An experiment was performed to compare the effect of gray component replacement on color separations created with the Hell 399ER scanner software compared to two desktop algorithms. Three different GCR methods were tested; the scanner method where the separations were performed and films output on a mid 1980's model Hell 399ER laser scanner using Hell's first generation GCR algorithm and two types of desktop methods where the scans were done in RGB, color separated on the desktop, and films generated on an Agfa Selectset 5000 imagesetter. The two desktop methods used were RIT Research Corporation's RGB-CMYK transform and Adobe Photoshop 2.0. All scans and proofs were made in the Color Separation Lab at the Rochester Institute of Technology (RIT) in Rochester, New York. The desktop separations were made and film output in RIT's Electronic Prepress Lab. For each separation method, three levels of GCR were performed; a non- GCR (0%) separation, one at 50% GCR, and one at 80% GCR. In addition, the separations performed on the Hell Scanner were done with Under Color Addition (UCA) off and on. After separation, the scans were output to film and proofed using 3M Matchprint II. Two proofs were made. One contained the six separations performed on the Hell scanner and the other contained the six separations made using the desktop software. The proofs were then measured using a spectrophotometer and the results compared using the L*a*b* color space. All color comparisons were done using the non-GCR separation as the color reference. Each AE* measurement represents the color difference between a patch on the non-GCR target and the corresponding patch on the target produced using GCR. The result of the experiment was the rejection of the stated hypotheses that there would be no significant color difference between the output produced from the scanner separations and the desktop separations at two levels of GCR. The experiment showed that the algorithms used to perform GCR on the desktop produced less color variation than the algorithm used in the Hell 399ER scanner at both 50% and 80% GCR. The results also showed that in almost every case, the amount of color variation increased as the level of GCR was raised. It could not be determined whether or not Under Color Addition (UCA) had any significant influence on color variation for the separations performed on the Hell scanner. Based on the results of this experiment, color professionals using a desktop production workflow should be encouraged to take advantage of the benefits of GCR without fear that the color of the reproductions will be compromised.
Library of Congress Subject Headings
Color separation; Color printing; Colorimetric analysis
Department, Program, or Center
School of Print Media (CIAS)
Orino, Thomas, "A Colorimetric performance comparison of gray component replacement algorithms" (1998). Thesis. Rochester Institute of Technology. Accessed from
RIT – Main Campus