepress) and press
Communication between production tools (Photoshop, layout, etc.)
Communication between CSR and client (not only ask "how many copies" but also talk about colour!

I can say that we’ve made a lot of press test, demon and installations with our deviceLink solutions and we never failed to satisfy clients (prinitng) and clients of those clients (customers).

Louis Dery
TGLC inc.
___________________________________________________________________________

K in commercial printing/history of Matchprint
Posted by: "Paul Foerts"
Thu Aug 28, 2008 5:10 pm (PDT)

On Wed Aug 27, 2008 7:46 pm ((PDT)), Dan Remaley wrote:

Bla, bla, bla - give me some numbers man, and I'll tell YOU how you
print. The main press manufacturer's
(from Europe), KBA ManRoland, Heidelberg can't even measure midtone
gain, where the control is needed.
The reason Komori got it right is that GATF made their color bars. No
brag-just fact! Please, don't take my word for it, challange yourself and read articles
in TAGA and system Brunner.

Some pieces are missing in this puzzle...
The fact that a 50% patch is not used in print control strips (from Europe) is because of a lawsuit...

The fact that GATF has been selling "European" color bars for years has also to do with a "settlement"... (or was it an agreement?)

The lack of a 50% patch in color bars has never caused any harm to anybody, except maybe to those who exchanged "dot gain" numbers without proper reference values.

Prepress professionals have never used only one reference point for the reproduction curve. Brunner did put 25 tone values (for CMYK+) on his "Eurostandard-Testform".

Please explain why your "midtone control" tool is any better than what we have today for press control...

Paul Foerts

On a side note: in search of a Troglodyte... As it is getting more difficult to sell tools to the graphic arts industry, companies may dump their evangelists and hire top selling artists. These people have a record of selling anything! This is real! This is hot... (Is it still hot?) :-)) Go to www.cloaca.be

and enjoy!
___________________________________________________________________________

Re: K in commercial printing/history of Matchprint
Posted by: Dan Margulis
Thu Aug 28, 2008 5:12 pm (PDT)

Terry Wyse, responding to an anonymous message that I had forwarded from a prepress manager, writes,

Notice that his complaint is not really about any real change in QUALITY ...

He stated "Quality has been replaced with 'good enough' and 'maybe the client won't notice.'" Perhaps the fact that QUALITY was not capitalized fooled you into thinking he was not stating that there is a quality issue. Or perhaps you read "good enough" at face value, whereas I think most people would read it as being a pejorative phrase.

....but is more about losing one more bit of his "craft" that he calls prepress.

His "craft" is unaffected, the whole process is transparent to prepress. His folk manipulate the images as normal; the separation or reseparation happens downstream. What he is complaining about is that he is judged by how his work appears in print, and he is well aware that the quality thereof will suffer.

I can tell you this, I could relay stories about printers that have
implemented this technology and, not only did they save ink, but in
fact they're printing quality IMPROVED and they're now much better at
matching their proofs that have been set to match an industry standard
specification (GRACoL2006 Coated1).

Possibly so, but I daresay these printers are also committed to process control. If it's your common garden variety of commercial printer, their clients will rue the day that they decided to jack up the black.

Undisclosed in the original post (but disclosed below) is that the application is a high-circulation newspaper, which makes a difference. If that is known, now the quality loss is an absolute certainty. No newspaper printer, however quality-conscious, can hope to have the precise control of the black inking that this technology would require. An excellent commercial printer, yes. A newspaper, not a prayer.

The sales idiot that quoted "25% ink savings" is a dumb @ss in my
opinion. You should be presenting this technology based on the color
quality improvement and that ink savings (potentially) is only the
icing on the cake and shouldn't be the primary goal.

Well, as Louis pointed out, 25% ink savings *is* possible. Only however, if a *VERY* heavy black generation is used, one that in many cases permits the weakest CMY ink to reach zero. If that's the case, it's idle to pretend that it's about anything but money. We all agree that a somewhat heavier black, given adequate process control, gives certain advantages along with certain drawbacks. At some point, though, there is no benefit in making the black even stronger, and if we do there are no advantages but only problems. That point has long since passed. Making the black THAT heavy, in a newspaper environment, is a guarantee of muddiness and C-minus quality at best.

Note the further comment on motivation from the manager below in the last paragraph. Having some connections in the newspaper industry myself, I can point out that the last year has been absolutely disastrous--and those words are not too strong--for newspapers, which haven't been doing well for a long time. Display advertising has cratered, classifieds have gone to craigslist, and circulations are down. Massive layoffs are everywhere and there is no end in sight. The chances of a newspaper making significant hardware (or training) expenditures are very small.

When the manager suggests, therefore, that these ink savings may save somebody's job, he's probably not kidding. Desperate times call for desperate measures. Perhaps, therefore, he (and we) can forgive the quality hit under the circumstances.

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

Now, let me don the moderator hat. From time to time I post here, without a name being attached, correspondence I have received from others. An example of this is the photographer's comment posted at the same time as the one from the newspaper's manager. That photographer is not, to my knowledge, a list member. I don't know whether he would object to my posting his name, but he was unaware that anything was going to the list, and I have no intention of causing embarrassment by requesting a permission to put a name to it.

The newspaper manager, however, *is* a list member. We have a firm policy here of not accepting anonymous posts. The reason is that other lists have run into big problems from inflammatory posts from folks who are brave enough to use nasty language but not brave enough to identify themselves. I have refused requests for anonymity from people who thought their questions were so simple as to be professionally embarrassing, and also from those who were afraid that they would be flamed on the list.

About the only exception to this policy is where a poster cannot speak freely if the name is used. That does not mean because he is afraid of being flamed here. It means that the comments can create professional difficulties if they are seen outside of the list, ordinarily by a boss or a client who may find them inappropriate. The comments of this manager, IMHO, clearly fall into that category, and accordingly I deleted his name before forwarding the first post.

Earlier today I received a second response from the same individual, and then a separate message saying I was free to post to the group or not as I saw fit. As I think there is valuable information in it, I choose the former course. However, I would like to point out that it contains criticism of other list members by name. This makes me uncomfortable when the poster's own name is not available but I am really disinclined to start editing people's text. I believe that both Terry and Dan are big boys who have heard worse before and can live with the following one-time post.

So, I make the following public comment to the poster. If you take from the above that I'm calling you a coward and decide to make your name public to teach me a lesson, you're a fool. Saying some of the things said below can cause real trouble for you and your company if certain clients see them. OTOH, if you wish to forward further material for this thread, I'll still withhold the name, but ask that anything reflecting negatively on other members be omitted.

Dan Margulis

************************
From:
Date: August 28, 2008 11:13:38 AM EDT
To: Dan Margulis
Subject: FW: [colortheory] Re: K in commercial printing/history of Matchprint

Good morning Dan,

I am fully aware that posting my name and employer would probably add weight and credibility to my opinions and statements. However, both of the asinine responses to my original post are why I ask that you filter my postings. Hell, I am more thick skinned than most of my contemporaries. I just don't want to deal with the added troubles.

I used to have some respect (in perspective, mind you) for Dan Remaley. And I never really cared about or for Mr. Wyse. Who obviously has not been busy with any of his clients in the last ten days! Gee, wonder why?

I do have some guilt about the situation here. After all, a good two thirds of the commercial and advertising work that passes through our hands is that damn SWOP (v.2) profile embedded or otherwise, and I have no qualms about a quick profile-to-profile conversion to make life easier in the press hall.

If these folks don't care to provide a properly prepared file for newsprint, on what grounds could they complain if it prints well and better than originally submitted? So long as they are signing their checks and returning with more advertising and money, what's the harm?

On the other hand it galls the hell out of me to automatically perform a CMYK-to-CMYK conversion on EVERY page file that is passed through to the RIP! Nobody, save for you Dan, has any business telling me what profile or black generation I should be using on a given image file.

This crap is going to improve quality as these clowns have responded? Bull[feathers/D.M.]! Just let the black ink density float for a few hundred, let alone a hundred thousand, copies and see where all the savings go. We'll be doing print-and-losses, make-goods, and discounted space ads for years to come.

For the record, my production bosses all the way to the top of the corporation know the score. It's not about quality. It's about keeping the doors open and people employed. We're buying a software product that has an ROI of a couple of months IF it works as promoted that MAY keep us in business longer than the competition.
___________________________________________________________________________
.
Re: K in commercial printing/history of Matchprint
Posted by: "RJay Hansen"
Thu Aug 28, 2008 5:13 pm (PDT)

No offense to Louis, but it would seem to me his take on this issue would hardly be unbiased...

RJay Hansen

On Wed, Aug 27, 2008 at 10:25 PM, Terence Wyse wrote:

Glad you could join the discussion Louis!

Tell us, what is YOUR take on the notion that these device link
conversions are inherently damaging and that, as some have suggested,
that it's tantamount to a breach of contract to re-separate a client's
supplied CMYK images. Loved to hear what you have to say and I'm sure
others would too from the perspective of a vendor providing these
solutions.
___________________________________________________________________________
.
Re: K in commercial printing/history of Matchprint
Posted by: "Louis Dery"
Thu Aug 28, 2008 5:29 pm (PDT)

Hi Rjay,

I understand your point.

I invite you to visit our web site and download the sample GCR images (on the PerfX Device Link page) and free tool to compare ink savings (PerfX image Ink] between our solution and the others. We are so confident that we provide one of the best tool (Device Link for ecxample) that we provide the free tool to compare! Try to find such reference form other solution provider. If you know one, please let me know.

Best regards,

Louis Dery
TGLC inc.
www.tglc.com
___________________________________________________________________________

Re: K in commercial printing/history of Matchprint
Posted by: Dan Margulis
Thu Aug 28, 2008 9:19 pm (PDT)

Louis Dery writes,

The whole idea is communication!
Communication between client and prepress/printing
Communication between proofing (prepress) and press
Communication between production tools (Photoshop, layout, etc.)
Communication between CSR and client (not only ask "how many copies"
but also talk about colour!

My question is, in view of the importance of communication noted above:

Do you, as a matter of policy, strongly advise those using your software to communicate this fact to their clients, so that any clients wishing not to partake of its benefits will be able to make an informed decision?

Dan Margulis
___________________________________________________________________________

Re: K in commercial printing/history of Matchprint
Posted by: "Louis Dery"
Thu Aug 28, 2008 9:19 pm (PDT)

Hi Dan,

You said heavier black channel (GCR) has no benefit and produces more problems… It depends on the way GCR has been calculated. If your reference is to the GCR made with Photoshop, I agree! But outhere, many ICC DeviceLink solutions and others have all their way to calculate what everybody call GCR. They are not doing it all properly. Some do and many don’t. That is why I would not call "heavy black" as a generic way to do GCR.

About ink savings and GCR, Customer reported that they started with conventional SWOP seps on press for catalog printing and they spent many hours to get client’s colours. They decided to remake plates with using devicelink made by us and went back to press.

The client was aside of the web press and sign all the copy right after that!

Resumé: make clients happy = money

my 2 cents.

regards,

Louis Dery
TGLC inc.
___________________________________________________________________________

Re: K in commercial printing
Posted by: Stephen Marsh
Fri Aug 29, 2008 5:15 am (PDT)

Jeremy, I am a little late to the party, sorry about that! Here are my belated thoughts on your original post.

Jeremy Stephenson wrote:
Are they even among the options listed in the Custom CMYK ink
options, or will I have to create the inks myself in the Custom Ink
options?

Custom CMYK had it's last major change in version 5, many current CMYK "standards" are not supported by the old Custom CMYK (for example, Custom CMYK "SWOP" is not the same as TR001 which was more commonly used for SWOP).

Do not expect similar results as an ICC profile if you choose to create the inks yourself with the same values as the ICC profile produces for solids/overprints/stock.

Where do you get the inkset/stock measurements from? A reference file that would normally be used to make an ICC profile?

Measuring the Lab values for these CMYK solids/stock in a file tagged with the correct ICC profile? Would you measure with Relative Colorimetric intent set in Color Settings - or Absolute Colorimetric? How would this affect the results?.

Custom CMYK is fantastic at what it does, although one should not expect too much of it.

There seems to be two camps here - Dan and everyone else! Everyone
else says use the supplied profile, and advise against other black
generation - Dan says pretty much the opposite.

A case of two vocal minorities - so what about the silent majority?

I also have the added issue mentioned before of 100%K (actually rich
black in most places) being the background colour of some pages. I
realise now that it makes it tough on the printing, but it's too late
to redesign the whole book now.

* Has your printer specified their preferred total ink weight for these large flat rich black panels?

* Has your printer specified their preferred colour build for these rich black panels?

* Has your printer mentioned what will happen if their prepress/press guy does not like the weight or build of these panels? Will you have to fix it, or will they "fix" it without telling you - and how would this impact on the final result?

* Are the rich black surrounds 0r0g0b and they then get converted to the same profile with the same settings (shadow build and limit will often vary with RelCol or Perceptual rendering intent options)? If you convert using the FOGRA profile, does the shadow total ink and channel build match the recipe supplied by the printer? Are they built in CMYK mode? Are the rich black panels raster or vector or mix of both and will these variables affect output?

I am asking Stamford (the printer) if they can print a second K ink.
That would make life a lot easier.

Perhaps in some ways it would be "easier" and perhaps not in others (extra cost and more complexity for prepress and press).

The other option would be to use UCR like Dan suggested, though that
jeopardises my black and white images.

Firstly, what is stopping you from using Custom CMYK Max GCR with UCA, then assigning the FOGRA profile to this separation and or softproofing with this profile? I doubt that the opponents of image dependent GCR generation can fault the use of more K and less CMY for four colour grayscale images or for when neutrals are more critical than colour.

For me, a good approach for colour images that contain neutral tones (not monotone images) is to use light K in the coloured areas and to use a heavier K generation with lesser CMY for the neutral areas. A hack would be to create a light K plate separation, isolate the neutral tones and then blend in a high GCR for the neutrals - ideally this would all be handled in a single conversion (I am not aware of an ICC profile that does this though).

Thanks for all your input. I am awaiting more info from the printer
to decide on my next step.

Jeremy, as much fun as it is to see the side topics raised in reply to your post thrashed to death yet again on this list, I personally would appreciate hearing what the printer has said and what you have done in the last two weeks since your original post - after all this topic was originally about your specific situation.

Sincerely,

Stephen Marsh
___________________________________________________________________________

Re: K in commercial printing
Posted by: Stephen Marsh
Fri Aug 29, 2008 5:18 am (PDT)

Terence Wyse:
If none of this is true, then he rolled the dice....and lost.

Mike Russell:
Hence the point of the skeleton black. The dice can be loaded so
that the probability of boxcars or snakeeyes is close to zero.

If as Terry muses, SWOP v2 is not an accurate description of the target, then this may be why quality is poor (how does the K plate of SWOP v2 compare to legacy Custom CMYK light GCR?).

The point raised by Mike may be for nothing, if the job is being reseparated in a manner that significantly alters the existing channel structure or ratio.

As has been noted, it depends on many variables when it comes to separation and reseparation (either with regular ICC profiles or DLP).

With all this talk of reseparating CMYK at the printer, it would appear that the "RGB only workflow" proposal of the 90's is being repeated again, in slightly different form without the RGB benefits and with the CMYK benefits perhaps destroyed. If my CMYK files that are *prepared for the target condition and image content* are going to be changed to "suit the target condition without my knowledge" - then I may as well just provide RGB and not bother with the CMYK stuff (I am in prepress, not photography).

Stephen Marsh
___________________________________________________________________________

Re: K in commercial printing
Posted by: Dan Remaley
Fri Aug 29, 2008 10:42 am (PDT)

You will be happy to know that their is a product out there that does exactly what want (below) automaticly. The product is <http://www.fineeyecolor.com> I think. It will have a larger impact on color than G7 or anything else that I have seen.

Dan Remaley
___________________________________________________________________________

Re: K in commercial printing
Posted by: Michael Jahn
Fri Aug 29, 2008 10:43 am (PDT)

@ Stephen Marsh

With all this talk of reseparating CMYK at the printer, it would
appear that the "RGB only workflow" proposal of the 90's is being
repeated again, in slightly different form without the RGB benefits
and with the CMYK benefits perhaps destroyed. If my CMYK files that
are *prepared for the target condition and image content* are going to
be changed to "suit the target condition without my knowledge" - then
I may as well just provide RGB and not bother with the CMYK stuff (I
am in prepress, not photography).

- in its simplest form, yes.

When one lives in a world of digital capture of images that are destined to be printed on a printing press, one could imagine that if one 'converted' the digital image from what was captured by the Bayer filter array into some "flattened" RGB - where one can push pixels around, do color corrections as required - and then place into a page - then this decision for the most appropriate separation could be made when one commits that image to some substrate.

I must say that while people do quickly "grasp" the concept - even agree - that when you prepare a color image or printing for a SWOP print condition, then later discover that that image might be used in a "Newspaper" printing condition - a room of us might all agree that 'they need something different than SWOP separations - well, lets keep that sort of thinking - then move that same image to Vinyl billboard, or even digital billboards - then think of something like an inkjet device where they use a light and dark component of a the same hue - like CcMmYyKk for example...

We can all agree that perhaps it was ill advised to move it to a specific CMYK only to have to try and re-separate it into a new CMYK for some other press condition.

So, yes. I agree with you Stephen - now that we have mostly digital images captured and mostly LCDs, Adobe 1998 RGB and sRGB - both invented to overcome things we no longer are encumbered by - may need replacing.

Michael Jahn
___________________________________________________________________________

Re: K in commercial printing
Posted by: "Romano, John"
Fri Aug 29, 2008 12:05 pm (PDT)

Hi Dan Remaley

Hey tell us more about this new product !

And do you work for them ?

What can you tell us about G7 ?

Also how many G7 press runs have you preformed ?

Regards

John Romano
___________________________________________________________________________

Re: K in commercial printing
Posted by: "Louis Dery"
Fri Aug 29, 2008 12:06 pm (PDT)

Hi Dan [Remaley],

FYI
I invite you to take a look at the GCR sample images at the following
URL:
http: //www.tglc.com/english/PerfX/ICC_DeviceLink_benefits.html#Demo_GCR

With our GCR approach, almost all vivid colours are made with 3 colours!

Louis Dery
TGLC inc.
___________________________________________________________________________

Re: K in commercial printing
Posted by: "Andrew Webb"
Fri Aug 29, 2008 12:06 pm (PDT)

<http://www.fineeyecolor.com/icemaker.html>

anyone tried it? I just requested the trial version of the IceSaver Beta.

/asw

Andrew Webb

Creative Director
Serious Retouching & Color
303.682.9119/303.819.0480
___________________________________________________________________________

Re: K in commercial printing
Posted by: "Romano, John"
Fri Aug 29, 2008 2:43 pm (PDT)

Thank you Terry for giving us the correct Link !

Well how about that, its a Color Server / Ink optimization software.

Guess what ? Its reseparating your precious images for Ink savings and GCR...Oh my bad it was PCR.

Must say it looks interesting but I would want to TEST it before commenting on it. Dan did you Test it ? How did it go ?

So if all this reseparating is so bad why are there so many companies making color servers ?

How many printers do you think use these ? They wouldnt be making them if there we not selling !

Heck I know Photographers and PrePress houses that use them.....so I can tell you there is a Huge Market.

Oh yeah and they work too !

Regards

John Romano
___________________________________________________________________________

Re: K in commercial printing
Posted by: dan remaley
Fri Aug 29, 2008 9:02 pm (PDT)

Hey tell us more about this new product !

The website is <www.fineeyecolor.com> It uses the power of software to act like the 'knobs' of an old Hell scanner, adjusting the amount of GCR throughout the tone scale -only where necessary.

And do you work for them ?

No, I do not. I got to meet the developer while at GATF. His name is Ernest Miller, he, like I, does not understand the term "pleasing color".

What can you tell us about G7 ?

I was there from the beginning, Don H. showed up (at GATF) several years ago touting these higher densities. I immediately questioned the high range of Magenta. My 40 years of experience at press side revealed that a Magenta printed at 1.46 or higher, casted everything RED. I know G7 has removed the Yelo and Magenta throughout the tone scale (smaller dots) to get rid of the cast, to remain neutral. Problem is, the press doesn't remain at a consistent density, it changes +/- as much as .10 throughout the run. When it does, at high density, it casts RED again, creating a problem for the pressman. He is at the 'edge' of control.

Also how many G7 press runs have you preformed ?

I usually end up at a printer when there is trouble. I lower the densities from G7, print to gray balance, measure dot gain, and create an environment for success, and allow for some variation. The G7 is great for the color management group, they can 'tweak profiles, etc. Proofers are pretty stable. On press it's a different story, G7 allows for little variation, it's better on brand new Heid. presses that are computer controlled. Most of the G7 guys I talk to admit that they drop the Magenta density a bit. Even Don backed off his original 1.50 position, I believe its now 1.45. I know they don't believe in density measurements, everything is Lab, well, except when you have to make plate curves and plug in the density of the tints (dot area measurements) of the tone scale. I guess it balances out, System Brunner doesn't believe in printing to Lab - my money's on Brunner.

Dan Remaley
___________________________________________________________________________

Re: K in commercial printing/history of Matchprint
Posted by: "Romano, John"
Fri Aug 29, 2008 5:16 am (PDT)

Dan

Do you actually think a Printer would listen to you ?

Oh I dont want you to use your Device Link workflow to Make the Job print better and save ink.....They would laugh you out the door.

Do you think you could tell say Quebecor World plant NOT to use DVLs or any of the other larger printers....Good luck on that.

Bottom line is they do not destroy files, Maybe you should download a Demo from Louis or Alwan Colors Color Hub. You can get a 10 day demo for free to try.

Do something extreme like take seps that were built for a sheetfed #1 coated and flip them to a Web Uncoated and look at your channels.

I have tested the Alwan product and I can tell you there is no destruction and I have nothing to gain by this statement, I just happen to know !

Regards

John Romano
___________________________________________________________________________

Re: K in commercial printing/history of Matchprint
Posted by: "Louis Dery"
Fri Aug 29, 2008 5:17 am (PDT)

Hello Dan,

Keep it simple! What I means is color information is important and needs to be communicated properly, no matter the way you are doing it.

Louis Dery

On Aug 28, 2008, at 10:11 PM, Dan Margulis wrote:

My question is, in view of the importance of communication noted
above:

Do you, as a matter of policy, strongly advise those using your
software to communicate this fact to their clients, so that any
clients wishing not to partake of its benefits will be able to make
an informed decision?
___________________________________________________________________________

Re: K in commercial printing/history of Matchprint
Posted by: Dan Remaley
Fri Aug 29, 2008 10:41 am (PDT)

Hi Paul, here's my logic. . .or lack of.

Some pieces are missing in this puzzle...
The fact that a 50% patch is not used in print control strips (from Europe)
is because of a lawsuit...

Don't know anything about that, but I can't believe that % tints could be in a patent? - I thought Europe used 20/40/60 because of their legacy 'positive' film. FYI- Europeon printers were always touted as being 'better' than North Americain printers. The real reason is that positive films (choke) going to plate. Negative films (expand) going to plate, so we darken and they lighten.

The fact that GATF has been selling "European" color bars for years has also
to do with a "settlement"... (or was it an agreement?)

At one time we sold a Brunner color bar, long ago. All of GATF products, test forms color bars, etc. are 25/50/75 - at least the 12 years I worked in that group.

The lack of a 50% patch in color bars has never caused any harm to anybody,
except maybe to those who exchanged "dot gain" numbers without proper
reference values.

The 50% patch is very sensitive to change - ever print purple? 50% Cyan / 50% Magenta - killer - no one can print it, (that's was Compac Computers logo color - good luck. Now since the 50% is more critcal, wouldn't it be a better place to measure than say 40% or 60%?

Prepress professionals have never used only one reference point for the
reproduction curve. Brunner did put 25 tone values (for CMYK+) on his
"Eurostandard-Testform".

No question, and Printergy will place all the values, but as you said, the 'specs' are based on 25/50/75. Ideally you would have 25/50/75 on the color bar, however because we need more patches of solids for each key - what % would you measure?

Please explain why your "midtone control" tool is any better than what we
have today for press control...

Same as above. . .it's where the change takes place. . .same reason we measure midtone gray . .values in the 50C/40M/40Y region. (For everyone)please send me an email at <danremaley@comcast.net> and I'll send you my process control reference guide.

Dan Remaley/former GATF
___________________________________________________________________________ .

Re: K in commercial printing/history of Matchprint
Posted by: Dan Margulis
Fri Aug 29, 2008 2:43 pm (PDT)

Louis Dery writes,

Hello Dan,

Keep it simple!
What I means is color information is important and needs to be
communicated properly, no matter the way you are doing it.

I agree on the importance of simplicity, which is why I phrased my question in a form that could be answered with a simple "yes" or "no". Since it was not answered, permit me to repeat it:

"Do you, as a matter of policy, strongly advise those using your software to communicate this fact to their clients, so that any clients wishing not to partake of its benefits will be able to make an informed decision?"

Dan Margulis
___________________________________________________________________________

Re: K in commercial printing/history of Matchprint
Posted by: Dan Remaley
Fri Aug 29, 2008 9:03 pm (PDT)

I don't know if I understand this thread? Is the position that we don't change the customers files? Let me take a swing at this -(even a .300 hitter is considered great).

We NEVER used the customers files - remember the overlays on paste-up boards that artists painfully designed? We threw them away in pre-press, they never fit. today we have digital files, some like word, no kerning going on here. Images are never in gray balance, even the photo's aren't within 2 f-stops of the correct exposure! I always tell the printers "you are the professionals in this business, repair this stuff". No matter what changes you make the proof will show the results and the customer can accept or reject the results.

Here's a classic example I use in my seminars: Your customer comes to you with a job designed with a 2" bar that runs across every page of a 36 page Kromekote or other expensive paper. The 2" wide bar is made of 50C/40M/40Y. What would you do? I can tell you that you can't print it form to form and match it. You would CHANGE the file to 25% c / 20 m / 20y and maybe 20% k so that you could control the color. Same applies to 4/color process. This is the only business that the customer supplies the 'raw' materials (files, photo's, fonts <sometimes>). If you wanted cabinets for the kitchen, you wouldn't bring the trees for them!

Dan R.
___________________________________________________________________________

Re: K in commercial printing/history of Matchprint
Posted by: J Walton
Sat Aug 30, 2008 5:25 am (PDT)

On Fri, Aug 29, 2008 at 6:41 PM, Dan Remaley wrote:

I don't know if I understand this thread?

I don't think you're alone. A lot of people don't seem to understand the thread, or the question that Dan posed several days ago:

Do you, as a matter of policy, strongly advise those using your
software to communicate this fact to their clients, so that any
clients wishing not to partake of its benefits will be able to make
an informed decision?

This was asked after it was stated that communication is really the key to quality results. Which would seem to suggest that before changing something that the client gave you in a substantive way you would communicate that.

Is the position that we don't change the customers files?

No, the positions seems to be to communicate, except for the matter of Device Link conversions.

Let me take a swing at this -(even a .300 hitter is considered great).

I think you're great even if you strike out on this one - and for the record, I am not the umpire.

We NEVER used the customers files - remember the overlays on paste-up
boards that artists painfully designed?

I remember paste-up boards, but since they were camera-ready we didn't do anything to change them unless the customer asked for it.

Wethrew them away in pre-press, they never fit.

Sounds like you had some sloppy customers, which is not an uncommon thing to have. I would not have gotten away with anything like that.

today we have digital files,
some like word, no kerning going on here.

True, but I highly doubt a printer who converts everything on the RIP is going to go in and fix letter spacing. A good example of garbage-in, garbage-out.

Images are never in gray balance, even the photo's aren't within 2 f-stops
of the correct exposure!

Again, those are some sloppy customers. That's a whole other end of the market from where I've spent most of my career.

I always tell the
printers "you are the professionals in this business, repair this stuff".

I quite agree with that. Just like a mechanic is the professional in their business, and their job is to repair things. But I would be upset if a mechanic did work to my car without telling me.

No matter what changes you make the proof
will show the results and the customer can accept or reject the results.

Very true - no argument there...except that if the Device Link conversion is done correctly it could be difficult to pick up on the contract proof, even if it caused trouble during the press run. But it is quite easy to bounce a job if it doesn't match the proof - and the final result is all that really matters.

Here's a classic example I use in my seminars: Your customer comes to you
with a job designed with a 2" bar that
runs across every page of a 36 page Kromekote or other expensive paper. The
2" wide bar is made of 50C/40M/40Y. What would you do?

The "Kromekote or other expensive paper" tells me that we're dealing someone who is looking for a quality end-product and would appreciate some kind of heads-up. The 50C/40M/40Y tells me we're dealing with a designer who knows just enough about printing to be dangerous, but not enough to be effective. The 2" bar running across every page is just bad design - if the project has gotten this far maybe the client won't care if the grays shift. But how do you really know for sure what a customer really wants you to do?

I assume that if the answer to everything is communication that a phone call would be the next step.

I can tell you that you can't print it form to form and match it.

That would be the basic subject of the phone conversation.

You would CHANGE the file to
25% c / 20 m / 20y and maybe 20% k so that you could control the color.

Quite so, and in this case, because you are fixing an obvious mistake from someone who clearly doesn't know what they are doing, fixing it without telling them saves them some embarrassment. Another example would be a client supplying Adobe SWOP TIFFs for a newsprint job - it's clearly a mistake. But what if someone with the savvy of Dan or Stephen submitted a job and intentionally tweaked the black to suit the image? That's where I agree with this statement: "What I means is color information is important and needs to be communicated properly, no matter the way you are doing it."

I don't think anybody is saying that Device Link profiles are evil in themselves. In fact, it's quite the opposite - the technology has tremendous potential. But the way it is being implemented by some people just seems fishy to me.

J Walton
___________________________________________________________________________

Re: K in commercial printing
Posted by: Dan Margulis
Sat Aug 30, 2008 8:38 am (PDT)

In reply to Stephen's suggestion that a separation algorithm could use different levels of GCR intelligently within the same image, Dan Remaley writes.

You will be happy to know that their is a product out there that does
exactly what want (below) automaticly.

If it works, that's great, ditto with the one mentioned by Louis. It would make printing considerably more reliable. It also might minimize the kludge discussed by Stephen (and shown in my books, too) of merging images that have different GCR levels in order to prevent color shifts or registration issues in critical areas, while permitting changes on press elsewhere. This is a practice that happens more often than people think, particularly in catalog work.

The problem is that all the effort you're going through is a complete waste of time if you encounter some printer who is so hard up for cash and so disinterested in preserving quality that he wipes out your black generation by reseparating your file. Which is what this thread is about.

Dan Margulis
___________________________________________________________________________

K in commercial printing
Posted by: "Paul Foerts"
Sat Aug 30, 2008 8:39 am (PDT)

On Fri Aug 29, 2008 10:42 am ((PDT)) Dan Remaley wrote:

You will be happy to know that their is a product out there that does
exactly what want (below) automaticly. The product is <www.fineeye.com> I
think. It will have a larger impact on color than G7 or anything else that
I have seen.

This is about PCR -> Programmed Color Reformulation...

Looks to me as the rebirth (redefinition / reformulation?) of PCR -> Polychromatic Colour Reduction (Crosfield) also called extended UCR (about 1990) ... ?

The least one could say is: wrong abbreviation used.

It is obvious now that the ICC route is questioned by more than one developer of post? ICC workflows. If the device independent "only" route, preached by some, would have been a success, there would have been no demand for device link profiles (in the Hell, Scitex, Crosfield, Dainippon etc. era known as color conversion modules) and other conversion tools as they are reincarnated these days...

Maybe there is still room for a device dependent workflow for those who want predictable color...

-------------

The fact that a 50% patch is not used in print control strips (from Europe)
is because of a lawsuit...

Don't know anything about that, but I can't believe that % tints could
be in a patent? - I thought Europe used 20/40/60 because of their legacy
'positive' film.

Mr. Brunner may tell you the whole history. (He was the first to use the 50% patches).

As not only 50 % tone values are present in real images, this issue is more of academic nature.

Paul Foerts
___________________________________________________________________________

Re: K in commercial printing
Posted by: Terence Wyse
Sat Aug 30, 2008 6:47 pm (PDT)

There's more than one "GCR" workflow/device link product that can "intelligently" adjust TAC and GCR. One that I'm very familiar with is Alwan CMYK Optimizer. It uses a "Dynamic Device Link" strategy that does pretty darn sophisticated image analysis based on pixel area to determine how TAC and black generation should be performed. It will literally build device link tables unique to each image based on criteria that you give it at the outset. AFAIK, these sort of "dynamic" or "intelligent" algorithms work only in the context of a dedicated workflow application or "color server" product. Alwan's own static device link profile product ("LinkProfiler") is virtually identical to their color server product with one key exception, it is NOT dynamic in nature due to the fact that the character of the device link is "hard-wired" at the time it is created. This is in contrast to the color server product that builds image-specific device links on the fly during image processing.

Having said that, one question I would have for Mr. Dery is the "artificial intelligence" aspect of his device link product. Given that it's NOT a color server application but is instead software that builds static device links, how are your device links able to make "intelligent" (dynamic?) decisions on an image-by-image basis when the parameters of the device link have already been "hard-wired" into the device link at the time it was created? It seems a color server or workflow application of your own design would have to be in control of the device link creation process at the time of image analysis for this to actually occur but instead you actually tout as a *feature* that yours is not a dedicated color server product but is instead static device links to be used/inserted into other workflows where no such image analysis is present.

Care to comment?

Regards,
Terry Wyse
___________________________________________________________________________

Re: K in commercial printing
Posted by: Terence Wyse
Sat Aug 30, 2008 6:47 pm (PDT)

I completely agree with you on this Mr. Foerts. I for one do not believe the ICC device independent approach is the best for *all* situations. We've already discussed pressroom color management via device link profiles. Another case in point is inkjet proofing. Arguably the two "best" proofing products on the market today, ORIS Color Tuner and GMG ColorProof, do not use an ICC strategy at all (although you can if you wish). Instead, they each use there own flavor of what are essentially device link profiles. I've worked for a number of years with the GMG product, while at the same time using ICC- based proofing products, and can tell you in no uncertain terms that the product produces a superior visual as well as numerical/ colorimetric match. This is not necessarily an indictment against device independent ICC color management, only the realization that in CERTAIN cases a more proprietary approach has advantages. It's really about trading a certain amount of flexibility for a more focused/ dedicated color management solution where needed.

Regards,
Terry Wyse
___________________________________________________________________________

Re: K in commercial printing/history of Matchprint
Posted by: "Dan Margulis"
Sun Aug 31, 2008 9:26 am (PDT)

Terry Wyse writes,

Take either Monaco PROFILER or ProfileMaker, two "mainstream"
profiling applications....

No, thanks. You attacked Custom CMYK as "if not outright wrong, at least VERY suspect in my book" because it has foolish defaults. I pointed out that many other defaults, not just in Photoshop but in many other applications, are equally foolish.

Now you inform us that another app has certain sensible defaults. This is not shocking. Several hundred other examples of sensible defaults in Photoshop and elsewhere could be propounded. The question is how seriously we should take it when a default is not sensible. You have explicitly stated that Custom CMYK is "VERY suspect" because the defaults are bad. You won't take my contemporaneous example of a bad default, so try this.

The Shadow/Highlight command is likely the most important one introduced in the last three revs of Photoshop. I use it on close to half of my images. The defaults, however, are suitable only for badly underexposed photos. To me, this is no big deal: I just enter better numbers and save them as new defaults, just as is done in Custom CMYK.

If Custom CMYK is "if not outright wrong, at least VERY suspect in my book" then you're going to have to find Shadow/Highlight "VERY suspect", too. You can't have it both ways.

More to the point (and I think this is the CRITICAL bit), it's not
transparent to the user what the various ink sets are based on.
"Newspaper", is that SNAP, IFRA, what, and what version of SNAP or
IFRA would it be referenced to? We have no idea without, at the very
least, interrogating the L*a*b* values in the ink set and seeing if
they correlate to any current or legacy standards.

Yes. We have agreed on this many times on the list. There is no question that mastery of Custom CMYK is unreasonably difficult for the average user, and that, while it was an excellent module by 1998 standards, we deserve something better today. The market has been clear that it will not accept something *worse*, no matter how hard the Photoshop team tries to jam one down our throats. Third-party profiles that can't be manipulated *at least* to the extent possible in 1998, that's worse. The points raised in your paragraph above are certainly regrettable, but until such time as a credible alternative to Custom CMYK emerges, we are stuck with them.

Maybe that information is somewhere but it's not available in the UI.
And even MORE to the point, AFAIK none of the built-in ink sets are
based on CURRENT printing specifications embodied in ISO 12647 and/or
any of the current FOGRA and GRACoL/SWOP data sets. So, even if you
KNEW what the appropriate separation settings should be, you're still
unlikely to build a profile that's usable relevant to today's standard
printing specifications.

Always dangerous to say "*you're* unlikely to be able to" when the speaker means "*I'm* unlikely to be able to." If you're truly not able to do this, you lack the understanding to criticize the module.

If by "machine-generated", you mean a profile created from
measurements made with a spectrophotomer, then duh, yea, that's how
data sets are created from which to build profiles from.

Unless, of course, you have to create the profile based only on printed results of known images, without the benefit of anything that can be measured. This is a common, if not everyday, occurrence for those of us who do it for a living. How do *you* make a profile under these circumstances, when no swatches are available? Being able to generate a useful profile off images only, no swatches. should be a reasonable baseline requirement for anyone passing themselves off as a color management consultant IMHO.

Even the
L*a*b* values used in Photoshop's Custom CMYK must have came from one
of these "machines", it's just not clear what the printing conditions
were that generated these measurements.

I believe that the numbers were picked out of a hat and were entered manually. This would have been normal procedure at that time. I know that CMYK profiles at Linotype-Hell were produced in this fashion.

As far as the two profiles you mention, yea, I've checked them out
before and they are simply inappropriate profiles and appear to be a
duplicate of each other. What does this tell YOU exactly, that the
"machine" is at fault for measuring two identical press sheets or
proofs but then this data was used to build two identical profiles for
two quite different printing conditions?

It tells me that Terry Wyse has a double standard in evaluating profiling packages. Less than a year ago, you offered as a compelling reason for not allowing profiles to be edited the possibility that someone using a Custom CMYK-like package could generate misnamed profiles. This is actually rather difficult in Custom CMYK; the module resets the name whenever a change is made so that if the author wants to misname it it has to be a deliberate decision.

Nevertheless, you cite as an important reason for not allowing profile editing in Photoshop the fact that users may misname their profiles. If you insist that casual users--who have no known history of doing this--are to be prohibited from having access to something at least as good as Custom CMYK on the grounds that they might give their new profiles the wrong names, then you have to insist that Adobe itself, which *does* have a history of doing so, should be banned from releasing new profiles. You can't have it both ways.

I guess I would fault the
person "behind the wheel" for either being too lazy or too stupid to
know not to use the same set of characterization data to describe two
different sets of printing conditions. How in God's name does this
implicate "machine-generated" (your term) profiles?

It doesn't, unless a person has taken the view that the possible production of "too lazy or too stupid" profiles is a good reason to prohibit the use of something else. In that case, once again, logical consistency would require that person to insist that Adobe be prohibited from releasing future profiles.

Personally, I cannot approve of the adjectives used above. I am happy for any profiles that Adobe chooses to release, poor or otherwise. As these are prepared by programmers, and not by color experts or professional retouchers, it stands to reason that they may not always be the best. Of the four main ones, only one is marginally acceptable as is. But all four could be fixed up easily IF we only had at least as much functionality as we had to alter other profiles in 1998.

As to the rest of your comments below, I simply wish you would state
unequivocally whether you feel Photoshop's Custom CMYK "profile
creation tool" (my words) is up to the same quality standard as
dedicated profiling applications?

YOU are the one who stated unequivocally that I said that it was--when you know perfectly well, from many years of sparring over this topic, that I think no such thing. Similarly, in another post, you attack me for "the notion that these device link conversions are inherently damaging". Again, you know perfectly well that I never said or implied such a thing. The thread is about the use of these profiles to hose incoming files by reversing intelligently implemented decisions about black generation and replacing them with "good enough" color.

I am not prepared to "clarify" statements that have been intentionally misinterpreted, but note that J Walton correctly stated, "I don't think anybody is saying that Device Link profiles are evil in themselves. In fact, it's quite the opposite - the technology has tremendous potential. But the way it is being implemented by some people just seems fishy to me."'

Posting and crediting to me opinions that the author knows full well are the opposite of what I think, followed by demands for "clarification", are unfortunately a standard part of this list's history. You have not to my memory been desperate enough to do it until recently. That you now indulge in it is unfortunate because I and, I believe, many others here find your independent opinions valuable and worth responding to. I hope that you will return to that kind of posting in the future; if not I suppose the next step will be the usual, posting to the ColorSync list how Dan Margulis says that Custom CMYK is better than ProfileMaker, and how he says that device-link conversions are inherently damaging, and how he thinks that JPEG is higher quality than raw, etc., etc., followed by two weeks of sulking about what an unjust world it is for color management consultants that people who are interested in quality listen to him and not to us.

One thing is clear, at least to
me, is that you cannot generate a profile based on CURRENT North
American and International standards and printing specifications using
what's available in Photoshop today, period.

Again,the meaning of "you cannot" appears to be "*I* cannot". Some of us can, and do.

What I'm NOT saying is that any Photoshop user worth their retouching
salt needs to go out and drop 2 grand on a profiling application,
although if they were to do that, I strongly believe they would
benefit by it. What I'm saying is that a Photoshop user could be
perfectly happy using a standard suite of freely available profiles
that are based on current international printing specifications,,,

Agreed. They could also be happy (and many are) with Brightness/Contrast, or with master curves, or with Auto Levels. It's "Good Enough" color--for them. But "Good Enough" isn't good enough for many others.

Similarly, if a printer is given an untagged CMYK file and chooses to guess at what profile should be assigned to it, this may result in "Good Enough" color for some. If the printer decides to reverse the client's GCR decision, this may be "Good Enough," too. And if the printer, in the name of better communication, decides to keep the move secret from the client, maybe the color will still be "Good Enough".

While times are undoubtedly hard for color management consultants, and while it's relatively easy to sell a credulous printer a bill of goods, the question needs to be asked whether it is really a good idea for "color management" to become synonymous with "Good Enough".

....but if you need more than that, you need to consider your options.

Right. And thanks to the Photoshop team's intransigence and the unwillingness of color management consultants to force the issue, they're exactly what they were in 1998:

1) A free, stable, functional profile editor that's difficult for beginners to learn, and incompatible with existing third-party profiles, but which can generate the needed variants in seconds,
within Photoshop.

2) A $2,000 standalone piece of software that has more functionality but generates the needed variants in minutes, outside of Photoshop.

3) Use one of a gaggle of publicly-available profiles of generally poor quality on a take-it-or-leave-it basis.

And, since the options haven't changed since 1998, there's no secret as to what the high-end CMYK users have done and will continue to do. Go to Photoshop World next week, and you'll hear the speaker on CMYK file prep (and it isn't me) recommending Custom CMYK.

It's now been ten years. Don't you think that's long enough to have determined that the desire for quality is not going to go away? There's no great problem in stating that you've been wrong in the past; I do this all the time (although not on this issue; history has already passed judgment on it). So, you face a choice of whether to try to validate the original promise of color management to give the average user a shot at truly excellent quality, or to say that the possibility of getting some new business justifies putting yourself down in favor of "good enough" color and, thus, against history.

The difficulty with this position is that there are only a limited number of printers who can be persuaded to hose their clients in this fashion. As for non-printers, those for whom "good enough" is good enough probably don't require the services of a color management consultant, while those for whom it isn't are likely to shun anybody who, for example, says that reseparating a carefully prepared file produces "good enough" presswork.

Dan Margulis
___________________________________________________________________________

Re: K in commercial printing
Posted by: Dan Margulis
Sun Aug 31, 2008 9:27 am (PDT)

Stephen Marsh writes,

Jeremy, as much fun as it is to see the side topics raised in reply to
your post thrashed to death yet again on this list,

I agree that superficially this resembles past discussions (color management theorists coming up with some quality-damaging brainstorm that coincidentally provides work for color management consultants, assuring us that the result would be "good enough" if only the fantasy of all printers in the world being good ones were true, and informing us that if we do not accept the dysfunctional solution we are Luddites). At least in the past we were able to employ such parts of their recommendations as were useful and ignore the others. Here, as we have seen, the quality hit is imposed surreptitiously, as the printers who are trying to pull this fast one would never dare make it public. Consequently, it's a new topic and one that warrants discussion, although I hope we're nearing the end of same.

There is no basis in historical practice for printers to claim a unilateral right to alter a clients' black generation decision. Printers have no authority to do this without the client's explicit consent, any more than to open a client's image and attempt to color-correct it. If the printer has reseparated without the knowledge of the client, and the job has turned out muddy as a result, the client is within his rights to demand a rerun without charge.

Meanwhile, list members should be aware that certain printers are actually behaving in this irresponsible way. As a group, commercial printers know almost nothing about file prep of images and are not competent to try to outguess you as to what technique is likely to make a file print well. Therefore, even though reputable printers do not reseparate without client's consent, those who are careful about their black generations should be sure that their work orders contain a specific instruction that the printer may not significantly alter the balance between CMY and black without explicit permission, and indicating that failure to adhere to this instruction is grounds for rejection of the work.

Dan Margulis
___________________________________________________________________________

RGB-CMYK curving (was: K in commercial printing)
Posted by: Dan Margulis
Mon Sep 1, 2008 6:49 am (PDT)

Stephen Marsh writes,

With all this talk of reseparating CMYK at the printer, it would
appear that the "RGB only workflow" proposal of the 90's is being
repeated again, in slightly different form without the RGB benefits
and with the CMYK benefits perhaps destroyed. If my CMYK files that
are *prepared for the target condition and image content* are going to
be changed to "suit the target condition without my knowledge" - then
I may as well just provide RGB and not bother with the CMYK stuff (I
am in prepress, not photography).

That's a close analogy. In the mid 1990s *EVERY* industry commentator (but one) was on board with the ill-fated "RGB workflow". It failed for the same reason that this reseparation idea will: it damages quality. The differences are 1) that here, unlike the 1990s, the printers are behaving indefensibly, surreptitiously altering files to save a few bucks at the cost of quality and hoping that the clients are too stupid to notice; 2) one of the reasons that the 1990s "RGB workflow" failed is that it took away the advantages of correcting in CMYK (unless the operator wanted to convert to work there, then convert back to RGB afterward), whereas now it's being reseparated anyway so it doesn't matter what space we correct in.

Since this thread is threatening to get out of hand, let me segue into a new line that is more about what the group prefers to talk about. I am adequate at color correction myself, so it is sometimes hard for me to grasp how people with less experience learn how to be good at it. ACT has been phenomenally successful at bringing up people's skills over the years, and it's tough to change a winning formula. Nevertheless, I overhauled it at the start of the year so as to emphasize the picture-postcard workflow. I've now taught the new curriculum eight times and in doing so have gotten one very unexpected result, which I'll get to in a few paragraphs.

When the class began in 1994, it was a CMYK world. Not only was CMYK the nearly exclusive output space, but most files originated in CMYK. As years went on, it became more and more common for the start point to be RGB, and today it almost always is. Nevertheless, on the first of the three days of the class, we continued to work with CMYK originals, CMYK output. We did not start in RGB because that would have brought up the whole complicated question of how to separate for best press results, which is the topic of the other thread and which is IMHO too difficult for the first day of the class. Similarly, we did not correct in LAB until late on the second day.

CMYK output has always been, and still is, the place where the most money is made by image processing professionals. As I've noted here for several years, though, the number of RGB-oriented professionals has greatly increased, but an even greater increase is shown in those who have color-critical output in either one. I poll each class, and also at Photoshop World, where the results are similar, if slightly less CMYK-oriented. This year, my classes are the closest yet: 38% of students strongly oriented toward CMYK output; 32% strongly oriented toward RGB output; 30% with frequent need to do color-critical work in either.

In re-evaluating how to handle the first day, I wanted to introduce RGB and LAB earlier. I now am convinced that it's easiest to *evaluate* (not necessarily *correct*) the images in LAB, because the numbering system in the AB channels is so easy to work with. So the new structure is:
*All files are evaluated, by the numbers, in LAB
*Nevertheless, the first set starts and ends in CMYK.
*The second set starts and ends in RGB.

Now, the surprise: I had always thought that people growing up with RGB would be more comfortable working with it, and CMYK groupies likewise. Again, we do a CMYK set first, and then a similar RGB set. Unexpectedly, at least half a dozen RGB-oriented people have exclaimed after evaluating the second set, "I had no idea that working in RGB was so limiting." Granted, these people are not considering channel blending yet, which *does* work well in RGB. They are confining the comments to curving and sharpening.

Based on this fairly decisive feedback, I'm wondering if I should revisit the question of starting work in an artificial wide-gamut CMYK, so that the files can be converted to RGB later without any loss of color. I proposed that a long time ago, and some list members have said they use that workflow, but up until now I've thought that it was too cumbersome to bother teaching.

Just food for thought.

Dan Margulis

Back to top
___________________________________________________________________________

Re RGB-CMYK curving
Posted by: "Mike Russell"r
Mon Sep 1, 2008 1:46 pm (PDT)

My own experience tends to say to me that wide gamut CMYK would be an interesting teaching tool, but that it is probably not going to cause much of a ripple in people's work habits, compared to Lab. My customers are drawn from a rather wide swath of non-professional Photoshop folks, and tend to be less experienced, so YMMV.

Curvemeister started life as a lossless wide gamut CMYK (wgCMYK) curve plugin, with additional features suggested by Dan. RGB, Lab and HSB were included to round out the software a bit because Photoshop supported the color conversions internally.

I soon discovered that I needed to provide an online class to get across the concepts of neutral, shadow, highlight, selective contrast, etc, and point people towards Dan's books as a further reference. Lab was, by far, the color model that caused the most light bulbs to turn on for people. Folks who had used nothing but RGB would simply bear hug Lab and not want to go back. Dan's Lab book made this tendency even more pronounced.

Compared to Lab, wgCMYK has remained a bit of a step child. Though important for concepts like the unwanted color, and control of shadows via K, not many people take to it as their main working space. The wgCMYK color model continues to cause confusion for people just getting their feet wet in CMYK. For example, they are often disappointed that the retention of saturated blues is only an "illusion", since this color space is not designed for printing. This could simply be a defect in the way I present wgCMYK to new people, and I for one would be very interested in any thing Dan had to offer in this regard.

Mike Russell - www.curvemeister.com
___________________________________________________________________________ .

Re: Re RGB-CMYK curving
Posted by: Dan Margulis
Mon Sep 1, 2008 4:50 pm (PDT)

I'd say it would have to be limited to those who are quite serious about color. A lot of people, John Ruttenberg being one who comes to mind offhand, have suggested that people should be taught to correct in LAB from the get-go, because it's conceptually simpler than RGB or CMYK. That's true enough in a complete vacuum, where people have no experience at all in either of the other two. But in real life, most beginners have at least a passing familiarity with RGB, which throws the equation off. Also, some files aren't appropriate for LAB correction.

I can understand--today--why novices give a "bear hug" to LAB as you describe. But I didn't always understand it, because I'm not in their shoes. Right now, what I'm teaching is that going into LAB while there are still casts in the original is overly risky, so there should be some minimal curving in RGB first. But I'd be open to doing it in a different way, understanding that it would be a niche workflow.

Ultimately I think we probably will go beyond current models, because improvements in people's ability to manipulate images have progressed faster than improvements in Photoshop (or AFAIK any competitor). With computers being so fast nowadays, we don't have to stick to traditional colorspaces--we can invent them on the fly, if we like. Particularly, there are now so many ways of enhancing contrast outside of traditional models that I have to expect that sooner or later we'll get raw modules (or plug-ins like Curvemeister) that can stack up a lot of luminosity and similar corrections, blends as well as curves. Hopefully sooner.

Dan Margulis
___________________________________________________________________________

Re: Re RGB-CMYK curving
Posted by: "mdjager"
Tue Sep 2, 2008 4:02 pm (PDT)

Hi,

I'm just a lowly photographer who's been interested for some time with increasing my knowledge of color correction and just for my own work.

I started with Dan's PPS3 and back then I'm pretty sure I tried to skip over as much of the CMYK stuff as I could and start with the RGB stuff, which is where I felt much more comfortable. CMYK scared the daylights out of me!

I kinda stayed that way into ver.4 of Dan's book but tried harder to make some sense of CMYK. Mind you back then I also liked to work with Levels!

Then Dan's LAB book came out and I was in heaven! I started to do everything in LAB. I remember reading Dan comment, more than once, that LAB was a 'difficult' space to learn because it was so different from anything else. Personally, I thougth he was nuts! For me the concepts surrounding LAB made total sense and seemed to come naturally. Unfortunately, I also started to believe that RGB wasn't necessary anymore and therefore how could CMYK possibly be useful for anything.

Then the Picture Postcard Workflow came out and I saw the light for the first time ever! Channel blending, Shadow/Highlight, Overlay Blending, etc., etc. all started to make sense. Since the introduction of the PPW I've actually gone back to PPS5 and have begun studing it again. And an amazing thing has happened since the last time I opened its pages... so much more of it makes sense now!

Finally, finally I'm starting to feel confident when I open up a photograph in Photoshop and hit the Curves button (why did I ever think Levels was better!) Plus now I actually study the individual channels in an attempt to understand what's happening in the image as a whole.

Now, I still wouldn't want to meet Dan in a dark (greatly underexposed) alley armed only with a copy of Photoshop and a mouse. But I would feel confident taking on any first year photography student from any of the finer institutions here in my home town!

One final thought bothers me however, if Dan is still learning this stuff how can I ever hope be as smart as him!

Murray DeJager
___________________________________________________________________________

Re: Surreptitious Conversions
Posted by: Dan Margulis
Mon Sep 8, 2008 11:48 pm (PDT)

Greetings from the California desert, where I am recuperating from Photoshop World by hiking in 110F/44C degree heat and preparing for a class later this week in San Diego. In principle I am on vacation from the list and enter it only because of the request for anonymity in the forwarded post below.

As this is the second prepress manager who has contributed to the thread in this way let me reiterate ground rules. We do not generally grant anonymity unless it appears that the poster cannot speak freely without it. There is more question in my mind about whether the following qualifies for that than was the case with the first manager, who stated outright that the purpose of the surreptitious conversion was to save money and that management understood that quality would suffer. However, I suppose that clients and bosses of the following poster might object to the sentiments he expresses, so I give it the benefit of the doubt and delete his name.

Two reminders: when I am doing this I am personally vouching for these people having the jobs they say they do. If anybody else is planning to do this, you have to be sure that I can verify who you are.

Similarly, although there was an exception made for the previous post, right afterward I stressed that someone who is posting anonymously loses the privilege to make negative comments about other list members. The post must be entirely constructive. It's not fair to attack people who can't see the attacker's name.

Dan Margulis

*******************************************
From:
Date: September 8, 2008 3:31:46 PM PDT
Subject: Re: [colortheory] Surreptitious Conversions

On Aug 31, 2008, at 7:19 AM, Dan Margulis wrote:

I agree that superficially this resembles past discussions (color
management theorists coming up with some quality-damaging brainstorm
that coincidentally provides work for color management consultants,
assuring us that the result would be "good enough" if only the
fantasy of all printers in the world being good ones were true, and
informing us that if we do not accept the dysfunctional solution we
are Luddites). At least in the past we were able to employ such parts
of their recommendations as were useful and ignore the others. Here,
as we have seen, the quality hit is imposed surreptitiously, as the
printers who are trying to pull this fast one would never dare make
it public. Consequently, it's a new topic and one that warrants
discussion, although I hope we're nearing the end of same.

There is no basis in historical practice for printers to claim a
unilateral right to alter a clients' black generation decision.

Meanwhile, list members should be aware that certain printers are
actually behaving in this irresponsible way.

If this is actually posted, please do so anonymously. This is "inside printshop" testimony that may be misinterpreted by some.

I understand that it's past bedtime for this thread but I would like to comment on the subject of irresponsible and reckless printers. Dan, it would be wonderful to have the time to communicate with every print customer directly and intelligently. It would also be more rewarding to work on all print projects with this level of personal involvement. Perhaps "good enough" color became the substitute and solution for this time shortage. It's odd to say this since so much of this thread stresses communication. The color consultant mantra is communication, education. Would it have ever occurred to you that part of this communication and education would revolve around the subject of changing the color you had engineered for your pictures? Just how does one communicate this tricky bit, informatively and in less time - and without arousing suspicion? It sounds counter- productive, doesn't it? Especially since this is a conversation that you don' want to have happen in the first place. I think so too, so this shop does not do conversions, except when requested to make the RGB-CMYK conversion.

The "progress" of color management has moved printers into quite a difficult place. Prior to the color management initiative, a press curve or similar "calibration" was the only alteration made to a file submitted for print. This was done out of necessity. When printers began seeing more and more Photoshop users preparing files, then the press curve was adjusted to accommodate, as best it could, the new photoshop files as well as legacy files. That those separations had similar dot gain meant that only a small adjustment was needed - and no significant change was ever made for black generation. Black generation remained whatever it was in the customer's file. Then things changed.

Printers became labeled as irresponsible, antiquated neanderthals in trade publications if they were hesitant to adopt the ICC color management initiative. According to color consultants, change was necessary. A lot of print shops didn't see the necessity. It still isn't as far as I'm concerned. Nice tools, but not absolute necessities. This early period was a time of total exasperation for both printers and creatives. Sadly, the same state of perplexity exits today - only now there are even more complications, and fewer simple solutions. But remember, the old way was condemned.

If any change was really "necessary", it was only necessary on one end - upstream at the creative end of the flow. It was/is necessary for them to be able to make CMYK files. Those files can be honored. This puts the creative in the driver's seat. But now, everybody is responsible for color, with the exception of creatives who believe that color management will magically take care of any issues. Color management dumbed down the act of creating separations - we have been witness to this for some time now.

There have always been some very good reasons to avoid the kinds of complexities that have been offered by color management initiative. The very topic of this thread evidences the angst of creative whose files have been regurgitated without their knowledge. All of the technical problems with this scenario are known so I'll not list them. But on the printer's side, where margins are slim, the simplest solution of all was already in play before "color management" arrived on the scene. There was 1 workflow. The customer supplied CMYK, the separations were slightly but uniformly nudged, and that was that.

"Conversions" became another completely different workflow for printers to consider. If simplicity were their goal - "honoring" file as submitted, they would be condemned for not adopting "modern" print methods. Note that there are a couple of definitions of "honoring" a file - one involves conversion, another doesn't. The new color initiative effectively pitted printer against printer in a war for the "most modern printer" trophy. Process control had been king for ages and was sufficient for accomplishing quality printwork. Then conversions became the rage.

A lot of the newer print customers have grown up with this color initiative, but most still know very little about it. They understand that it means "push-button perfect color every time". They only know what they have read from its proponents, and if a printer hesitates or acts "uncomfortable" with the initiative - that printer won't get their business. I have personally been berated just for mentioning color management's real and actual shortcomings - by customers no less! After careful and delicate and time consuming consultation, the smart ones catch on, but the hyped-up ones don't ever get it. It doesn't matter that a printer might have more knowledge in this area than the customer if the customer has already bought into the hype of push-button color. But I'll tell you this: I can't remember the last time a customer said he just wanted "good enough" color.

In an effort to scratch everybody's color itch, printers are faced with an ever growing number of possible workflows and solutions. My gosh at the number of solutions. And now, with all of these solutions, printers can be condemned for re-separating and altering black generation. Ironic, isn't it? Condemned - either way.

What is to be thought of a print shop that does not automatically convert incoming files? Responsible? Irresponsible and behind the times? I guess it depends on who you ask.
___________________________________________________________________________

Re: Surreptitious Conversions
Posted by: Dan Remaley
Tue Sep 9, 2008 2:18 pm (PDT)

Thanks for sharing this Dan M. - I agree with the observations for the real world color. I don't even know the term. . . 'pleasing' color. . .

Dan R.
___________________________________________________________________________

Re: Surreptitious Conversions
Posted by: Dan Margulis
Thu Sep 18, 2008 8:09 am (PDT)

While I was away, Dan Remaley wrote, in reply to a prepress manager's comments about surreptitious conversions,

Thanks for sharing this Dan M. - I agree with the observations for the real
world color. I don't even know the term. . . 'pleasing' color. . .

FWIW I agree with the comments as well. Everybody agrees on the desirability of "communication" between printer and client as to what each is up to, but in the real world it is not always possible. Under these circumstances either side may be justified in making assumptions about what the other wants--if it is not obvious.

In the actual thread, though, the practice being discussed was a deliberate defiance of the clients' wishes, by a surreptitious CMYK>CMYK conversion that is designed to save the printer a few pennies on ink at the expense of the quality of the final product.

At Photoshop World I spoke to Taz Tally, who instructs on CMYK file prep and who consults actively, about this issue. He confirmed that he has occasionally encountered this type of surreptitious conversion. His view on this is substantially the same as mine, which is:

1) A printer who behaves this way is acting unethically. There is no historical basis for a claim that printers have the right to completely regenerate a client's channel structure, particularly when the client, by refusal to embed a CMYK profile, has made it clear that he does not wish such a conversion to be made.

2) The common practice of using platesetter curves to slightly lighten or darken individual channels has nothing to do with the issue of surreptitious conversions, which completely alter the relation between and the character of individual channels.

3) Typical printers are not competent in prepress and should not be trusted to make such conversions without clear demonstrations of their ability in advance.

4) A client who discovers that such a surreptitious conversion has taken place is justified in rejecting the work if he finds that it came out too muddy.

5) If the printer discloses in advance what he plans to do, then there is no problem, as the issue can then be discussed intelligently, and the two parties can decide what's best for both.

Dan Margulis
___________________________________________________________________________
.
Re: Surreptitious Conversions
Posted by: Michael Jahn
Thu Sep 18, 2008 11:30 pm (PDT)

Hi Dan,

sur·rep·ti·tious
---- done, made, or acquired by stealth : clandestine - acting or doing something clandestinely : stealthy.

Sounds so mean spirited !

I think my big disconnect here is that I may be in a different vertical - in a world where people are designing ads to then be sent for insertion in a magazine or newspaper. These users rarely do any of this channel magic that you are doing. Worse, in the case of a newspaper, most color ads enter the facility set up for SWOP, and absolutely would look horrible if some change were not done to the separations. Are these changes made "surreptitiously" ? Absolutely.

you wrote;
In the actual thread, though, the practice being discussed was a deliberate defiance of the clients' wishes, by a surreptitious CMYK>CMYK conversion that is designed to save the
printer a few pennies on ink at the expense of the quality of the final product.

I agree! If you were creating the image for something like a poster and were working with a print service provider that you know - who is using a particular commercial press set up in a way you are familiar - then of course, using your experience, you would insist they they honor your very carefully made separations "as is". If they did not, you would not pay them and find someone who understands what you are trying to do.

I will assume that you fully understand that "use my seps as is" request goes completely out the window (that is, the file will be re-separated) if they print this on some 'other' printing system (large format InkJet, HP Indigo, Xerox iGen or Kodak Nexpress) as these systems do not use the inks or densities your carefully constructed channels required.

It is in that moment - where the art and images that are separated - that I wish we could re-do forever - that is, come up with a "real world" image manipulation space (RGB/HSB ?) and then send THAT with an output profile.

3) Typical printers are not competent in prepress and should not be trusted to make such
conversions without clear demonstrations of their ability in advance.

I have no idea as to what a typical printer "is" - but we can agree that there are many printers who will always blame "bad separations" when you don't like their press sheet. It certainly goes the other way sometimes - as I described above, people send PDF/X1a CMYK pdf files to newspapers every day, and they were designed for a SWOP printing condition - and when this 'it was sent to us wrong' scenario happens, then 100% of the time it must be re-separated else it looks like mud.

So, I guess I remain on the side of 'we will continue to surreptitiously re-separate your PDF files as required" camp.

--
Michael Jahn
___________________________________________________________________________

Re: Surreptitious Conversions
Posted by: Henry Davis
Fri Sep 19, 2008 9:59 am (PDT)

On Sep 18, 2008, at 12:08 PM, Michael Jahn wrote:

It is in that moment - where the art and images that are separated -
that I wish we could re-do forever - that is, come up with a "real
world" image manipulation space (RGB/HSB ?) and then send THAT with an
output profile.

Sorry, I'm having trouble following this part. The final color space is where one will most likely apply some noodle or small correction, or perhaps a more bold adjustment to black. This final print version is what this thread has focused upon. The manipulation space (RGB/ HSB) you describe is not the final version that will print, right?. In the case you mention it is sent with an output profile - which I'm guessing would be CMYK if there is a rip involved. I can't understand how on earth that a yet to be separated version could be the best solution.

What's the problem with making the separation using the output profile, noodle as needed, and then send that as the print version? Then, If it turns out that further adjustments are desired, wouldn't it be more simple, direct, and less convoluted to make these adjustments to the final CMYK version rather than some prior color space?

There seems to be an underlying assumption that there never needs to be any further adjustments made to the final version - that the final conversion to CMYK via profile will yield the most desirable version of separation.

Getting really "real world" about this, the necessary noodling required for matching Pantone 4 color process simulations is wacky when attempted in anything but the final print version. And yes, there are cases in picture elements where specific color matching is critical - they are not just limited to vector elements.

I still can't help but thinking that the most simple approach is the best - supply the print version with no further conversion downstream. As you point out, this has been made less than easily possible and sometimes impossible in some systems. It has been made so on purpose. The conspirator in this matter is not always the printer. Developers and proponents who sell the idea of push-button color are the prime movers at the center of this plot.

Henry Davis
___________________________________________________________________________

Re: Surreptitious Conversions
Posted by: Michael Jahn
Fri Sep 19, 2008 4:43 pm (PDT)

@ Henry Davis

On Fri, Sep 19, 2008 at 8:29 AM, Henry wrote:

Sorry, I'm having trouble following this part. The final color space
is where one will most likely apply some noodle or small correction,
or perhaps a more bold adjustment to black. This final print version
is what this thread has focused upon..

Yes, you are correct ! - I apologize profusely for the confusion here. My bad !

- you are correct in your statement that the thread is about whether or not the final set of color separations submitted should be allowed to be modified.

I believe that the folks - like Dan - are saying "never ever modify our separations" - and while I have suggested this would be nearly impossible to comply with when one needed to print your separations on several popular digital print systems (iGen, Indigo, nexpress or nearly any and all inkjet printers) - what I described was "off" thread (crossed threading is NEVER good) in that I was describing a scenario that was more about the "Selective Color" thread - I am fairly sure there would be no method available in Photoshop that woould enable a user to modify channels in an nUp image (Like Hexachrome) without a special plug-in (like Aurelon CoCo or Pantone Hex-wrench)

I still can't help but thinking that the most simple approach is the
best - supply the print version with no further conversion
downstream. As you point out, this has been made less than easily
possible and sometimes impossible in some systems. It has been made
so on purpose. The conspirator in this matter is not always the
printer. Developers and proponents who sell the idea of push-button
color are the prime movers at the center of this plot.


Yes - agreed - I think an example might be in order - to retort if i might..

Imagine you get a PDF that is indeed a 6 color PDF - that is -- lets say -- designed to be printed in Hexachrome.
 But you do not have Hexachrome inks. In this case (of course) you have many choices such as converting back to some color space (like LAB) and re separating (pehaps into so other set of separations) - or something. to do that secretly and not tell the client, well, that would be probably a really bad idea, but it can (and has) been done.

In many instances, push button color should be embraced and promoted.

In far fewer, it will never ever work to meet a persnickety customers requirements.

Perhaps I live in a much faster paced vertical, where that magazine or newspaper forces more push button like conversions where economic manufacturing trumps artist deepy concerned with the nuances in a image.

I guess I became jaded with the advent of PageMaker and Quarks H&J engines, when I saw the fine art of typesetting become completely unavailable almost overnight. While InDesign offers some of the technology that was available to Xyvision, Atex and Penta, must publication typesetting is horrid.

*sigh*

--
Michael Jahn
___________________________________________________________________________

ANDTWOMONTHSLATER, the original poster reported in with the results of his job. As you can see, he would have been happier with what he got had he followed the advice I gave. -DM

Re: K in commercial printing
Posted by: "Jeremy Stephenson"
Tue Nov 25, 2008 9:39 pm (PST)

This is a reply to an old post that I started several months ago - in particular, to reply to some questions by Stephen Marsh...

I have been in the process of having a book of my photographs printed - my first real foray into CMYK offset printing. I started two different topics here enquiring about the use of custom CMYK and different methods of black separation, based upon what I've read in Dan Margulis' books. One question was regarding the issues of having large black background areas and the possible influence on the amount of black ink laid down in adjacent image areas.

I apologise for the long delay, Stephen - I was extremely busy trying to make the press deadline, but now I have received both digital proofs and copies of the final offset print job, I think I can make some objective comment on what I ended up doing and the results on paper.

The two topics I started effectively set off a couple of wars in the colortheory group, and I hope that doesn't start again! However, as a newbie, I feel I've learned a thing or two, and others can benefit from that.

My printer is based in Singapore, and instructed me to use the FOGRA cmyk profile found in photoshop for my separations. I was concerned about potential issues Dan writes about when using stock cmyk profiles, where if the black ink comes down too heavy, it will block up shadow detail and muddy/desaturate colors. He advises using custom cmyk to create a separation with lower black levels. The responses received on the thread went strongly in every direction - both for and
against using the supplied profile.

In the end, trying to juggle all the choices, and with the deadline looming, I decided to just use the standard FOGRA profile supplied in photoshop and advised by my printer. I did not feel confident playing with other settings, particularly with known bugs in the way Custom CMYK works. I decided that for my position, at least the standard profile was a known starting point, and I would just have to see how it went. I did not have the time to experiment more with other options.

My printer advised me to use 100K, 30C for rich black background panels, so I used that option (printing was on a matte art paper by the way). I ended up staying with 4 process inks rather than using a separate black ink for background panels - the price quoted for 5 inks added 50%.

When I received the proofs back from the printer, I was very happy with them - I felt they were a close approximation to what I'd seen on my calibrated 23" Apple cinema display, so no problems there at all.

Yesterday I received a copy of the printed body of the book - as yet unbound. The paper quality is very nice, and images are sharper than they were on the more glossy stock the proofs were printed on. However - the images were generally darker than the proofs - particularly in shadow areas, where I have lost some shadow detail, and the colors are generally less vibrant. Seems to be a little of what I was warned about when using standard profiles?

I have also noticed that the effect on color, and plugged shadows, is generally worst where there is a large surrounding black background - also what I was warned about.

However - my 4color black and white, and toned images look beautiful - full detail. Obviously the heavier black of the standard separation is healthy for those areas.

So, I am a bit disappointed in some pages - but I guess that's life, given the choices I made on this job. The printer's claim is to match the proofs to 85% - whatever that means, and however that is measured. I'm sure given such leeway, what I've received falls within their tolerances. And it's quite possible that my customers will be quite happy with what they see, never knowing the full potential that I knew was within those images.

I think that had I tried fooling around with custom CMYK, I would not really have known what factors worked or did not. Now at least I have a visual reference for the issues I can expect using a standard profile, and I can work around those in future to try for better results.

I suspect the black printed heavier than intended, and I also suspect the dot gain gave me a darker result than I was expecting from the proofs.

In future, I would do a lot more experimenting with the printer beforehand, asking them to do some samples on both their proofing machine and the press - and maybe even try profiling their press conditions. But I am now confident that I do not want to trust a standard profile, or even the standard proofs as supplied by a printer too literally.

Oh - by the way - I did decide to have one file printed with 5 inks - the standard cmyk, plus a second k for background black. This is for the dust jacket, which has a lot of black in it too. I felt that as this is the first view of the book, it had better be as good as possible. My printer did suggest that there would be negligible difference between using 4 and 5 inks, but I decided to do it anyway for the dust jacket. I have not yet seen final prints of this as it is still being printed, but as mentioned earlier, I did see negative impact from black background in other areas within the body of the book. As a test, I am having the dust jacket printed 5 ink, and the hard cover, which has an identical design, is being printed with 4 inks. I will report on any differences when I see the final product.

Thanks to everyone for their various input - it has been helpful - from both sides of the fence.

Regards,
Jeremy Stephenson
___________________________________________________________________________

Re: K in commercial printing
Posted by: Stephen Marsh
Wed Nov 26, 2008 8:08 am (PST)

Jeremy Stephenson wrote:

I apologise for the long delay, Stephen - I was extremely busy
trying to make the press deadline, but now I have received both
digital proofs and copies of the final offset print job, I think I can
make some objective comment on what I ended up doing and the results
on paper.

Thanks for getting back to the list Jeremy, feedback is important and it is rarely given.

The two topics I started effectively set off a couple of wars in the
colortheory group, and I hope that doesn't start again! However, as a
newbie, I feel I've learned a thing or two, and others can benefit
from that.

I am sure that Dan will hate to be proven correct! Dan writes about this and illustrates the topic in his books and has made postings here. He is not noting this issue for himself. If the printer is not radically changing the supplied colour build/K ratio of your images, then the suggestions that Dan made would have reduced these issues. Jeremy, I understand why you went the way you did and that you are not complaining when you were advised differently upfront. This has been an illustrative exercise for you, which you can keep in mind for similar future jobs.

Yesterday I received a copy of the printed body of the book - as yet
unbound. The paper quality is very nice, and images are sharper than
they were on the more glossy stock the proofs were printed on.

I presume that the offset printed book is using halftones, while the proof was a stochastic dither inkjet output.

However - the images were generally darker than the proofs -
particularly in shadow areas, where I have lost some shadow detail,
and the colors are generally less vibrant. Seems to be a little of
what I was warned about when using standard profiles?

Quite likely! I presume that process control was good in some areas, from what you mention below about 4C grayscale images. That being said, this is is not a process control issue, it is a known effect of offset litho printing that should be accounted for in prepress.

It is not so much about a "standard profile", it is more about using a one size fits all separation profile. It is also about looking at the colour build of the files critical areas and evaluating how this will translate on the press when the image is included in a page layout, not to mention how the pages will be imposed for printing and how this may further change results from the expected result. This obviously takes experience and some skosh, it is not built into the CMM or ICC profile.

A lighter GCR separation and shadow build for these "framed" images, if not all images was originally recommended from experience, not theory. Experience may also guide one in making a minor tweak to the K channel, on top of using the lighter separation. The theory of GCR is sound and in proofing it should be hard to see the difference. The practice of ink on paper often tempers the theory. Dan wrote a couple of Make Ready articles on "The Unlucky Expert" which was all about such defensive driving tactics (which are not foolproof, they just increase your odds).

However - my 4color black and white, and toned images look beautiful
- full detail. Obviously the heavier black of the standard separation
is healthy for those areas.

The GCR in such a one size fits all profile is not designed for helping a press operator maintain a neutral 4C grayscale images.

Do these images appear neutral, or are they warm/cold casted? Does the cast vary, between images or within the same image over different tonal ranges?

If casted, is this more apparent in the highlights to midtones than the midtones to shadows?

Did these images have the solid black + cyan tint panel around them?

As a test, I am having the dust jacket printed 5 ink, and the
hard cover, which has an identical design, is being printed with 4
inks. I will report on any differences when I see the final product.

I suspect that the differences may well be due to stock. Is the dust jacket and the hard cover the same substrate (I am guessing not)? You will need to compare the 4C vs 4C+K Bump on the same paper, otherwise the "test" is comparing apples to oranges.

Sincerely,

Stephen Marsh
___________________________________________________________________________

Re: K in commercial printing
Posted by: "Todd Shirley"
Wed Nov 26, 2008 8:08 am (PST)

Hi Jeremy

Thanks for reporting back your results. It is an interesting story and real world results are always useful. I just have a few comments.

So, I am a bit disappointed in some pages - but I guess that's life,
given the choices I made on this job. The printer's claim is to match
the proofs to 85% - whatever that means, and however that is
measured. I'm sure given such leeway, what I've received falls within
their tolerances.

I guess you get what you pay for. I assume you went with a printer in Singapore because of the price, which is reasonable, but "to match proofs to 85%" sounds like a ridiculously wide tolerance. With that much leeway you could have gotten back much worse results and still been "within tolerance". The print jobs we prepare for our clients usually have pretty big budgets and go to pretty good printers with proven track records, and if their results were even at 95% of their proofs the client would be unhappy. Although I shouldn't be, I'm a little amazed that a printer could produce "proofs" and then say we are only going to be able to match these 85%. That's not really a proof at all! I expect every printer I deal with to be able to match their own proofs, even the cheaper ones!

In future, I would do a lot more experimenting with the printer
beforehand, asking them to do some samples on both their proofing
machine and the press - and maybe even try profiling their press
conditions. But I am now confident that I do not want to trust a
standard profile, or even the standard proofs as supplied by a printer
too literally.

Well, press proofs and print runs to create a profile cost a lot, and it doesn't sound like you have a tremendous budget for that kind of thing. And if a printer can't even match their own proofs, then there is no reason to believe that they have the necessary process controls to make one print run match another. I would seriously doubt that if you got a reprint of this book in 6 months it would match the first run, so that makes profiling somewhat pointless.

Please don't let one bad experience make you throw out standard profiles and trusting printer's proofs. The foundation of quality printing is that a printer CAN match a known standard profile (especially one that they recommend!) and that they CAN match their own proofs. I understand that their are many printers who can't accomplish even these very basic goals, but I know for a fact that many many printers can do these things on a regular basis, and it really should be the bare minimum. I know real world considerations force us sometimes to use lower quality vendors, but unless your budget was seriously miniscule, I would expect that you should be able to get better results for the same money.

Here's the thing—second guessing the printer and not trusting their proofs will only make the problem worse. The hallmark of a crappy printer is that they have no process control, which means that any given print run is to some degree unpredictable. They don't want to spend the time and money to make sure they can get consistent results. Therefore, the next time you print with them they MIGHT match their proofs perfectly, and here you have lightened all the images because you thought it would print dark. Or maybe they ended up printing LIGHTER than the proof, and your pictures are all blown out. Maybe if you printed several jobs with the same printer and began to see a pattern emerging, then you could make a reasonable adjustment, but to draw any conclusions form a single print run is useless.

On the other hand, since I can't see your results, perhaps they aren't really that bad. If you know anyone who does have more experience with 4/C printing, you should show them the proofs and printed pieces and see what they think of the quality. Based on your 85% comment, I'm assuming that I wouldn't find the results acceptable, but maybe I would. Good luck with the next job!

-Todd Shirley
___________________________________________________________________________
.
Re: K in commercial printing
Posted by: Michael Jahn
Wed Nov 26, 2008 10:11 am (PST)

Hi Jeremy,

a question (if I could ?)

you wrote;

"My printer is based in Singapore, and instructed me to use the FOGRA cmyk profile found in photoshop for my separations."

-- Which one of the FOGRA profiles ? (coated 27, Coated 39, Uncoated 29, Web Coated 28, something else ?)

A few "red flags" (things that made me go "hummm..") appear to have happened here that makes this suggested setting immediately suspect.

- you write "printing was on a matte art paper" - then later you wrote "The paper quality is very nice, and images are sharper than they were on the more glossy stock the proofs were printed on.." - is "matte" equaly to "uncoated" - perhaps this is not the case for you...but I think everyone on the list certainly understands how paper will change how a single image (and a single set of color separations) print.

Each paper stock possesses the following characteristics: surface texture, brightness, color, whiteness (if you specify a white sheet), opacity, grain direction, weight, bulk, caliper, and size.

Uncoated and coated paper have different surface textures. In the papermaking process, uncoated stock has been compressed between metal rollers (calendared) only to a limited degree, yielding vellum, antique, wove, and smooth surfaces (from rough to smooth, depending on the amount of calendaring). Coated paper varies from roughest (matte) to smoother (dull) to smoothest (gloss), also depending on the amount of calendaring. Papermaking machines can even impress such textures as "linen" and "canvas" on paper. The smoother the paper, the better the "holdout" (the better the ink sits up on the surface of the paper rather than being absorbed into the fibers)

So - reading your detailed email - there was a noticeable difference on the paper stock (perhaps even the printing condition) between proofs and the press work, yes ?

The FOGRA profiles represent very specific print conditions. These are not one size fits all - and while I am not interested in some protracted debate about which one you should have used (only the printer should tell you that) and what they suggested accurately represents the printing condition they have - well it sounds like that this may not have been the case.

To better understand the differences ;

http://www.color.org/FOGRA.xalter

Beyond this single selection (of some Forgra profile), as you know, there are several other selections that are important (like, for example - in Photoshop, under the edit menu, when you selected Color Settings, what was used for Conversion Options > Intent ? Was Black Point Compensation on or off ?

All I might add here is that it seems to me if the the proof is on glossy stock and they printed on matt stock, or the other way around, that this makes it very tricky for you to actually see what you will actually get unless they really are simulating the print condition extremly accurately, which does not seem to be the case here !

Good luck, and hope this helps yu for next time.

--
Michael Jahn
___________________________________________________________________________

Re: K in commercial printing
Posted by: Terence Wyse
Wed Nov 26, 2008 4:58 pm (PST)

Comments/questions:

1) They said to use the "FOGRA cmyk" profile. This is pretty much meaningless. There are a multitude of Fogra data sets available (Fogra 40 or 41 is the latest I believe) for all different printing conditions and paper types. Did they mean "ISO Coated" which is built from the Fogra 39 data set? That would be a good assumption but ppossibly an incorrect one...

2) You said "printing was on a matte art paper". This says to me "uncoated paper" for which the ISO Coated/Fogra 39 profile would be inappropriate for. There's probably nothing in the Fogra data sets that is a good match to to your matte art paper but Fogra 29 (ISO Uncoated) would probably be the closest. In any case, ISO Coated would have too high of a total ink limit for that paper unless you used the 300% total ink variant of that profile that is provided by the ECI. Even then, I think 300% is still a bit high for an uncoated matte art paper.

My point is "Fogra cmyk profile" doesn't really mean anything...or it means SOMETHING, just not anything specific. : -)

Regards,
Terry Wyse

______________________________
Terence Wyse, WyseConsul
Color Management Consulting
G7 Certified Expert