Dan Margulis Applied Color Theory - Reverse Engineering Seps

Date: Thu, 17 Oct 2002 09:07:32 -0700
From: Lee Varis
Subject: Reverse Engineering Seps

Hi all,

Just wondering if anyone had a strategy for reverse engineering a separation set up. I'm trying to figure out how to provide files for a client that has always had problems with digital camera files at this one prepress house. I provided seps based on scanty info from said house--they output, made Fuji proofs and the resulting prints were flat (not what I expected, I had a black in those files that was 70c, 60m, 60y, 90k!)

They corrected the files (or said they did) and everything printed great. The info I was given basically matched Photoshop's default SWOP set up (300 TIL, 100k, 20% dg) and I was just a little suspicious because the client showed me versions of these same files, printed before with the same prepress house that looked dark and muddy (like the dot gain was higher than expected) so I was asked to correct for this and put some life into the files. I re-separated from the original RGBs with a higher dot gain curves and reconfigured the black plate to increase contrast and pop out the black in the shadows (these were shots of motorcycles with engine parts with dark line detail in low tones).

The problem I'm having right now is that the corrected digital files seem to have disappeared - the sales rep fro the prepress house said he needed them for billing. So I don't have the corrected version only the original version of these digital files. Right now I'm thinking that this place messed up the original files so that they could charge to correct them but I can't really know until I can compare my original to their corrected file--or can I?

Any thoughts on this?

regards,

Lee Varis
http://www.varis.com
888-964-0024

Date: Fri, 18 Oct 2002 11:17:18 -0000
From: Stephen Marsh
Subject: Re: Reverse Engineering Seps

Lee Varis asks:

> Right now I'm thinking that
> this place messed up the original files so that they could charge to
> correct them but I can't really know until I can compare my original to
> their corrected file--or can I?

Hi Lee, I would like to help - or at least try to offer some thoughts, but I am having problems understanding the situation (after reading your post in full numerous times)...perhaps it's a Friday thing and my brain has gone off for the weekend before me.<g>

So the prepress houses files can't be found???

Is there film or plates? Without knowing how the final seps look, it will be hard to second guess things - although sometimes the h/tone dot proof or press sheet can give you clues when you check things under a glass.

From what you say, if you use a set-up similar to what they 'say' - results are not great...so perhaps this place makes minor use of the separation features and does all the work post sep in CMYK, radically departing from the separation source to arrive at the final 'great' final image?

Regards,

Stephen Marsh.

Date: Fri, 18 Oct 2002 08:29:13 -0700
From: Lee Varis>
Subject: Re: Re: Reverse Engineering Seps

Stephen Marsh wrote:

> From what you say, if you use a set-up similar to what they 'say' -
> results are not great...so perhaps this place makes minor use of the
> separation features and does all the work post sep in CMYK, radically
> departing from the separation source to arrive at the final 'great'
> final image?

I'll restate my problem, here's the chronology:

I'm hired, from outside the loop entirely, to correct some mediocre digital camera files so that they print better (these are not my images but I'm going from the RGB originals). The Ad agency has printed these before--they came out dark and muddy but they used them anyway. This time they are going to print them bigger and they want them to look better. I ask for specs from the prepress house. What I get back sounds like Photoshop defaults - not particularly helpful - the original job was probably separated using Photoshop defaults (though I can't track down anybody who knows and I don't have the original seps) I make new seps with a higher dot gain setting and a "snappier" black thinking that I should get lighter and more contrasty proofs back. Instead I get flat muddy images back on the Fuji proofs. The ad agency freaks, the prepress house says no problem we'll fix it, agency pays prepress house big money to fix files (they don't even call me). I finally get a look at the proofs and I'm suspicious... did they proof my files or did they muck up my files (inadvertently or otherwise) before they proofed. Now they can't seem to deliver the corrected proofed files !!!! Now I'm even more suspicious. I only have my original files, a set of bad Fuji proofs and a set of good Fuji proofs -- don't have the film that was used to finally print the "good" files. Is there any way I can tell from examining the "proofs" if they printed my original file or if they "doctored" it? I'm still hoping I can get a look at either the film or the final digital file but so far nobody seems to know what happened to them... hmm... does that make you suspicious?

I've seen this sort of thing before. A prepress house would prefer that the client continues to use film for photo jobs so they (prepress) can continue to make and charge for scans...good money on large catalog projects. If the client wants to shoot digitally the prepress house looks for some other method to recover lost profit so they make sure that none of the "digital" photos print properly and require lots of "color correcting"

regards,

Lee Varis
http://www.varis.com
888-964-0024

Date: Fri, 18 Oct 2002 20:58:00 -0700
From: "Steven Smith"
Subject: RE: Re: Reverse Engineering Seps

Hello

This is my first time to respond. and I am certainly no teck, but it seems to me the only clear way to tell would be to take your file to another Pre press house that you trust (is that an contradiction in terms?) and have it Proofed again. I know this has a price tag but without the corrected files everything is conjecture, you get into that calibration thing and that ICC thing and that Profile thing and then that finger pointing thing. You know they have already said its your problem and now you want to say it's their problem when we all know you can take the same file to 10 different Proofers and get 10 different proofs.

Date: Sat, 19 Oct 2002 08:11:00 EDT
From: Clay Tomas
Subject: Reverse Engineering Seps

Lee,

Hopefully, someone can help you by reading your files and proofs. Still, if inquiring minds want to know...One image to film and matchprint or a high end dependable proof from a known source should tell the story.

I had a similar experience once with a brand new client. I shot all files with a MegaVision T32 camera and supplied 5 sample seps to an out of town printer, through a print broker. I spoke with the printer prior to finishing 30 some full page images, same light and background. I made sure the backgrounds read the same numbers in a central area of the files.

The printer was going direct to plate and using Rainbow proofs. Unfortunately for my client, I never saw the proofs before they printed like mud. They worked on the files and reprinted at their cost. I later got one 8.5 X 11 matchprint pulled with four cropped sample images. Each image was consistent in color and contrast. My client wished his book printed that well. He paid for the matchprints.

My point, they may not be trying to screw you. They may be inept. If you can't tell from the proofs, get your own proofs.

Good luck,

Clay Tomas

901-761-4983

Date: Sat, 19 Oct 2002 11:23:23 -0400
From: Jerry L. P'Simer
Subject: Re: Re: Reverse Engineering Seps

Lee Varis wrote:

>>I'm hired, from outside the loop entirely, to correct some mediocre digital camera files so that they print better (these are not my images but I'm going from the RGB originals). The Ad agency has printed these before ? they came out dark and muddy but they used them anyway. This time they are going to print them bigger and they want them to look better. I ask for specs from the prepress house. What I get back sounds like Photoshop defaults - not particularly helpful - the original job was probably separated using Photoshop defaults (though I can't track down anybody who knows and I don't have the original seps) I make new seps with a higher dot gain setting and a "snappier" black thinking that I should get lighter and more contrasty proofs back. Instead I get flat muddy images back on the Fuji proofs. >>

It sounds to me like the specifications that you received from your client were specs that they came up with themselves or at least they were the specs used for the original separations. That does not mean that they were correct by any means for this particular printer. You do not mention whether or not the new specs came from the printer themselves or from your client. If the printer in question does not follow strict adherence to SWOP standards and use there own version of process control, (very common) and you do not have accurate information about the actual print conditions than, I would say, that you would have about a one in fifty chance in preparing an accurate separation for this printer. Even if the specs came from the printer it does not necessarily mean that they accurately describe their print capabilities. As Stephen suggested they may only be a starting point to enter into the CMYK color space and redefinition may occur after the fact. I personally do not have access to profile creation or editing so I am stuck using the built in methods for creating color separations and than edit the file to meet whatever specification are required. This does require very specific and accurate information to actually receive the expected results at the time of printing. It also helps to have an accurate method of proving the results before going to press.

>>The ad agency freaks, the prepress house says no problem we'll fix it, agency pays prepress house big money to fix files (they don't even call me). I finally get a look at the proofs and I'm suspicious... did they proof my files or did they muck up my files (inadvertently or otherwise) before they proofed. Now they can't seem to deliver the corrected proofed files!!!! Now I'm even more suspicious. >>

In my view I believe that the agency was correct in asking the pre press house to correct the files because they are familiar with their own processes and have a much better chance at producing the expected results than a third party retoucher unfamiliar with the present print conditions. I also do not find it suspect that they are unwilling to provide the final separations to you so you can see where things fell down. If the printer corrected the files for the agency than they have no obligation to you to provide these separation. You might try asking the agency if they can get their hands on them for you. Since they paid this printer to correct these files they will have a better chance of getting their hands on them then you would. I actually doubt that foul play was afoot but I wouldn't rule it out either. Too often I receive separations from outside sources that are garbage even for SWOP printing and have to spend a considerable amount of time correcting the images. (I'm not saying that this is the case with you so please don't take offense to this) In the past we have been accused of sabotage and were more than willing to provide the proper evidence to our client who did not prepare the original seps, but we would be unwilling to provide proof to the third party that produced the files in the first place. (sad but true)

>>I've seen this sort of thing before. A pre press house would prefer that the client continues to use film for photo jobs so they (prepress) can continue to make and charge for scans...good money on large catalog projects. If the client wants to shoot digitally the prepress house looks for some other method to recover lost profit so they make sure that none of the "digital" photos print properly and require lots of "color correcting" >>

From my perspective I do still prefer to scan from film but the world is changing and this does not occur very often anymore. I prepare several catalogs a year that are shot completely digital. To say that a printer or separator would deliberately sabotage files because they were shot digitally in my view is rather paranoid. Producing files for lithographic printing is a very unique and precise process and not all digital captures are equal nor are the skills of the photographers. Those who do not prepare files for lithographic print conditions consistently and modify their procedures accordingly will usually not receive the expected results every time. This takes patience experience and a good understanding of the print specs required for producing quality work every time out.

Regards,
Jerry P'Simer

Date: Sat, 19 Oct 2002 13:27:28 -0400
From: John Castronovo
Subject: Re: Re: Reverse Engineering Seps

Lee, I think you should get your own proof from one of your files, but use a different house. It's the only way to know if what you did is close to being correct. Then you might be able to prove yourself to your client, but you'll never prove yourself to the guy whose livelihood you almost stole away.

I believe that if the prepress house had an opportunity to "diminish" your work, they would, not because they're evil, but terrified of competition in a shrinking market. Most places certainly wouldn't offer up the least amount of help to your files even if cost them nothing. The attitude goes something like, "Hey, if the client is gonna' give us seps, then they get whatever comes out. We just press the button". Knowing this, the agency was correct to let them fix the files because it's the surest way to get around the blockade before the press.

Also, I've seen a few places so locked in to their own work flow, and not calibrated to standards, that they're incapable of instructing anyone from the outside on how to make separated files for their shop. For them, they *have* to adjust the files to get them to prove correctly. Your client might still want to, or have to, use them for reasons we can't imagine.

Sad, but true. Get you own proofs next time before submitting your work.

john castronovo

Date: Sat, 19 Oct 2002 13:27:32 -0700
From: Peter Figen
Subject: Re: Re: Reverse Engineering Seps

Lee,

Why don't you make or have made a custom profile for that proofing system? With proofing systems, especially those for DTP, adhering to no known standard, it's extremely difficult to prepare files with the techniques we used to use everyday. This is what I've had to do with at least half a dozen commercial printers in and around Los Angeles, all of whom claim to be calibrated to SWOP or something like, but none of whom really are.

Peter Figen

Date: Sun, 20 Oct 2002 21:26:59 -0400
From: Jim Rich
Subject: Re: Reverse Engineering Seps

Lee,

The next time you send files out for proofs, why don't you slip in a known image that you have used previously. Then you will have a better idea of who is screwing things up or if they are inept.

Jim Rich

,

Date: Mon, 21 Oct 2002 08:54:53 -0400
From: Dan Remaley
Subject: RE: Reverse Engineering Seps/Proofing

This proofing issue is the number one problem in the industry, with digital proofing we have lost our 'stake' in the ground, Matchprint, Waterproof etc. GATF make a great product for this purpose, it's called a GATF Proof Comparator. This target allows us to make a visual reference as well as a colormetric one. Include it with your images and 'measure' your proofs.
Dan Remaley/GATF
412.741.6860

Date: Mon, 21 Oct 2002 13:25:21 -0000
From: Stephen Marsh
Subject: Re: Reverse Engineering Seps/Proofing

> This proofing issue is the number one problem in the industry, with digital
> proofing we have lost our 'stake' in the ground, Matchprint, Waterproof etc.
> GATF make a great product for this purpose, it's called a GATF Proof
> Comparator. This target allows us to make a visual reference as well as a
> colormetric one. Include it with your images and 'measure' your proofs.
> Dan Remaley/GATF

I agree Dan.

Having a known colour bar or image as has been previously suggested seems like a great idea to me. I have seen inkjet proofs which have had the common 'magenta flood' which does not show in the nozzle check or head alignments...unless the person running the proofer really knows the image, they may think that this is how things should be and is not clued into the fact that there are problems until the peson who knows the image says something...unless they first chase their tail correcting a problem that is not in the file and wondering why the next print is off too.<g>

As many wide format inkjets use a scatter type system (in publishing anyway) - having a colour bar/image content in a tall strip would work well, so that the wasted 'gutter' space between pages is used and you can have full page depth for the patches etc. With luck the person running the proofer will get to know these memory colours (and a physical sample that is known to be good can always be used for visual or machine comparrison) and will spot an issue on the colour bar more so than the 'unknown image'.

Although this seems obvious -- there seems to be little talk of such things which seem very critical for the process control side of digital proofing (it seems that some treat inkjet proofing as a push button solution, when they did not have the same attitude to traditional proofing).

Stephen Marsh.

Date: Mon, 21 Oct 2002 17:43:00 -0400
From: John Rawlins
Subject: Re: Reverse Engineering Seps

As a commercial printer, let me comment on some of the posts to this subject. Did anyone consider that the original files just might be junk and it is neither the service bureaus' or the printers fault at all?

Let me say that we as the printer, do not alter anyone's files just because we really wanted to do the seps ourselves. Nor would any pre-press house, that I know of. If proofs look bad, 999 times out of 1000, its purely a case of "junk" files of scans and digital captures, produced and provided by people that think they know what they are doing... but don't. Now if you are taking your files to an insta-copy shop, or a printer that has very limited process color experience, or limited proofing capability, maybe something else could be going on there... but otherwise no.

There is an "illusion" that has manifested itself recently, that if a printer, or pre-press house supplies a "profile" for the provider to use, that it will mysteriously "fix" any scan or digital image and make it beautiful. Wrong. It has to be right to start with, so the printer can reproduce it. Profiles are to aid in "tuning" color, not making it from scratch. Kind of like trying to tune up a car that does not run... you just can't.

There is also another mystique, that each and every commercial printer out there is so different in how they handle and print a file, and if you don't "profile it" for them, it just won't print correctly. Wrong here too. Again, lets throw out the insta-printers and the one and two color printers just trying to print process color. Assuming we are dealing with a reputable printer that has been around for a while, he is probably Ripping the files with either Scitex Brisque, Rampage, Harlequin, or Agfa RIP. He is outputting either Kodak, Fuji or Agfa plates, and he is printing the job on a Heidelberg, Komori, or Mitsubishi press. All inks are basically the same... There is only one manufacturer of the base flush (pigment). They all buy it from the same place and blend them with varnishes to make ink. All blacks are the same. All Cyans, for all practical purposes are the same. Yellows will either be chrome (red), or clean (to the green side). Magentas will either be Rhodamine or Rubine cast. A good scan, or digital capture will print beautifully, using any combination of the available inks. Now if we were matching specific wood tones, or fabrics, the scans or seps. would in fact have to be done for the specific inks. Changing inks, or correcting for the wrong inks, would be evident by slight tone or shade shifts that would not be acceptable for this type of work. But, for everyday commercial work, the range is far wider than most imagine. Dot gain... most printers 12 to 20%, and with that, depending on the image, the lay person may not even see a discernable difference between 12 or 20% gain. If you are dealing with a printer showing gains much higher than that (on coated stocks) you really need to be looking for another printer, not trying to compensate for it.

We get in from clients some very nice quality work that we have no problems with whatsoever. We also get our share of junk. And when those clients see their proofs... its the old "what did you do to my file?" "It didn't look like that on my monitor" "Why doesn't it look like my laser?"

Education here is the key... buy most don't want to hear it. I guess I went beyond my .02 cents worth.

John Rawlins

Date: Mon, 21 Oct 2002 10:34:47 -0500
From: Al MacDonald
Subject: Re: RE: Reverse Engineering Seps/Proofing

Dan,

Can you provide some specifics to this inclusion in a workflow?
--
Al MacDonald
Shaughnessy MacDonald, Inc.
Elk Grove Village, IL 60007
http//www.shaumac.com

------

> From: "Remaley, Dan">
> This proofing issue is the number one problem in the industry, with digital
> proofing we have lost our 'stake' in the ground, Matchprint, Waterproof etc.
> GATF make a great product for this purpose, it's called a GATF Proof
> Comparator. This target allows us to make a visual reference as well as a
> colormetric one. Include it with your images and 'measure' your proofs.

Date: Mon, 21 Oct 2002 15:35:53 -0700
From: Peter Figen
Subject: Re: Re: Reverse Engineering Seps

John wrote:

>>There is also another mystique, that each and every commercial printer out there is so different in how they handle and print a file, and if you don't "profile it" for them, it just won't print correctly. Wrong here too. Again, lets throw out the insta-printers and the one and two color printers just trying to print process color. Assuming we are dealing with a reputable printer that has been around for a while, he is probably Ripping the files with either Scitex Brisque, Rampage, Harlequin, or Agfa RIP. He is outputting either Kodak, Fuji or Agfa plates, and he is printing the job on a Heidelberg, Komori, or Mitsubishi press. All inks are basically the same... There is only one manufacturer of the base flush (pigment). They all buy it from the same place and blend them with varnishes to make ink. All blacks are the same. All Cyans, for all practical purposes are the same. Yellows will either be chrome (red), or clean (to the green side). Magentas will either be Rhodamine or Rubine cast. A good scan, or digital capture will print beautifully, using any combination of the available inks. Now if we were matching specific wood tones, or fabrics, the scans or seps. would in fact have to be done for the specific inks. Changing inks, or correcting for the wrong inks, would be evident by slight tone or shade shifts that would not be acceptable for this type of work. But, for everyday commercial work, the range is far wider than most imagine. Dot gain... most printers 12 to 20%, and with that, depending on the image, the lay person may not even see a discernable difference between 12 or 20% gain. If you are dealing with a printer showing gains much higher than that (on coated stocks) you really need to be looking for another printer, not trying to compensate for it.>>

Most of the many variables you list here are EXACTLY why custom profiles are necessary. Sure, if you get a film and proof from most anywhere, you CAN match it, but with DTP, it's the house digital proofers that you have to aim for, as that's what they're going to try and match, and it's those digital proofers that are all over the map. Some clients might not see the difference and some might not care, and still others don't mind three rounds of proofing, but none of that is acceptable for any of my clients.

Peter Figen

Date: Tue, 22 Oct 2002 01:08:05 +0100
From: Richard Kenward
Subject: Re: Re: Reverse Engineering Seps

John writes

>We get in from clients some very nice quality work that we have no
>problems with whatsoever. We also get our share of junk. And when
>those clients see their proofs... its the old "what did you do to my file?"
>"It didn't look like that on my monitor" "Why doesn't it look like my
>laser?"

With respect to John and several other contributors to this thread, I feel that I really must put in a word here for Lee.

To my knowledge Lee is a very competent operator who has been producing digital files for print for a long time. I for one would be most surprised if he would be producing 'bum' work, even though I understand that he was sorting out someone else's poor quality digital files.

I suspect Lee is far too modest to say this himself.

Cheers

Richard Kenward

Date: Mon, 21 Oct 2002 21:22:08 EDT
From: Clay Tomas
Subject: Re:Reverse Engineering Seps

To John,

You sound like you have had some frustrating experiences from the print side, and I am sure they are justified. Peter, brings up the question of how good are the proofs, and what does the printer match?

I began shooting digitally in 1992 at a highend prepress and print shop. For seven years I converted and color corrected files prior to sending them to the computer graphics department. I have been doing the same on my own for three years and only one print job was bad. I sent the printer 5 files prior to converting the other 30 or so and really received no feedback. One file was sent to a magazine and printed fine. Matchprints were fine. I later saw rainbow proofs, under my 5000 K lightsource, and they were all over the place, ugly green. My files were consistent in common areas.

Where did it all go wrong? My files??

At my old employer we ran film. If any test patches were 2% off we notified someone and it was looked into and adjusted. If I had a product that was 70C, 50M,and 40Y on my file, I expect that good film output will read close. Not noticeable to many people. Is it unreasonable to expect that same file to produce the same densities if burning a plate for DTP? I realize there may be many factors I am unaware of and I do not claim to be a prepress or print specialist. I always ask my clients to get test proofs prior to my conversions, most don't want to spend the money. Except for one time, I have never had a problem. Yes, I on occasion say I wish that one had a little more contrast, this one was a little lighter and so on, but all in all I can't complain. From past posts here and on another user group I get the impression Lee Varis has delivered many files to many satisfied customers, so I for one gave him the benefit of the doubt.

Printers going direct to plate would have to have a tighter system it seems to me and a proofing system which is color correct. Otherwise, what are we correcting and why?

Anyones thoughts?

Clay Tomas
Clay Tomas Photography

Date: Mon, 21 Oct 2002 18:48:48 -0700
From: Peter Figen
Subject: Re: Re:Reverse Engineering Seps

> Printers going direct to plate would have to have a tighter system it seems
> to me and a proofing system which is color correct. Otherwise, what are we
> correcting and why?

Clay,

It seems to me that it's more as the printers move to DTP, with its digital proofing, that they are not maintaining any sort of correlation to their old proofing systems. I spoke to another one in L.A. today and they acknowledged that but could offer no reason. I can only surmise that it's either the presses run better at their new calibration, or they know that they are going to be getting legacy files that won't proof well the first time, requiring more rounds of proofing, but I'm not quite that cynical....yet.

Actually, the impression I get is that they just don't care. They tell me that their scanners have been calibrated to their proofers, and that's all that matters, the rest of their customers supplying their own files be damned. This phenomenon does not seem to be occurring at places that send out lots of ads to magazines, which by and large HAVE gone DTP, but with no huge difference in how legacy files print.

Peter Figen

Date: Mon, 21 Oct 2002 22:29:11 -0400
From: "john castronovo"
Subject: Re: Re:Reverse Engineering Seps

From: Peter Figen

> Actually, the impression I get is that they just don't care. They tell me that
> their scanners have been calibrated to their proofers, and that's all that
> matters, the rest of their customers supplying their own files be damned.

That's it in a nutshell for most of the printers I've ever dealt with.

john c.

Date: Tue, 22 Oct 2002 00:29:10 EDT
From: Dan Margulis
Subject: Re: Reverse Engineering Seps

After a couple of weeks in places where e-mail doesn't exist, I come back to one of the more interesting threads.

Lee raises the age-old question of whether a prepress house is sabotaging his work. Probably not. The consequences of getting caught are too high. In any event if they're competent enough to sabotage, then they should have been competent enough in the past so that their work wouldn't have been so muddy that you were called in the first place.

Jim Rich has it nailed: if you're sending CMYK files for proofing to a new prepress house, printer, or whatever, it's an important insurance policy to also send a file or two where you're familiar with the expected result. Even thumbnail size images can indicate whether there's something bogus going on.

But, too late for that now. A couple of possibilities, caused by incomplete information: we all seem to be assuming "SWOP" as the printing condition but Lee hasn't stated this for a fact. A prepress house may have an arrangement with a particular supplier to use nonstandard proofing methods.

More important is the question of whether this "Fuji Proof" is a Fuji Waterproof, an analog process requiring CMYK film, or a Fuji Pictrography proof, a digital proof that requires, among other indignities, that the file be converted internally to RGB in order to reconvert it to CMY. That leaves a lot of room for strangeness, particularly if, as you appear to be doing, a nonstandard black is in use. The Pictros that I've seen by default assume a rather cheesy black which may account for the flatness you perceive.

Also: what color is the proofing stock? If it's significantly off-white, that really contributes to a flat look, if you're used to a more brilliant substrate.

And: did you embed a profile? If so, and if its a Pictro, there have been reported cases where this caused a problem.

Since apparently there will be some loss of sleep if this isn't resolved, and one would prefer not to go to the expense of proofing the job elsewhere, the indicated procedure is to try to reverse engineer not one of the supplier's seps but rather their proofs of your own sep.

That is, open your CMYK files and (if necessary) go to Mode: Assign Profile>Don't Color Manage. Now open Edit: Color Settings>Working Spaces>CMYK>Custom CMYK and enter the dot gain curves dialog. Do whatever it takes to make the images on your monitor look like the flat, muddy proof the prepress house provided. Then, using those same settings, open some CMYK images you've printed elsewhere and see whether they look radically different.

My guess is that with the big black you've constructed you may have fallen victim to underestimating black dot gain. 5% (or even more) extra dot gain in the black isn't uncommon. So, as a first effort, try 25% dot gain generally and 35% in the black. If that still doesn't look muddy enough then something is seriously hosed with the proof and you should consider proofing it elsewhere. Similarly, if no matter what you do with the dot gain curves you can't get a preview to more or less match the proof, something funky is going on.

Dan Margulis

Date: Tue, 22 Oct 2002 00:13:09 -0700
From: Lee Varis
Subject: Re: Reverse Engineering Seps

Hi all,

On Monday, October 21, 2002, at 09:29 PM, Dan Margulis wrote:

> Jim Rich has it nailed: if you're sending CMYK files for proofing to a new
> prepress house, printer, or whatever, it's an important insurance policy to
> also send a file or two where you're familiar with the expected result.

Well... The agency still doesn't have the files back but I think I'm a victim of miscommunication and ignorance on the part of the agency not the printer! I did something like this before for this agency with some other similar files. These printed flawlessly and I was the big hero! This time around I was working with files that seemed flat, unsaturated and dark - I made what I thought was a reasonable adjustment - BUT... I just found out that a different printer printed these new files and a different prepress house prepared the film.... ARRRRGGGGH...!!! I was given specs for the wrong prepress & printer!!

I was compensating for a press condition that the 2nd printer didn't have... their dot gain wasn't the same... They were printing Sheet fed with a fairly high black limit. My files were weak in CMY but the big black (meant to compensate for a weak black on the other press) actually dulled the color more than snapped contrast. I had made the black plate too contrasty and with the weak CMY the shadows looked noisy - The printer and agency just assumed that it was due to bad digital camera files (they were bad but they could have been prepared to print better If I had known the true specs)

> My guess is that with the big black you've constructed you may have fallen
> victim to underestimating black dot gain. 5% (or even more) extra dot gain in
> the black isn't uncommon. So, as a first effort, try 25% dot gain generally
> and 35% in the black.

I had separated with a light GCR and a lower black limit, then boosted the black point in the black plate to get a higher number to hit the expected ink limit. I think the new printer could take the black but really needed more CMY than I delivered. I got pretty reasonable (but somewhat light) neutrals and highlight detail was there, but color seemed lacking in the proofs which made everything look just a bit flat.

I think the moral here is to NOT assume that the production manager of the agency understands that there is a difference between a web press and a sheet fed press or that one printer may print to different specs than another. My contact at the agency was the creative director - he doesn't really know.. but I expected his production manager to get me the info from the print/prepress that was handling the job! They actually changed their mind after I had already prepared the files... "After all, CMYK is CMYK, isn't it"

I was thinking that the prepress house had some goofy automatic transform that they always applied to digital camera files separated through Photoshop and I had supplied files that didn't need the transform -- instead, the client simply switched to a printer running a different press with different dot gain and ink limits!

The funny thing is that the first proofs weren't horrible - just a bit flat and weak AND the shadows were not as strong as I expected given that I thought I was bumping the black (within what I thought the ink limits were)

Oh well... now I'm sure this prepress house thinks I can't prepare CMYK properly... they actually print closer to SWOP TR001 !

Date: Tue, 22 Oct 2002 07:41:31 -0400
From: John Rawlins
Subject: Reverse Engineering

Clay Tomas Wrote:
>You sound like you have had some frustrating experiences from the print side,
>and I am sure they are justified. Peter, brings up the question of how good
>are the proofs, and what does the printer match?

Peter Figen Wrote:
>>Most of the many variables you list here are EXACTLY why custom profiles are necessary. Sure, if you get a film and proof from most anywhere, you CAN match it, but with DTP, it's the house digital proofers that you have to aim for, as that's what they're going to try and match, and it's those digital proofers that are all over the map. Some clients might not see the difference and some might not care, and still others don't mind three rounds of proofing, but none of that is acceptable for any of my clients.>>

Major misconception here guys... A pressman's #1 job is NOT to match the proof. His job is is to put the plates on the press, put it in register, level the fountains, and bring the ink levels up to standard densities. Now at this point, we, and/ or the client go to a view booth and compare the sheet to the proofs and/ or originals. The better the proof, and the more controls used in the making of the proof, the closer it will be to the press sheets. At this point we can juggle densities to tweak the fine points, but for all practical purposes, if we are printing good, and we are at standards, we are reproducing the files as provided to us. In our case, our digital proofs are very close to what we are printing. Now we could go down the street to competitor "A" and have him print the job... it would look almost the same. Same with vendor "B", and same with vendor "C". So we have nailed down the fact that THIS IS what the files really are and how they reproduce. Having said that, have the same 3 competitors make a digital proof of the files and all 3 of theirs and ours would not match. We have invested a considerable amount of time, money, and energy getting our proofs as close as possible, so there are no "surprises" at press time. So my question... If the file could print at 4 different vendors and the end results look the same, is it really correct to advocate a client to supply different files to each to compensate for the output to match the proof or the proof for the output? If my proof device is profiled to match my press, that is all I need. If the proof I make of your file looks like crap, its because it is... Its not because my proof device made it that way or I don't know how to print.

The digital revolution created as many new problems as it solved. We have been digital for over 12 years, and converted to total digital workflow over 3 years ago. We no longer even have a frame to burn film plates, nor do we accept film, or supply film proofs. Accepting digital proofing (especially inkjet) is a very hard concept for many clients to live with. We were told we could never do what we were doing and that some clients would never buy it. But they do. Cost is everything now. To us and to the clients. We quote all jobs with inkjets only. Kodak Approvals or Digital Matchprints are shown as an option line item. Few spend the money for them, but enough do that we still need to hang on to them... for now. This business is constantly evolving and we try to keep up with it. Processes change, technology changes. Inkjet proofing will get better, but for now we all work with what we have.

John Rawlins

Date: Tue, 22 Oct 2002 07:52:23 -0500
From:Henry Segalini
Subject: Re: Reverse Engineering

Different strokes for different folks. If we don't match the proof we made, we eat the job. Our press people's responsibility is to match the proof.

Our Customers don't care about standard densities. They just want the job to match the proof.

Date: Tue, 22 Oct 2002 09:02:02 -0400
From: Dan Remaley
Subject: RE: RE: Reverse Engineering Seps/Proofing

The GATF Proof comparator is an eps file that can be output to any proofer, or run to film for film based proofing. As a workflow, include it with your other images, as a sample for verification of the proofing system. Whatever you choose to 'read' Lab, dot gain, gray balance , should be consistent from proof to proof, job to job. The color proof is probably the most important step in the color reproduction process, the customer 'signs' off on it, the scanner or photographer bases his color correction on it, the press operators follow it for a reference. From my 30 years experience, who makes these proofs?? The lowest paid - newest hired person! How's that make you feel! Without something to measure - not very good! On November 4 I'll be presenting my Process Control Workshop here in Pittsburgh, I wish you all could attend. Please call with any questions/comments.

Dan Remaley/GATF
412.741.6860x450

Date: Tue, 22 Oct 2002 09:13:05 -0400
From: Dan Remaley
Subject: RE: Reverse Engineering Seps

The reality is that most printers web and sheetfed print close to SWOP numbers. The reason is that the 20-22% midtone gain for web is based on 133 line screen, so when we print sheetfed at 150-175 line screen, the 'sharper printing' sheetfed press work is printed at a higher screen ruling causing higher (SWOP) dot gain! The only 'problem' is with CTP - where the plates are linear and the midtone gain is near 15% (or wherever they want it to be - by moving curves). This creates a problem, a designer looks at a Pantone book for his screen builds, the Pantone book is printed with 22% gain - how will you match that with a press printing 15% gain????- You won't! So the CTP-linear people 'build' their own Pantone book to match their 'new' print condition. Color management doesn't work without excellent process control.
Dan Remaley/GATF
412.741.6860x450

From: Lee Varis

> Oh well... now I'm sure this prepress house thinks I can't prepare CMYK
> properly... they actually print closer to SWOP TR001 !

Date: Tue, 22 Oct 2002 09:35:07 -0400
From: Scott Olswold
Subject: RE: Reverse Engineering

Taking this way off course, I agree with Henry Segalini. When I worked at a printshop, it was the floor's job to match the proof. Sometimes we were lucky and it was our proof we were matching. Sometimes we were lucky and it was a Matchprint or similar analog proof. Towards the end of my sojourn there, the reps started delivering off-site inkjet proofs. Getting those to match was sometimes fun (from what I hear), but not impossible. From an overly simplified perspective, what John advocates as a pressman's responsibility is very much less an art and more just a factory worker (remember that whole time-motion study stuff from the Gilbreth's?). Ick.

Scott Olswold

Date: Tue, 22 Oct 2002 09:51:33 -0400
From: Jim Ray
Subject: RE: Reverse Engineering

Whose job is it to deliver what the customer wants (expects)? When I sign off on the proof, that is what I expect to get (within reasonable limits).

Jim Ray

Date: Thu, 24 Oct 2002 18:11:25 EDT
From: Dan Margulis
Subject: Re: Reverse Engineering Seps

Lee writes,

>>I did something like this before for this agency with some other similar files. These printed flawlessly and I was the big hero! This time around I was working with files that seemed flat, unsaturated and dark - I made what I thought was a reasonable adjustment - BUT... I just found out that a different printer printed these new files and a different prepress house prepared the film.... ARRRRGGGGH...!!! I was given specs for the wrong prepress & printer!!>>

Welcome to life in the real world!

>>I was compensating for a press condition that the 2nd printer didn't have... their dot gain wasn't the same... They were printing Sheet fed with a fairly high black limit. My files were weak in CMY but the big black (meant to compensate for a weak black on the other press) actually dulled the color more than snapped contrast. I had made the black plate too contrasty and with the weak CMY the shadows looked noisy.>>

That's a good analysis of what did happen, and an excellent illustration of the limitations of what was (in effect) heavier GCR. What happened is, you put the bomb in the pressman's hands and now are upset because it happened to go off.

If you're determined to get a heavier black in the shadow at all costs and aren't totally confident of the printing conditions, you want to affect only the tail end of the black curve, not the middle. And you want to reduce the CMY component by means of Selective Color>Blacks, not by curves.

>>I think the moral here is to NOT assume that the production manager of the agency understands that there is a difference between a web press and a sheet fed press or that one printer may print to different specs than another. >>

No, assuming that is the once upon a time, not the moral. You have to accept that agencies and printers often screw up and when they do you will get the blame, unless you take precautionary measures. Anything else is a fairy tale.

There are two possibilities of dealing with it: 1) complain about color management and how printers should give out profiles; 2) use common sense.

The moral is: use a correction style that will make the image look better under virtually all printing conditions. That's what by-the-numbers does. Note that there is a big difference between "better" and "best possible". You can't get best possible without being sure of the printing conditions.

What happened in your case is that you were shooting for best possible, but your correction style left open the possibility of a very bad result indeed under certain circumstances. If you had corrected in a more standard way you would still have been unpleasantly surprised by the printed results, which would have been slightly too light. But slightly too light is something that clients live with all the time. Drab, muddy, and colorless, they don't like.

Dan Margulis

Date: Thu, 24 Oct 2002 16:51:15 -0700
From: Lee Varis
Subject: Re: Re: Reverse Engineering Seps

Dan wrote:

> ... You have to accept
> that agencies and printers often screw up and when they do you will get the
> blame, unless you take precautionary measures. Anything else is a fairy tale.

Yes... I got sloppy!

> There are two possibilities of dealing with it: 1) complain about color
> management and how printers should give out profiles; 2) use common
> sense.

LOL.. hits it right on the head...

> What happened in your case is that you were shooting for best possible, but
> your correction style left open the possibility of a very bad result indeed
> under certain circumstances. If you had corrected in a more standard way you
> would still have been unpleasantly surprised by the printed results, which
> would have been slightly too light. But slightly too light is something that
> clients live with all the time. Drab, muddy, and colorless, they don't like.

Yes, although I did compensate for the increased black with a Selective Color>Blacks move it still was too radical a correction because I separated to a lower ink limit and played with curves too much... it messed up the GCR. It might have worked... BUT, in the end the gamble didn't pay off.

Hopefully everyone on the list can learn from my mistake... like Icarus flying too close to the sun... I got burned by my own ambition!

regards,

Lee Varis
http://www.varis.com
888-964-0024

Date: Mon, 28 Oct 2002 10:33:44 -0500
From: Dan Remaley
Subject: RE: Re: Reverse Engineering Seps

Hi Lee, this is exactly why SWOP was established! (Specifications for Web Offset Printing) by the way, the numbers 'work' for sheetfed as well. The AD Agencies got tired of seeing their important color corrections turn to junk on press. Without a specification, a target, GRACoL, something, we can only "hope" they print correctly. The large color trade houses always scanned to SWOP, gray balance, dot gain, density, TAC etc. This way every job doesn't become an 'original'.
Dan Remaley/GATF
412.741.6860x450

> I did something like this before for this agency with some
> other similar files. These printed flawlessly and I was the big hero!
> This time around I was working with files that seemed flat, unsaturated
> and dark - I made what I thought was a reasonable adjustment - BUT... I
> just found out that a different printer printed these new files and a
> different prepress house prepared the film.... ARRRRGGGGH...!!! I was
> given specs for the wrong prepress & printer!!

Date: Wed, 30 Oct 2002 09:47:26 -0500
From: John Rawlins
Subject: Reverse Engineering Seps

Dan Remaley wrote:
>Hi Lee, this is exactly why SWOP was established! (Specifications for Web
>Offset Printing) by the way, the numbers 'work' for sheetfed as well. The AD
>Agencies got tired of seeing their important color corrections turn to junk
>on press. Without a specification, a target, GRACoL, something, we can only
>"hope" they print correctly. The large color trade houses always scanned to
>SWOP, gray balance, dot gain, density,

In actuality, SWOP was established by the printers to tell the agencies what to do, not the agencies to tell the printers what to do. When It was first published in 1974, few people even knew what a scanner was, let alone had one. The SWOP standards where purely related to standards for supplied advertising material (film) and their related proofs (mostly press proofs and progressives). The larger web (primarily publication) printers of the day were having a tough time matching the supplied film to the supplied proofs once they got on their presses. They could pull up a 16 page signature that may have had 12 full page ads on it... each one supplied film from 12 different agencies/ and or film houses. Some matched, some didn't. Some were close, some miles away. Jockeying ink densities was really not an option since forcibly matching one ad could destroy 2 others. By all the printers agreeing on a standard ink set, and standard density range to run in, the established guidelines film houses to proof by. Basically saying, if you do yours like this, and we do ours the same, we will all be a lot closer.

In no way did it tell the printer how to print, or the film house how to make film. All it did was establish the standard for supplied advertising material to publications. It had nothing to do with commercial printing. If a commercial printer elected to use SWOP standard inks, that was fine, but he could also run whatever he had to. And, as in any business or manufacturing process, some use the good stuff, others the cheap stuff. Some do a good job, and get paid well for it, some do a poor job, that you go to because you want to save money.

SWOP has evolved over the years and has been revised 6 or 8 times, each time encompassing more of the current workflow standards established. Most commercial printers have in fact come around to the SWOP hue & density standards.

John Rawlins

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