Methodology
The data collected for this project was amassed in three parts: through the measurement of drawdowns, a press trial and a qualitative survey. Drawdowns were conducted using a hand-proofer on MacTac white polyethylene terephthalate (PET) label stock with a Siegwerk L39 series UV ink set.
Over the course of conducting drawdowns, ink formulations were adjusted in order to most easily match Flexographic Image Reproduction Specifications & Tolerances (FIRST) and industry standard density and L*C*h°. Violet, magenta and orange inks were extended using Siegwerk UV Extender Varnish. Beginning anilox volumes were chosen according to recommended aniloxes for film substrates, and were adjusted due to limited inventory and to match FIRST and industry standard densities and L*C*h° values for CMYK+OGV inks. Resulting conditions are demonstrated in Figure 1.
Drawdowns conducted included the following: CMKOGV inks alone, CMYKOGV inks and their respective primary pairs, and CMYOGV inks and their pairs over black (in line with the research published in November 2016’s FLEXO Magazine). After conducting these drawdowns, the samples were measured (for density, L*C*h°, L*a*b* and opacity) using an X-Rite eXact spectrophotometer. Measurements were averaged between three samples to minimize any anomalies created by the variable drawdown conditions.
After gathering measurements, four optimized EG sequences were determined using the four methods defined by previous researchers at Clemson (Figure 2), which indicated an appropriate placement of orange, green and violet inks amongst their corresponding primaries based on opacity, overprint chroma and a simplified gamut space calculation (overprint triangle areas and tetrahedron volumes).
After determining a print sequence from the four proposed models, Esko 8-stage 29-by-29 targets were imaged two-up on DuPont DPR 0.067 digital photopolymer plate material using an Esko CDI Spark. Files were RIPed using conventional (120 lpi) and stochastic (Monet for Flexo at 220 lpi) screening on a 4,000 dpi RIP. Plates were mounted on 18-in. repeats using 3M 1320 stickyback (determined to be the most appropriate after conducting a benchmark test). Plates were run on an Omet Varyflex 530 flexographic printing press located at Clemson University’s Sonoco Institute of Packaging Design and Graphics, as the modular, inline configuration of this press most easily allowed individual color stations to be rotated as needed for each of the four sequences to be printed.
For each sequence, the same MacTac PET label stock was used, and aniloxes were chosen based on those used during drawdowns and the available inventory at the Sonoco Institute. The use of UV inks ensured more stability and ink specs were kept within 1.5 Delta E 2000 between each print sequence run. After ensuring color within tolerance and even impression, the press was run at 300 fpm to allow it to pull art into registration and run for 150 feet. In conjunction with the use of Esko 8-stage targets, the Esko Equinox sequential run strategy was utilized, so that orange, green and violet could be isolated and analyzed with their respective CMY pairs. Five samples of CMYK, CGYK, OMYK and CMVK were taken from each sequence to be measured.
Measurements from each set of samples were taken using an X-Rite i1iO table and averaged using Esko Equinox Profile Creator to create corresponding EG profiles. The resulting profile gamuts were analyzed using Esko Color Engine Pilot using the Gamut Checker and Gamut Viewer tools. The Gamut Checker tool allows the user to determine any changes in the number of achievable Pantone colors, while the Gamut Viewer tool allows the user to visualize any changes in the numerical volume of the gamut. Both of these analyses were used to determine “best” and “worst” EG print sequences. In addition to this analysis, the patches embedded in the color target which contained 100 percent overprints of each CMYKOGV color and its pair over each other and black were specifically analyzed to determine the accuracy of initial drawdowns to the press conditions.
The best and worst sequence profiles were applied to a sample of artwork, which simulated a consumer package that may be found in a normal store. This artwork was imaged to digital photopolymer plates in the same process as the characterization plates. Using the same printing conditions as the characterization run, this artwork was printed using both sequences and samples were taken.
The samples taken were then used to conduct a qualitative survey among Clemson University students. Participants were asked to observe the samples to distinguish any differences in color and, if they were able to see a difference, which sample they preferred.
Print Sequence Results & Discussion
Figure 3 (above) illustrates the resulting gamut volumes of each sequence printed in cubic L*a*b* units (shown on the y-axis). In addition, the difference between these sequences is expressed along the x-axis in terms of percent increase, where the smallest gamut represents 100 percent of the color space.
This result is anticipated when compared with the results from the original study conducted at Clemson. As expected, the sequence predicted using the mathematical volume of a tetrahedron-shaped color space yielded the largest gamut, while the sequence predicted using only ink opacity yielded the smallest gamut. In addition, the percentages of gamut increase between the smallest gamut and largest are similar—7.7 percent compared to 6.6 percent here. However, unlike the original study, there is no real median gamut size, with the two largest gamuts having similar percent increases and the second-smallest gamut having only a 1.1 percent increase.