JUser: :_load: Unable to load user with ID: 166

Color scanning: Hitting color, part 2

March 18, 2007

I am a big Edward Demming fan, and his comments about process control are right on. Demming said that when errors in production occur, “It’s the process, not the people.” So, let’s control the process. To follow a chronological order, color scanning is first, followed by proofing, platemaking and presswork.

At some point, you are going to need access to Photoshop for the two exercises I am explaining in this article. It shows differences that are difficult to communicate in words alone. And, if you haven’t read my free “Process Control Reference Guide” yet, contact me at dremaley@piagatf.org. I’ll send you a copy.

Color scanning has changed dramatically since the invention of the electronic color scanner. The early color scanners were expensive and required a lot of training. The new flatbed scanners, with the use of Adobe Photoshop, are more user friendly. This change has been difficult for the industry. The fundamentals of color separation have been abandoned, and new people and techniques are adapted every day. Consider this: A Hell or Crossfield color scanner sold for around $200,000. These scanners had nothing to “view” color—no color monitors, no color screens, nothing. The only “viewer” was a small screen with measurements of density and dot area, with no capability to actually “see” the color image. Now we have color monitors and preview screens. Has the quality of color become better or worse than before? I’ll let you decide.

These early color scanners came complete with a series of transparencies made with different emulsions (Kodak, Fuji, Agfa, etc.). These images were scanned, proofed and evaluated for color, and “custom values” were entered into the scanner as a benchmark.

Following this principal, PIA/GATF has introduced a series of photos (high key, low key, flesh tone, saturated color and neutrals) available as prints, transparencies or digital RGB. These are excellent to set up and monitor your scanner’s ability to capture and repeat color settings. By scanning and proofing these images, you can decide if your scanner settings are correct, and then you can measure them regularly. These common images can be tracked for changes in the scanner or proofing process.The process of color separation is to provide images that can be reproduced in your color reproduction system—your printing press. This product is listed in the “Process Control Catalog” available at www.gain.net.

Color fundamentals The fundamentals of color scanning:

  • Tone reproduction
  • Gray balance
  • Color correction

The ideal way to set up the tone reproduction curve is to print a test target, such as PIA/GATF’s test form, or a similar form that can show how the press prints under a “standard” condition. What is a “standard” condition? In my mind, it’s a set of numbers (like SWOP) that has a numerical setting with a measurable tolerance and defined reference. Another standard condition may be your own specification, written with known values for density, dot gain, etc. If you haven’t established your specifications, use SWOP as a reference guide.

A little secret: Most scanners and Photoshop settings use SWOP numbers for the RGB to CMYK conversion. The numbers for press are as follows:

Density Yellow 1.00 Magenta 1.40 Cyan 1.30 Black 1.60 Dot gain (TVI): Yellow-18; Magenta-20; Cyan-20; Black-22.

Tone reproduction is the ability to reproduce an original as accurately as possible within the restrictions of your color reproduction system. A simple example is a black-and-white halftone printed on newsprint: The original has a density range from zero to 2.0, but newsprint cannot reproduce that wide of a density range. Newsprint’s maximum density is around 1.20. To reproduce this original, we compress the shadow tones and adjust the midtones to make a smooth transition. This applies to process color on a sheetfed or web press, as well. The dot gain adjustment also is a function of tone reproduction. With known dot gain values, we can customize Photoshop for our in-house conditions or we can select SWOP from the pull-down menu settings.

Exercise 1

Just so you understand what is happening in the Photoshop settings, select newsprint from the (color settings) pull-down menu, then select an RGB image, and convert it to CMYK. Save this CMYK file. Now set Photoshop (color settings) for SWOP coated, select the same RGB image, and convert to CMYK. Bring both images (newsprint and SWOP) to the screen, and notice the “weight” reduction between these images. SWOP is 20 percent in the midtone and newsprint is 30 percent; the newsprint separation has more “weight” removed for the print condition (it appears lighter). Under “custom” you could set it for any amount of gain to match your print condition. The last setting in tone reproduction is the “total ink” limit—the darkest area of color cannot exceed a total of 240 percent (for newsprint), 320 percent for SWOP. This means the total percentage of C, M, Y and K cannot equal more than 320 percent.

Gray balance is the heart and soul of color reproduction throughout the process. In scanning, gray balance means that all areas of the scan are neutral in color, highlight, midtone and shadow. This is very important, because any color correction will be misguided if correct gray balance is not achieved in the beginning. To establish good gray balance, we again refer to the press and its print condition. Without a press run, here are the numbers according to SWOP, which can be entered into the eyedropper tools in Photoshop:

For black-and-white, the numbers are: highlight, three percent; shadow, 95 percent (coated stock).

Your numbers might be different, depending on your press condition and process, but these are great starting points.
Color correction is the reduction or addition of color in overprint areas or color areas that are a problem in the original. All color correction must be done after gray balance is achieved.

The judgement of the separation should be made on a color calibrated monitor and/or the dot percentage values in the Photoshop information window.

Our methods must be accurate and consistent for measurable color reproduction. Images supplied from digital cameras require the same methods for faithful, quality color reproduction. With digital photography, there isn’t a color print to “compare” or reference. An item called a MacBeth color checker is an excellent item to photograph in your first exposure, to quantify the lighting and camera exposure. It has color patches that are referenced to a Lab value, therefore you can see what the original color is to look like.

Now that we have established our standards and tolerances, it’s time to help the process along with the use of GCR. GCR stands for Gray Component Replacement, subtracting yellow, magenta and cyan in all trichromatic colors and replacing them with black. It can be done as a percentage (i.e. 70 percent, which equals the total amount of black replaced) on a high-end scanner, or in Photoshop (as light, medium, heavy or maximum). There are many advantages to using GCR. It “hides” some of our color reproduction problems.

  1. 1. We use less of the expensive colored inks and more of the (less expensive) black ink.
  2. 2. The black contains nearly all the detail, (the colors become more or less saturated), so misregistration is less noticeable.
  3. 3. We have less color shifts at press because the values of Yellow, Magenta or Cyan are moved away from the midtone (where color shifts rapidly).

Black now becomes very important. It has a full tone value from highlight to shadow with lots of midtone areas, vs. a traditional separation black that has a “ghosted” or “skeleton” black. When the black changes at press, the images become “lighter” or “darker” but not “color casted.” Smaller presses with limited controls and web printing benefit greatly from a GCR separation.

Application of GCR is the last step after all gray balance, tone reproduction and color correction has been completed.

Exercise 2

To help you understand this method, set Photoshop (color settings, under custom) for maximum GCR, open the same RGB image we used before and save it as CMYK with maximum GCR. Now open the (CMYK) non-GCR image, and compare the black channel of the GCR image—note the “full” tone of the black printer. The most dramatic difference is in the color channels. Bring both Magenta printers to the screen—visually and with the eyedropper tool, we can examine the noticeable difference between these images, such as the very little Magenta in the GCR image. If there is less of this color at press, then it can’t shift as rapidly as the other (non-GCR) image. Compare the Cyan and Yellow, as well.

If these GCR images are supplied to the prepress department, they can’t be “color corrected” because of the absence of Y-M-C values. Similarly, they cannot be adjusted or corrected at press. This brings about a whole new discussion in color reproduction: Can you “color correct” at press? Do you want to “color correct” at press?

My position is that I supplied the press with the best color rendition of the original, with correct gray balance and color correction. When scanning, I know any change in gray balance affects all the colors, and the same is true on press. If we increase magenta on press to improve flesh tones, for example, we will ruin grays, blues and purples.

We have defined the process, measured the attributes and confirmed the output. Now, even with different images, we will be able to repeat the process and control it.

A correctly made color proof is the most accurate method for color analysis; that’s our next step in the process control system.

Dan Remaley has 30 years’ experience in color lithography. He is senior technical consultant of process control for PIA/GATF. Contact him at (412) 259-1814 or dremaley@piagatf.org.

Leave a comment

Make sure you enter the (*) required information where indicated.Basic HTML code is allowed.