Dickeson--Cyan, Yellow, Magenta, Black Magic

"Where's the payback?" I asked myself at the Sunbelt Computer and Graphics trade show in Atlanta in early March. Here was a new product introduction called CSR V.1, being offered by QIP Inc., that baffled me.

I watched as one of the representatives fanned out the color bars of a dozen or so printed press sigs on a flatbed scanner. He lowered the cover and clicked a button in the software. All of the bars appeared on a computer screen exactly as he had spread them.

After about a minute or so, a window opened on the screen with the sample set data for density, dot area, dot gain, print contrast, trap, hue and grayness. Data for each segment for each of the dozen bars appeared. Accompanying the data were charts with curves and plots of all sorts together with the identifying number of the sample set.

Sophisticated Software
What confused me was the absence of a traditional densitometer making the readings and recording the information. "How did you get the readings from all those samples for each of the black, cyan, magenta and yellow segments on each of those sets at once?" I asked.

"From the scanner information," the rep answered.

"But suppose you didn't have those sheets at the same space intervals and got them cocked or staggered on the scanner bed. What would happen then?" I asked.

The rep responded by putting the sheets down at random angles with varying distances between them and pressed that button once more. Same results. They appeared on the screen at those various angles and distances apart, and the numeric results, graphs and charts were displayed.

"You see, Roger, the scanner software programs find the color bars you lay on that 11x17˝ glass bed wherever they are and however they're displayed," the rep said.

"How wide do those color bars have to be?" I asked.

"Width doesn't matter," the rep replied. "Quarter of an inch, eighth, 16th, 32nd, whatever you want. You print 'em, and the scanner finds 'em and measures 'em."

A group called Apogee apparently developed the programs that identify and analyze color bars using some of the same techniques used by military intelligence to identify and count tanks, trucks and guns from aerial or satellite photographs. Apogee programs gather the scanner data from the color bars and pass the information to the Color Metrix programs.

Color Metrix utilizes programs and techniques it has developed over the past few years to analyze and process the raw data into graphs, charts and tables significant to the printing operations.

Really, it was some "gee-whiz" technology. But I'm beyond being snowed by all of the technological magic of today. So I asked myself aloud, "What do you do with that information?"

Information that doesn't lead to decisions and actions is worthless. You've got the gain, contrast, density, hue, etc. "What do you do with it?" I asked a guy standing next to me.

"Tell you the first thing I'd do with it," he fired back. "I'd take those charts and tables back to the color separation crowd and tell them, 'This is the real world. This is what that press can print on that paper using those inks. Now make your color separation scanners match those real numbers!' Then I'd take it to the digital proofing folks and say, 'This is your product specification. These numbers are what your proofs must reproduce.' "

BR>Good thought? What it suggested to me was that you could take, say, 10 sample sets of 10 sigs, compute the information, and come up with a realistic approximation of press process capability. Sorry, gang. No more smoke, mirrors and quality illusions. WYSIWYG. What you see with these numbers and curves is what you get from the process.

Don't like it? Then buy better paper, press, inks, blankets and press controls to move the process level up to what you want—or what you can get your customers to pay for. Stop hollering at the press people and diddling ink keys to match proofs you can't possibly match. Think of the wasted press time and materials this might save. Agree?

Hold on a moment. Suppose I took those 10 sample sets, computed the averages of the data, found the standard deviation of the samples from the average, multiplied it by three, then added and subtracted it from the average to arrive at upper and lower control limits.

That sounds like SPC (Statistical Process Control) Shewhart Charts, doesn't it?

So I asked one of the reps if they were doing that. "No," he said. "Rog, we haven't programmed for that yet. We can do it any time."

But then, I wondered, would printers use it? "Never underestimate the ignorance of the public" is a guiding principle of marketing. I hate that kind of thinking. It's wrong. I say, "Never underestimate the intelligence of printing people."

If they don't see it at first, teach 'em about investigating special causes when a number penetrates a control limit in order to improve the process. If it's right, printers will grab it and run with it! Would that save time and materials and make for happier customers? What do you think?

Then I thought about all the times, when I was a printing CEO, that a publisher had demanded a credit for quality because the ad agency was beating him for a credit or "make good." Suppose I handed the customer a spreadsheet with all the numbers from all the samples from their job from such a system and let them see the process capability, the oscillations within the control limits and maybe an instance of a special cause where we shut the press down to investigate.

It's embarrassing sometimes to let facts get in the way of a strongly held opinion—but sometimes you must. It would sure take the illusion out of samples, wouldn't it?

Maybe QIP has something there. Time will tell.

—Roger V. Dickeson

About the Author
Roger Dickeson is a printing productivity consultant based in The Woodlands, TX. He can be reached via e-mail at rogervd@bigfoot.com or rogervd@pdq.net; or via fax at (281) 362-7572.