What Color is a PMS Color - Really

[Bill W]

Greetings,

Take PMS 279 - something our customer wants us to match. Here are some values from an Xrite 528 spectro, and a i1 via MeasureTool measuring a new pantone book; Photoshop; and EyeOne Share Pantone Library:

528:
65/10 - L=62.18, a=-5.61, b=-43.68
50/2 - L=58.75, a=-2.38, b=-49.58

i1:
65/10 - L=61.6, a=-11.4, b=-44.7
50/2 - L=58.9, a=-2.6, b=-49.1

Photoshop:
L=58, a=-5, b=-46

EyeOne Share Pantone Library:
L=57.5, a=-1.7, b=-52.5

Now I know that the correct color is the one that matches the swatch supplied by the customer and if we measure it with the same device and illuminant in our shop we will be able to meet the customers needs.

However, it would be nice if the tables in Photoshop and the one in EyeOne Share matched - one would thing they would. Additionally, I am curious why the 528 and the i1 were relatively close under 50/2, but not so under 65/10.

Thanks to anyone that responds.

-Bill-

 

[twooodogz]

This brings me to a question I have not yet found an answer to. G7 is highly LAB dependent so I have been looking for a way to verify my instrument is reading the proper values. I have seen many T-REF cards that verify density where is a standard reference for LAB values?

 

[gordo]

PhotoShop is the closest. These are the Pantone values for PMS 279 extracted from a RIP:
PANTONE 279 C L=58.1, a=-5.02, b=-45.82

best, gordon p
my print blog here: Quality In Print

 

[meddington]

. Additionally, I am curious why the 528 and the i1 were relatively close under 50/2, but not so under 65/10.-

I would say that the two devices are exhibiting a bit of metamerism failure. The spectral curves of the sample as measured by two devices must be fairly similar in the area emphasized by D50 lighting, but different enough under D65 lighting to cause more significant deltas. Which one is right? well, unless you can prove that one or the other instrument is out of spec or otherwise deviant from the norm, they both are.

 

[meddington]

I have seen many T-REF cards that verify density where is a standard reference for LAB values?

I can't say that its an unequivical standard reference, but there is a LabRef available that comes with expected lab values for 10 or so different patches. Lab values are rendered from an xrite 938 (crossed referenced to a Milton Roy). Thing is, depending on the stock used to print the cards, and the lamination used to protect it (glossy or matte) a given instrument can report different results. Ceramic tiles would be more ideal, but this doesn't help predict deviation on actual substrates all that well.
No easy answer I'm afraid.

 

[dimitri]

metamerism could be an issue for the D65/10; is the eyeone io uvcut or not; that could be another difference with all the blue light on D65... but that is a speculation. I do not know if UV filters affect D65 more than D50.

what I do not get is why there is difference between two digital libraries? Aren't they both licensed from the same company? They should be identical. Is Photoshop on D50/2 and if so, are there preferences to change that?

and another related thought: Why not use D50/10 as a 'standard' for spot colors? Theoretically, D50 is the standard for the printing industry, and 10 degrees is meant to provide more accurate measurements for areas wider than 4 degrees.

As for the LabRef card, you got to make sure you order one for the right observer,one that matches your device. Then, I would say that due to the issues that mike mentioned it can be very useful as a means to compare multiple devices, in the case that you cannot match the measured values of the patches.

 

[twooodogz]

So I could get this LabREF from xrite since my device is a 528?

 

[dimitri]

Is a 'trick' question?
your X-rite should be able to hit the values of the LABref. These cards come with fairly good instructions on their use as well.

labref color - Google Search

and this presentation explains instrument variability very well too:
http://www.piagatfcolor.org/documents/2008handouts/RadencicG_Intro C Measurement.pdf

 

[encompus]

Here are my values without measurement but just referencing 279:

ColorMunki Design L*57.8 a*-4.0 b*-46.2

i1 Share L*56.1 a*-2.6 b*-49.4

Color Picker L*58.10 a*-5.0 b*-45.85

I am guessing it has to do with the library data. The data has to come from measurement data sets - most likely averaged. Therefore there must be a Delta average for the data sets coming from PANTONE. I do wish X-Rite's applications at least shared the same data set.