'Safest' way to convert PMS - CMYK

[jethrodesign]

Hi, I'm wondering what the 'safest' or most reliable method is for converting PMS spot colors to CMYK when we might not know the end printer?

I was reading up on the difference in the conversion in InDesign (and Illustrator) when checking the "Use Standard Lab Values for Spots" box in the ink manager. So I understand how they come to different conclusions (sometimes pretty radically different) but don't really understand which is 'better'.

I know this is a bit of a subjective question, but would love to hear from folks who have experience here.

Most of the time, we are not aware of the printer who will end up printing a certain piece. And in the cases when we do actually know, I've found that the majority of the printers I've spoken to don't have a custom profile for their press (or don't know what that is) and usually tell me they calibrate their presses to SWOP standards or something.

For the current project, it's going to one of those large gang-printing shops (4by6). They state that they don't recommend using ICC profiles and that using a standard SWOP2 profile would give an approximation of their presses.

- So if all we know is 'use SWOP', and I have my document set to use US WebCoated SWOPv2, in general would using the LAB conversion give me a safer match or trusting the Pantone conversion?

(The numbers are very different, for instance our dark blue PMS 539C is either 100C, 49M, 0Y, 70k using the Pantone conversion, or 100C, 77M, 47Y, 51K using LAB conversion. Those are pretty radically different numbers and kind of make me nervous).

THANKS!!!

[Lou Dina]

Jethro,

In general, I never use spot colors in a design unless my ultimate intent is to actually print with premixed spot inks. I recently completed a job where my customer's logo is specified as a bright blue (Pantone 300C), but they told me they did NOT want to print the brochure as a 5/C job (ie, CMYK + PMS 300C).Pantone 300C lies outside of the normal CMYK gamut. I told them this and they just told me to get as close a match as I could with CMYK.

I used Pantone Color Bridge PC (Process coated) to select the best CMYK equivalent to 300C based on their current recommendation. They recommended 100C/42M/0Y/0K. On my calibrated monitor it looked like a reasonable match, so that's what I went with. I also have two swatch books and looked at CMYK equivalents to 300C in both, just as a check. (By the way, the Pantone solid to process guide I have suggests 100C/44M, but I chose the color bridge numbers, since I'd rather err on the side of less magenta). I set up InDesign with color management turned on, and set the CMYK profile to SWOP2006_Coated3v2 (IDEAlliance profile, which conforms to the latest G7 criteria).

As in your case, the final printer had not been selected, and was not selected until after the job was completed. So, I converted the job to PDF using PDF/X1-a, but specified SWOP2006_Coated3v2 (Grade #3 paper, with a 300 TIL). I felt this was fairly safe. The job printed fine, but that's a crap shoot when you don't know who the printer will be. I was cut out of the loop after the job was submitted to my customer.

To cover my butt, I sent high quality printed proofs on a nice, bright, coated paper from my calibrated inkjet (custom profiles), which I set up to simulate the above SWOP profile. I also put in writing that the printer's final job should come very close to my proofs if they did everything right and met industry standards. I also told them my PDF had stripped out all profiles, converted everything to the SWOP profile (preserving CMYK numbers, so I didn't convert black plate data into CMYK), etc. I also supplied the original InDesign file with tagged files, fonts, images, etc. I didn't want to get hung if the job ended up looking like crap. As it turns out, the customer selected a decent printer, provided my proofs as a guide, and the job printed fine (this time).

I sometimes wonder if Pantone looks at their sales figures, decided they need a boost, and releases new swatch books with new numbers just to pump up sales! Your color (539C) is a deep, dark blue. Here are the some of the numbers I came up with:

1.  100/77/47/51 gives 15L, -7a, -19b in the PS color picker as the Lab equivalent.

2.  100/49/0/70 gives 15L, -9a, -32b (same density, much bluer)

3.  Pantone Color Bridge PC gives 100/58/23/77, with a Lab equivalent of 8L, -11a, -24b, darker tone, but color in between the two above.

4.  Pantone solid coated guide gives a Lab equivalent of 15L, -9a, -21b.

The good news is that this color is so darn dark, that a bit of color shift one way or the other may not be all that critical. Pick your poison.

Good luck.

Lou

[jethrodesign]

Thanks for the detailed reply. Yeah, we usually don't deal with spot colors, but on occasion we have logos that are specified as spot color (as they may be used in spot or process jobs) so we have to make it work.

It's discouraging the discrepancy in color numbers achieved, and as you noticed, the Pantone conversion number appears much more blue. But this color is dark so it may not show as much.

The other color in this specific job, though, is PMS 631C which is a much lighter cyan color.

The LAB conversion gives me: 68C, 7M, 16Y, 0K.

The Pantone conversion is: 67C, 0M, 12Y, 2K

And the Bridge conversion is: 73C, 0M, 11Y, 0K

So it looks like the Pantone and Bridge conversions might be a bit more 'pure', but that 7% Magenta can make a pretty big difference in how the color swings on press.

It almost feels like just picking numbers out of a hat with this!

With this job, I'm trying the LAB conversion numbers as this printer is ganging the job anyways, so we've already warned the client not to expect much as far as color control. But they're cheap, so that's what the client wanted.

I'm much more concerned about future jobs where the client is much more demanding. And where it really comes into play is when we design a logo for a company and we need to send them master files in different color spaces (PMS, CMYK & RGB) so they can use them for different purposes. Which conversion process is going to be safest????

THANKS AGAIN!

[Lou Dina]

I feel your pain.

A few thoughts...anyone using a gang printer has to expect to get whatever comes off the other end. All the jobs are run together and there is no tweaking or control. Gang printers, in particular, state that their aim is "pleasing color" and they definitely do not guarantee color match. Gang printers are all about price, so be sure to inform your customer. I think there is a sign that says, "Quality, Price, Speed — Pick Any Two".

Another thought is to put all your concerns and criteria in writing when submitting a job, with a strong disclaimer. I stack the deck in my favor by submitting accurate, hard copy color proofs to a good SWOP standard, and telling the customer that the final prints should match my proofs if the printer does his job right.

Another thing is that my standards are far higher than many customers, so even when I am not happy, they often are, as long as Aunt Matilida doesn't come off the press green. (Then again, sometimes they prefer that Aunt Matilda come off the press green!)

Unfortunately, quality requires time, care, expertise, attention to detail, good equipment and good coordination and communication every step of the way. If that's not in the cards, then the customer has to live with the consequences. Accurate color proofs help, because I can get the customer to sign off on the colors. Then, if they choose the printer, cutting me out of the loop, I wash my hands of the result right then and there. Of course, I still want the job to come out great, but I won't be held accountable if they choose to go down their own road. I warned them up front. I just try to take a defensive posture when put in this situation.

One other thought...I find blues are generally more acceptable to me, and my customers, if they err on the side of less magenta rather than more. So, given a choice, I'll tend toward using less magenta to guard against incorrect gray balance on press.

Lou

[Printer_Rick]

jethrodesign wrote:

(The numbers are very different, for instance our dark blue PMS 539C is either 100C, 49M, 0Y, 70k using the Pantone conversion, or 100C, 77M, 47Y, 51K using LAB conversion. Those are pretty radically different numbers and kind of make me nervous).

The first build is Solid to Process.

One good feature about Solid to Process, you will notice that almost all of the builds are comprised of only 2 or 3 primaries. Hardly ever 4.

The colors are fully saturated, then darkened with black when necessary. (Black also desaturates because CMYK is not a cube like Lab)

The Lab conversion technique introduces the yellow gray component to the 539 color. In theory no problem with that, and if the press does its job it should be OK. That is the color management solution.

But with this Lab conversion you have four ink variables for the critical color swatch instead of only 3. In any production system, an additional variable is more chance for error.

I would use both color management and a Pantone build library. Create a blank CMYK in PS and assign the appropriate profile. For example Lou's suggestion of SWOP2006 grade 3.

Go to the color picker. Pick the solid 539 and hit enter.

Go back to the picker again. Change to the Solid to Process library. This identifies the closest STP match to the solid Lab in your current CMYK color space. It happens to be 5395 (FWIW) using relative colorimetric.

Going to process coated you end up with DS 327-1. This is 35C 0M 0Y 100K. So not a great result using that library.

There are other solutions you could try too. Some people use Saturation intent for solid colors, which may yield a more pleasing result. Others create a specialized max GCR version of the CMYK space and convert the Lab to that, ensuring that black is used as the gray component.

Bridge may be OK to use too. It is the new Pantone build system after all. However like the Lab conversions, many Bridge builds use 4 colors. Another problem with Bridge, it is not the Adobe default, so it's not widely used.

Something else too when using Lab source for solids. If the job is on a dull coated sheet you want to use Matte. The Solid Coated is for a gloss sheet. There is not a huge difference as with uncoated, but there is a difference nonetheless.

[Guest]

For more information, you can also consult

http://help.adobe.com/en_US/InDesign/6.0/WSa285fff53dea4f8617383751001ea8cb3f-7025a.html. 'Earlier versions of InDesign' should also be read here as 'earlier versions of the printed job or PDF' I think. Since if you change the values, this will also have it's affect on the printed output or in the PDF, if you convert the InDesign-file to PDF first.

[Lou Dina]

Thanks for the Help link, Didier. Good info.

That helps, but the original problem still persists, since different color libraries (from the same vendor) specify different Lab values for a given spot color. I guess that is to be expected to some extent, especially since so many spot colors are out of CMYK gamut. But, you would think Pantone would keep at least Lab values consistent for a given color on a given paper grade over the course of many years. I'd expect Pantone 300C to look the same on #3 coated stock from one year to the next, since it uses the same ink formulation on the same paper grade (as long as it is printed carefully and consistently with industry standard inks, papers and process control). That should result in Lab values that are a very close match from one year to the next. So, the designer is still faced with the problem of which set of Lab values to select.

Of course, Pantone 300M, printed on an uncoated #5 sheet, may well have different Lab values than 300C printed on gloss stock. It does make it confusing for the designer, and also for very discriminating clients (luckily, they are not always as picky as we designers are).

My current solution is to look at the Pantone Solid Coated library, and note the Lab and CMYK equivalents in my given CMYK working space, using Relative Colorimentric. If the color is out of gamut, I click the OOG warning and see what colors I get. Then I look at the Color Bridge PC library to see what colors I get. They're usually different. I might even look at the Solid to Process Library. The one that gives me the best "eyeball" match on my carefully calibrated monitor is the one I usually use, but I also exercise some judgement based on past experience.

For example, 2935C is one of the corporate logo colors of my previous employer (out of CMYK gamut, of course).  Here's info from three different Pantone libraries for coated stock (USA inks and standards):

Color Bridge PC

100c, 52m, 0, 0   43L, -13a, -50b This color is printable and is a decent match to the swatch book

Solid Coated

100c, 65m, 2y, 0    36L, -4a, -63b (this color is OOG so you have to bring it in gamut to evaluate)

100c, 65m, 2y, 0    39L, -6a, -49b (this is the Lab equivalent when brought into gamut—prints way too purple from lots of experience)

Solid to Process Coated

100c, 46m, 0, 0    45L, -17a, -50b (this one is too light and errs a bit toward cyan, but helps avoid purple shifts on press)

So, what did I use? From experience in printing this color over and over on corporate brochures, fliers, post cards, posters, etc, on lots of different presses, I settled on 100C, 50M, 0Y, 0K (between Color Bridge and Solid to Process). It errs slightly on the cyan side, but I find that much more acceptable that a purple shift if the magenta is laid down too heavy.

That illustrates the mess they have made out of specifying spot colors. Sort ot takes the science out of it, doesn't it? Well, nobody said it was perfect, and it is a good guide to get us closer to where we want to be, faster. But, judgement, knowledge and experience still come into play. I did a job for a customer recently, and their corporate logo uses 300C, very close to my old corporate color, 2935C. So, I purposely erred on the side of less magenta based on years of struggling with my own corporate color. They were happy with the printed job, fortunately. They selected the printer themselves after the job was designed and submitted. I submitted it as a PDF/X1-a file, stripped out all profiles to prevent conversions, and specified the original design intent as SWOP_Coated #3v2. I also submitted proofs and cautions about color matching, to let myself off the hook.

We have a ways to go.

Lou

[Printer_Rick]

The Pantone Lab values are fixed and should account for the paper white. 293 C is L29 a7 b-66. If I have a Pantone coated swatch book (not out of date), and read the color with a spectro it should be within 2 DeltaE (CMC) of the book Lab numbers.

The problem with processing a CMYK color back into Lab is having to use Absolute rendering to account for paper white.

To better explain, assume you are reasonably sure of the destination CMYK. But there's no way you will be able to read the press sheet with a spectro to get the printed Lab value.

You could take the CMYK build, map it to Lab using Absolute. The numbers should closely match the printed result.

[jethrodesign]

Thanks for all of the helpful dialog! It's sort of looking like there is no clear winner when it comes to choosing the best conversion solution. I wasn't so much worried about this job, it's more to understand for the future so we would have a clearly defined methodology.

- And for the sake of understanding, if we're usually printing on uncoated stock but WITH a U/V coating, would coated PMS values be closer than uncoated in this scenario?

I wish our conversions were a bit more consistent like the example Lou was stating (only 2 primaries being used, with only one really shifting). But in our case, we are seeing 2 or 3 of the primaries changing, sometimes pretty radically. So it makes it a bit tougher to 'average out' I guess.

But IN THEORY, would this be the order of preference for conversion?

1) Use the LAB conversion with accurate press profiles provided

2) Use Pantone Bridge PC conversion

3) Use the older Pantone CMYK conversion (default in InDesign and Illustrator)

THANKS AGAIN!

[Printer_Rick]

jethrodesign wrote:

- And for the sake of understanding, if we're usually printing on uncoated stock but WITH a U/V coating, would coated PMS values be closer than uncoated in this scenario?

Uncoated is closer. The UV follows the printing. Are you sure that it's uncoated with UV? It will be a duller color appearance with a shiny coating on top.

jethrodesign wrote:

I wish our conversions were a bit more consistent like the example Lou was stating (only 2 primaries being used, with only one really shifting). But in our case, we are seeing 2 or 3 of the primaries changing, sometimes pretty radically. So it makes it a bit tougher to 'average out' I guess.

But IN THEORY, would this be the order of preference for conversion?

1) Use the LAB conversion with accurate press profiles provided

2) Use Pantone Bridge PC conversion

3) Use the older Pantone CMYK conversion (default in InDesign and Illustrator)

There is no winner.

Bear in mind, with option one you can create alternate CMYKs with different black generations. In InDesign, you could choose to use a different rendering for solid colors too. Be very careful, though. There are jobs where InDesign swatches need to match Photoshop colors. In such case different renderings could cause a problem.

With option 2, for uncoated, Bridge has a separate library.

The coated, matte, and uncoated designations for the solids can get EXTREMELY confusing. Consider for a moment that you’re actually printing the Pantone, not CMYK.

A corporation specifies 292. It’s a very specific can of ink. But the ink has no Lab designation. It has to have the paper to get that.

So is the statement “our corporate color is PMS 292” really good enough? If the company means 292 C – what if the job goes on uncoated? The better solid uncoated match is 2915 U.

Now throw CMYK into the mix. Do you map 292 C Lab to the uncoated ICC? Or 292 U? Or 2915 U?

If there’s no way of knowing for sure, I personally would default to the C library and map that to the destination (because if you had to guess, the other person is looking at the C book, the most popular one).

As for Bridge, I would generally avoid using Bridge coated numbers for uncoated jobs, because Pantone specifically designed different Bridge builds for coated and uncoated. That was not the case with Solid to Process, where the same build values were printed in the coated and uncoated books.

Again I should reiterate, no winner when it comes to builds. The modern solution is Lab – CMYK. But there is a chance, somewhere down the line, some crotchety old dude is looking at a 10 year old solid to process book, and wondering why different numbers were used on his print job.

[Lou Dina]

Jethro,

-  And for the sake of understanding, if we're usually printing on  uncoated stock but WITH a U/V coating, would coated PMS values be closer  than uncoated in this scenario?

I do almost all coated stock, so I can't really answer that one. I suspect things will fall somewhere between uncoated and coated if using UV, but that is merely a guess.

But IN THEORY, would this be the order of  preference for conversion?

1) Use the LAB conversion with accurate press  profiles provided

2) Use Pantone Bridge PC conversion

3) Use the  older Pantone CMYK conversion (default in InDesign and Illustrator)

My experience has led me to choose different methods depending on a number of variables. (Sorry, I know you are looking for a simple, easy answer that fits all scenarios).

If I was printing on uncoated stock with a low total ink limit of 240–260, I'd be inclined to stick with a solution that reduces total ink load. For dark colors, this means more black ink (heavy GCR). This will result in less color shifting on press, but may also have a tendency to have a little less vibrance than a light GCR. But, it definitely reduces ink load, which may be a big consideration, depending on your final paper, press, ink.

CMYK is widely known to be deficient in reproducing blues, and there can be a tendency to shift toward purple, so I am always cautious if there are important blues in the job or if they represent a large area. I will select the method that tends to err on the side of less magenta, rather than more.

Light Pastels are another tough challenge for CMYK, especially if printing on a dingy, uncoated stock that has any yellow bias. The only way to get those light pastels is to spread out the dots and make them small, so the paper exerts a big influence on color, saturation, etc.

Another factor that weighs into my decision is the commercial printer. If I don't know who will print the job, I tend to be very defensive, pick a middle of the road standard profile (I've been usning the IDEAlliance SWOP2006_Coated#3v2 profile for my coated work). The final press may be able to utilize 320, 340 or higher TIL, but if I prepare the file that way, and it heads to a 280 or 300 TIL press, I'll have a problem. I also strip out the profiles to prevent unwanted conversions and hope the printer comes close to SWOP G7. If I am working with a color managed printer that I know and trust, I get their profile and design the job around their press and specifications. In that case, I feel more confident pushing things a lot farther, knowing that my hard proofs will match theirs (from experience with that printer). I still err on the side of caution with blues. The Solid to Process uses way too much magenta in the blues to suit me, so I'd use the Color Bridge or Solid Coated books for blues. I also prefer to convert all my images to the final CMYK space in Photoshop instead of InDesign.

The easy answer…I'd probably use Color Bridge to select my colors, since it is the latest guide and seems better with blues (at least to my preference). If I had important colors I had to hit, I would do an accurate hard proof and make sure the colors are acceptably close.

Using Lab is fine, but which Lab value are you going to select? The Lab values from different swatch books are different.

One final thought—CMYK is unlike RGB, HSB, LAB and other tri-stimulus spaces. With three coordinates, there is only one way to define a given color (assuming you have nailed down your working space...sRGB, Adobe RGB, etc). With CMYK, you have four colors to work with, and can create the same color with many different combinations, and they may all print exactly the same.

Lou