Traceability to Standards – does it matter?

One of the many issues surrounding closed loop color control is the issue of traceability to standards. Some folks seem to place great value on it, and certainly it sounds good. The idea that the accuracy of a device used in production can be quantified relative to a known, presumably â??perfectâ?? source, somehow suggests that device must be more accurate and therefore anything itâ??s used for must somehow turn out better.

But does it?

QuadTech has a pretty sizeable machine shop. We have fully automatic machining centers that boggle my mind when I think about the stuff they can do. We have traditional non-automatic tools as well, and even have the usual manual measurement tools. Like micrometers.

As a mechanical guy, I understand micrometers. You calibrate them using gauge blocks â?? little metal blocks made very accurately to size â?? that are in turn calibrated by someone else, and so on until you get to the Big Cahuna gauge block owned by NIST in Washington. Using this system a few times a year we make sure our micrometers are as accurate as they can be. This is required by our quality procedures, as it is at most competent companies.

The problem is that if you use a micrometer to measure the diameter of a cotton ball, the accuracy of the device relative to the accuracy of the measurement becomes irrelevant. Cotton balls, being pretty squishy, donâ??t need to be measured to .001 inches, they only need to be visually correct â?? assuming anyone anywhere is measuring them at all. But this is where the analogy falls apart because folks definitely are measuring color.

When youâ??re measuring the color of a web flying by at thousands of feet per minute, it doesnâ??t matter if the surfboard on the cover is within .1 DeltaE or .2 DeltaE. Whatâ??s important is that itâ??s good enough for the customer.

Once the job is past make ready, closed loop color control is really just cruise control â?? the target values have been determined, and the system seeks to maintain them, albeit via some sophisticated algorithms in some cases. With that comes measurement and data collection, of course, but thatâ??s for analysis and process improvement, not for control. Now, you could argue that much higher levels of accuracy have value in these measurements because they could signal subtle shifts in color. My answer to that would be to look at a graph of density versus time, and think about how much noise is in that data, and also the accuracy of the other variables needed to make use of the data.

Another way to look at it is to consider when things donâ??t go well. Suppose that for whatever reason the color doesnâ??t pass muster. Is anyone going to contact the manufacturer to check the calibration? Is a detailed trail of documents showing the pedigree of a sensor going to solve the problem?

Not likely, because the reason why the color is wrong isnâ??t because the system made it wrong, itâ??s because either the press cannot hold the color or the color isnâ??t possible to reach given the plates, paper and ink (I was astonished to learn that you can not actually print every color you want to â?? a fault of the way printing works) or the wrong color was set in prepress.

So, does traceability to standards really matter?

One thought on “Traceability to Standards – does it matter?

  1. One would think traceability matters since there is a need to compare print done in one plant or process with print done in another. The question is how accurate will it be or can it be.

    Measuring colour is problematic on its own. On a newsgroup called sci.engr.color I once read that there was a dirty little secret in the industry. It was about the problem of having different colour measuring instruments reading the same values. It seems that it is common to have discrepancies even with calibrated instruments. If there was a requirement to get a truly accurate reading, one had to go to a national lab where there was a multimillion dollar instrument. Not very practical. The point is that one should not expect high accuracy on any reading. Just hope for a practical level of accuracy.

    Also if one is measuring with an RGB device, then you can not get an accurate colour value because RGB can not measure colour. Applying a profile to get colour values is not guaranteed. It only gets you close. A good guess. To measure colour, you must measure the spectrum.

    Then again there is the problem of colour. Having a small Delta E value does not ensure a match. People do not see colour quite the same way. In some abstracts of papers, it has been stated that the variation between real observers is about DeltaE = 2. This difference in observers is called “observer metamerism”.

    Colour science does not guarantee that you see colours as they are defined by the science used in graphic arts. If you have the same measured colour in an image in two locations, the colour can be perceived differently based on the surrounding colours. It is an optical illusion but one that can be in any image.

    The trick is in knowing how much accuracy is the practical limit.



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s