Spot color to CMYK conversion


Hi all - I've been reading this forum for a long time and occasionally comment, but this is my first question. Please hang in there with me, I had a hard time writing it up as short as possible and still including vital info.

Two questions:
1. What is the proper way to convert spot colors into CMYK? As a general rule I do not warn customers when I am doing it unless there is a drastic change. I prefer to convert in Illustrator using SWOP because I feel Illustrator handles color the best.
2. Why would the same swatch look different in one file vs. a newly created file. It converts to a different CMYK breakdown as well. I copied and pasted the swatch out of the old file into the new file.

I received an inDesign file for print. It was set up in CMYK with PMS 021. I converted everything to CMYK and created an output file in Illustrator (.ai) to make films out of, my normal work flow. All colors look good and output the way they are supposed to. Get it on press and all the colors are printing like they should until I grab the packaging printed by a vendor and the oranges are off. The vendor apparently converted 021 as well but through InDesign and their result was a light orange (C=0, M=50, Y=100, K=0) where mine was a dark orange (C=0, M=80, Y=100, K=0). To give you a better idea of how different the colors are, mine is closer to PMS 7579 (page C 23) and theirs is close to PMS 2012 (page S2 C). Checking some color discussions online I feel that the correct conversion is mine. Checking their files though it seems the customer wants the lighter orange (C=0, M=50, Y=100, K=0) since that is all over their website. I checked their logo in a few different files, and everywhere it shows up as a light orange even though the swatch is called 021. I attached a screen shot of PMS 021 in a blank Illustrator document compared to how their logo looks in their file. As you can see they are drastically different. I would say they just created a new swatch and called it 021, but when I copied their logo into a new document the 021 showed up in the correct darker/vibrant color. The original file was an Illustrator 3 .eps, which is my only guess as to why the colors change between files. There do not appear to be any unusual Transparency settings on their end.

This disturbs me, because the customer supplied files to us and they printed two very different ways. I strongly believe my version is technically correct but looking over all their work on the website it's clear they want it to print a light orange, which is not PMS 021. How am I supposed to guess at what they want if they are supplying art this way and everyone else happens to print it different than I do?

BACKGROUND: I've been in Prepress for about 14 years. This is the only job I've had in this field and everything I know is self-taught. I am the only one in my department. I work at a screenprint shop with some offset, but we print on plastic rather than paper so our presses and inks are completely different than in other shops. We have no color bars, registration marks to double check the color. I have the CS6 suite but no other post prep software like PitStop. I set everything up manually in my software, I do all trapping by hand and all films are created by me through InDesign. I check all output percentages and compare to the original file to make sure nothing has been changed in the conversion before creating films.


  • Colors.jpg
    6 KB · Views: 481
Between CS5 and CS6 InDesign and Illustrator started defining spot colors with LAB values instead of CMYK values. The difference when converting to actual CMYK can be drastic. That's why, if customers are picky about their color, I recommend they convert spot colors themselves rather than assuming the conversion of a specific spot to CMYK will always come out the same. Not that they always listen, it's just what I recommend ;)
Last edited:
are the pms colors maybe from the pantone plus libraries vs. the old pantone libraries? in addition to going to LAB they also changed the formulation on some of the colors, though i found out the other day that some of the old values are retained. but one of our colors changed pretty drastically.
Not sure if this is true but it seems to me that in the old days, printers could not so accurately print the colour in images but they knew exactly how to come close to a CMYK spot colour because they printed those catalogues that showed how they would print different combinations of CMYK screen values.

Is it now the case that it is easier to print the colour of the image but they have more problems with CMYK solid representations? :)
The perils of blindly following Δ deltas over the human observer are very apparent with CMYK conversions of Orange 021…

EDIT: Lets take a look at the (rounded) L*a*b* values for PANTONE 021 C and PANTONE+ 021 C, notoriously OoG (Out of Gamut) for common CMYK offset press conditions:

(Old) PANTONE 021 C: L63 a63 b85

(New) PANTONE+ 021 C: L61 a66 b85

Stephen Marsh
Last edited:
PANTONE+ Coated library files are now Lab colour based. If one is after a “one size fits all hard wired recipe” (there is no such thing), then there are the PANTONE+ Color Bridge books which are CMYK based like the old Coated library files used to be.

PANTONE+ Color Bridge CMYK values for 021 C: m65 y100

Printing in ISO Coated v2/Fogra39 CMYK, these values equate to:
L63 a46 b76 (Relative Colorimetric)
L60 a42 b61 (Absolute Colorimetric)

Compared to:

L61 a66 b85 (original PANTONE+ 021 C Lab values)

What does this mean in Difference values?

This is an approx. ΔE76 value of 22
This is an approx. ΔE94 value of 5.7
This is an approx. ΔL* value of 2
This is an approx. Δa* value of -20
This is an approx. Δb* value of -9

The plain english take away conclusion? All of the numbers and colour analysis above agrees with the human observer - the m65 y100 F39 colour values are not even close to matching the original spot colour! :]

This is just setting the scene for further analysis…

Stephen Marsh
Last edited:
If a colour is "Out of Gamut" in CMYK, there are no "correct" CMYK values. It comes down to the human observer, personal preference, this is now subjective, it is "art" and not "science".

The original post from GLMTart mention these CMYK builds (presuming F39 CMYK, Relative Colorimetric Intent):

m50 100y = L70 a32 b78
m80 100y = L56 a59 b65

The old Pantone Coated library had hard wired CMYK values for 021C of m53 y100.

The problem with creating a "bright" orange in press CMYK is that it is not really possible. We can't get a stronger positive b* value on common paper, as the Yellow value is already at 100%. Once we start going higher than say 50-60% magenta (depending on TVI), our "orange" starts looking "red". So historically, most people have settled on a value of around m50-60 y100 as the “best” CMYK orange that they can achieve.

m80 y100 in difference values to the original PANTONE+ 021 C Lab values:

Approx. ΔE76 value of 21.7
Approx. ΔE94 value of 6.6
Approx. ΔL* value of -5
Approx. Δa* value of -7
Approx. Δb* value of -20

m50 y100 in difference values:

Approx. ΔE76 value of 35.8
Approx. ΔE94 value of 13.9
Approx. ΔL* value of 9
Approx. Δa* value of -34
Approx. Δb* value of -7

Compare the numbers above, the m50 y100 value has higher ΔE values, so they are worse, right? I guess that depends upon the human observer! :]

Even though a Δa* value of -7 is in theory better than -34, in this particular human observer’s opinion the build of m50 y100 is "better", despite the numbers saying otherwise.

In theory, having Lab based colour library files and converting to CMYK using a profile offers better results. This is true in many cases, the more so when spot colours are “in gamut” for the CMYK target condition. However, when spot colours are “out of gamut” - it often comes down to “craft” or “art” over “science”.

The plain english take away conclusion? I agree with your customer’s historical colour build choice of m50 y100 as being a “better” orange (better as in a more acceptable orange in CMYK that does not even attempt to come close to the original Pantone colour).

Stephen Marsh
Last edited:
I do think the supplied colors come from an old formula. My concern is that the electronic version is the only one I have to go by. I verify mostly by percentages in a color rather than a visual and I would have tended more towards the darker color orange/red as a match rather than a light orange.
I do think the supplied colors come from an old formula. My concern is that the electronic version is the only one I have to go by. I verify mostly by percentages in a color rather than a visual and I would have tended more towards the darker color orange/red as a match rather than a light orange.

The thing about colour is that it is context dependent.

Imagine that you had a job which called for “bright” oranges and reds. Both were equally important. If you made the orange m80y100 and the red m100y100, then there is not much distinction between the two colours. In this case m50y100 would be more appropriate.

If the job only had orange as the primary colour, then say m70y100 would probably be more appropriate.

If you had a printed CMYK sample then you at least have a known target to match, despite what you think the colour might be.

Stephen Marsh


A 30-day Fix for Managed Chaos

As any print professional knows, printing can be managed chaos. Software that solves multiple problems and provides measurable and monetizable value has a direct impact on the bottom-line.

“We reduced order entry costs by about 40%.” Significant savings in a shop that turns about 500 jobs a month.

Learn how…….