Predictability of Spot Color Overprints Robert Chung, Michael Riordan, and Sri Prakhya Rochester Institute of Technology School of Print Media 69 Lomb Memorial Drive, Rochester, NY 14623, USA emails: rycppr@rit.edu, mprppr@rit.edu, sreehemanth@gmail.com Keywords spot color, overprint, color management, portability, predictability Abstract Pre-media software packages, e.g., Adobe Illustrator, do amazing things. They give designers endless choices of how line, area, color, and transparency can interact with one another while providing the display that simulates printed results. Most prepress practitioners are thrilled with pre-media software when working with process colors. This research encountered a color management gap in pre-media software s ability to predict spot color overprint accurately between display and print. In order to understand the problem, this paper (1) describes the concepts of color portability and color predictability in the context of color management, (2) describes an experimental set-up whereby display and print are viewed under bright viewing surround, (3) conducts display-to-print comparison of process color patches, (4) conducts display-to-print comparison of spot color solids, and, finally, (5) conducts display-to-print comparison of spot color overprints. In doing so, this research points out why the display-to-print match works for process colors, and fails for spot color overprints. Like Genie out of the bottle, there is no turning back nor quick fix to reconcile the problem with predictability of spot color overprints in pre-media software for some time to come. 1. Introduction Color portability is a key concept in ICC color management. Color portability between colors captured and colors displayed (or printed) is accomplished in an ICC-based pre-media software in two color conversions. As depicted in Figure 1, the first step is to translate color data captured into the RGB working space via the input device profile. The second step is to translate color data from the RGB working space into the monitor space (or CMYK printer space) via a monitor (or output) profile. Figure 1. Color portability in color management workflow There are two profiles used in a color conversion. The source profile provides the input-to-pcs (A-to-B) conversion, and the destination profile provides the PCS-to-output (B-to-A) conversion. A color conversion is also described as the A-to-B-to-A conversion. In a late device-binding workflow, pictorial color image data are converted and saved in the working space, such as Adobe RGB (1998). Thus, the digital file can be rendered to any number of output devices with the appearance of the monitor display. Color rendering intent of pictorial color images is either perceptual or relative colorimetric. In the context of pictorial color image reproduction, color portability means WYSIWYG or good color agreement between display and print as the result of correct pre-media settings. Why not an exact match between display and print? The reason is that color device is gamut-limited. In this case, monitor gamut is smaller than the working space, and printer gamut is smaller than monitor gamut. 1a. Predictability of pictorial color images from display to print Color predictability emphasizes the use of A-to-B-to-A color conversion to simulate printed color closely with a display. As shown in Figure 2, the source or simulation profile is the printer, and not the working space. Thus, color predictability is realized when (a) the display gamut is larger than that of the printer, and (b) the color rendering intent, used in the pre-media settings, is absolute colorimetric.
Submitted to the 35th Int l Research Conference, Sept. 7-10, 2008, Valencia, Spain 2 1b. Spot color portability Figure 2. Color predictability in color management workflow Spot colors are colors that designers choose from a color swatch library, e.g., Pantone solid coated library, when creating artwork using pre-media software. There are many custom color libraries and each library contains more than 1,000 spot colors. The fact that in-gamut spot colors can be accurately displayed on the monitor is because there is a spot color-to-pcs conversion available and the pre-media software acts as the API to complete the color conversion for monitor display (Figure 3). Figure 3. Spot color management workflow There are two ways to reproduce spot colors in hardcopy. The first method is to print spot color alone as specially formulated ink on a printing press (Figure 3). In this case, these are high chromatic colors, e.g., Coca-Cola red, IBM blue, etc. The second method is to print spot colors as CMYK composites on a CMYK digital printer. There is a RIP-based look-up table that performs spot color -to-cmyk conversion prior to hard copy output. Due to gamut limitation of the digital printer, spot color accuracy is compromised (Pantone, 2008). In this research, hard copies from spot color printing using specially formulated ink serve as the reference. 1c. Research questions on spot color predictability When a spot color is specified in pre-media software by default, colors underneath the spot color are removed or knocked out from the digital file as if the ink is opaque. Due to the advent of pre-media software development, the rule of "printing spot color alone as an opaque ink" has been changed. Today, pre-media software, such as Adobe Creative Suite, allows spot colors to overprint on top of other colors with different blending and transparency effects. As a result, spot color overprint features offer many pre-media color choices to designers. There are two research questions raised in this research. The first research question is, What is the predictability of spot color solid between display and print? The second research question is, What is the predictability of spot color overprint between display and print? If we can answer both questions with confidence, we will extend the understanding of color management from process color to spot color in the graphic arts industry. 2. Methodology This research takes the following four experimental steps: (a) building a viewing platform to enable displayto-print comparison; (b) validating the predictability of process color between display and print; (c) testing the predictability of spot color solid between display and print; and (d) testing the predictability of spot color overprint between display and print. To elaborate, the first step is to set up an experimental condition for display and print comparison. The second step is to validate the experimental condition by means of psychometric analysis. The last two steps are to answer the research questions regarding spot color predictability. 2a. Building a viewing platform to enable display-to-print comparison It is customary to view a monitor display under dim surround and a hard copy image under bright surround. ISO 12646 (2006) specifies that the ambient illumination of the monitor shall be low, i.e., less than 32 lux.
Submitted to the 35th Int l Research Conference, Sept. 7-10, 2008, Valencia, Spain 3 ISO 3664 (2006) specifies that the viewing illumination of the hard copy shall conform to ISO viewing condition P2 or 500 ±125 lux. In a pilot study, many observers found it difficult to evaluate display-to-print color match with the display under dim surround and print under bright surround. The inability of human visual system to fully adapt to both viewing conditions is a major cause of the difficulty. The work of Katoh (1995) elaborates on the phenomenon, known as partial adaptation. To overcome the difficulty of adaptation, this research modified a viewing booth to accommodate the displayto-print comparison in a common bright surround. As shown in Figure 4, a GraphicLite viewing station, manufactured by GTI Graphic Technology, Inc., was modified so that a hard copy (A) is placed next to a 20 Apple Cinema Display (B) surrounded by a gray mask (C). A second monitor, located outside the viewing booth, is used for logistic purposes. The 20 Apple Cinema Display is referred to as the soft proofing monitor in this research. Figure 4. Display-to-print comparison under bright surround To configure the viewing condition and pre-media settings between display and print under bright surround, the following steps were carried out in this research: 1. A digital printer was profiled using the IT8.7/4 profiling target and X-Rite ProfileMaker 5. A grayscale print, as shown in Figure 4, was made as a reference. 2. The soft proofing monitor was calibrated and profiled under a number of white point and gamma combinations using the X-Rite Eye-One Pro. In a preliminary test, the grayscale print was compared to a number of displays of the grayscale PDF file using different monitor profiles. The monitor profile with D50 and 2.0 gamma was found to yield the best display-to-print match. 3. Both the monitor profile and the printer profile were placed in the ColorSync folder. 4. In the pre-media color settings dialog box, (a) the soft proofing monitor ICC profile acts as the RGB working space; (b) the digital printer ICC profile acts as the CMYK working space; (c) the color rendering intent was set to absolute colorimetric. This color setting was synchronized in all Adobe CS components using Adobe Bridge. 5. The grayscale PDF file was opened in Adobe Acrobat Pro and displayed on the soft proofing monitor. Go to Advanced / Print Production / Output preview and verified that (a) the correct printer ICC profile was chosen under Simulation Profile, and (b) the Simulate Paper Color option was checked. 6. The observer was asked to adjust the intensity of the ambient light in the viewing booth so that the white point and intensity of the hard copy matched that of the monitor. 2b. Validating the predictability of process color between display and print The strategy for validating the experimental viewing setup is to test the display-to-print agreement using printed process (CMYK) color as the reference and multiple display conditions, as permitted by the pre-media software, as samples. Colorimetric coordinates of printed process colors, e.g., magenta and yellow, are inside the monitor color space. The display-to-print agreement is determined by paired comparison, a psychometric analysis technique (Rickmers, 1973). Below is a description of the validation process with an illustration (Figure 5) to show how the experimenter interacts with a print reference and two display samples in the modified viewing booth.
Submitted to the 35th Int l Research Conference, Sept. 7-10, 2008, Valencia, Spain 4 Figure 5. Paired comparison between a print reference and two display samples 1. Place a print reference (left-hand-side) in the viewing booth. 2. Prepare the digital file in four PDF Output Preview conditions, i.e., one with correct and the other three with incorrect Proofing Simulation profiles, as shown in Table 1. 3. Display two of the four proofing simulations in the proofing monitor (middle and right). 4. A total of 10 observers are asked to select one display that matches closer to the print reference. 5. There are six possible display pairs with four display settings. The paired comparison procedure is repeated for all possible combinations of display pairs in random sequence. 6. The above steps are repeated for a total of four references (or scenarios), i.e., magenta solid, yellow solid, M+Y (50%) overprint, and M+Y (100%) overprint, as shown in Table 1. Table 1. Sample preparation of process color Process Colors Display Samples Magenta Yellow M+Y - 50 % tint FOGRA 28 ICC profile Japan Newspaper ICC profile Eps_4k SWOPv2_ Siml ICC profile SWOP v2 ICC profile M+Y - 100 % tint Among the four display samples, the SWOP simulated Epson SP 4000 profile represents the correct simulation profile. In three out of four paired comparison scenarios, the results showed that the best simulation profile for display-to-print match is the SWOP simulated Epson Stylus Pro 4000 profile. The only exception is the M+Y (red) overprint where the SWOP v2 simulation profile was judged to provide the best display-to-print match. Upon colorimetric analysis, the colorimetric difference between the SWOP simulated Epson SP 4000 profile and the SWOP v2 profile is small. There are two lessons learned: (a) experimental errors occur when there is no real difference between the correct and the incorrect pre-media settings, and (b) such an error could be avoided by selecting proofing simulation profiles that are sufficiently different to begin with. 2c. Testing the predictability of spot color solid between display and print Spot colors are any number of specially formulated inks. Pantone color library provides color swatch books along with a unique number and a recipe. For example, Pantone 1788C is a bright red color and is mixed by 14 parts of Pantone Warm Red and 2 parts of Pantone Rubine Red. In this research, we created a test target consisting two spot color solids, their 50% tints as well as their overprints of 50% tints and solids. Two pairs of spot colors were selected in the experiment. The first pair, Pantone 1788C (red) and Pantone 7466C (turquoise), and their overprints are shown in the Adobe Illustrator CS3 (Figure 6). The second pair of spot colors was Pantone 493C (pink) and Pantone 577C (green).
Submitted to the 35th Int l Research Conference, Sept. 7-10, 2008, Valencia, Spain 5 Figure 6. Spot color solids and overprints These spot colors were printed in two different ink sequences using a Heidelberg Speedmaster 74 sheet-fed offset press and certified Pantone inks. Printed hardcopies then serve as references in the subsequent crossmedia color matching experiment. The procedure below only describes how predictability of spot color solid between display and print is tested. To test the predictability of spot color solid between display and print, four spot colors were created in the Adobe Illustrator (Table 2). For each spot color selected, four different input (or spot color-to-pcs) conditions are defined, i.e., (a) correct PMS number and correct substrate, (b) incorrect (adjacent) PMS number and correct substrate, (c) incorrect PMS number and incorrect substrate, and (d) measured CIELAB values from the print reference. Table 2. Sample preparation of spot color solids Spot Color Solids Display Samples High Chroma 1 (7466 C) High Chroma 2 (1788 C) Moderate Chroma 3 (577C) Correct PMS Color and Correct Substrate Hardcopy CIELAB Color Incorrect PMS Color and Correct Substrate Incorrect PMS Color and InCorrect Substrate Moderate Chroma 4 (493 C) The paired comparison procedure, as mentioned in Section 2b, is used to find out if a particular monitor display condition provides the best display-to-print color match. To hypothesize, the predictability of spot color solid should hinge on the pre-media settings for the correct PMS number on correct substrate yielding the best display-to-print match. 2d. Testing the predictability of spot color overprints between display and print To test the predictability of spot color overprint between display and print, the two high-chroma spot colors (Pantone 1788C and Pantone 7466C) were overprinted in two ink sequences; and the two moderate-chroma colors (Pantone 493C and Pantone 577C) were overprinted in two ink sequences. There are five pre-media settings to display overprint colors, i.e., (a) Overprint Fill, (b) Overprint Fill with Opacity, (c) CIELAB measurement from hard copy reference, (d) Multiply blending mode, and (e) Photoshop Duotone. Only four display samples were chosen from the five input or display conditions for each of the overprint references, as shown in Table 3. Specifically, Overprint Fill is the official overprint setting given by the Adobe Creative Suite software for any two overlapping color objects or layers. The rest of the overprint settings are chosen from their simulation possibilities, such as the Multiply blending mode, Overprint Fill with Opacity to account for ink trapping, and how duotone colors are defined in Photoshop.
Submitted to the 35th Int l Research Conference, Sept. 7-10, 2008, Valencia, Spain 6 Tint of Overprint Solid Overprint 50% Tint Overprint Table 3. Pre-media sample preparation of spot color overprints Type of Spot Color High chroma Moderate chroma High chroma Moderate chroma Overprint Fill Overprint Fill with Opacity Illustrator LAB from Hardcopy Multiply blend mode Photoshop Duotone x x x x x x x x x x x x x x x x The paired comparison procedure, as mentioned in Section 2b, is used to find out if a particular monitor display condition provides the best display-to-print color match. To hypothesize, the predictability of spot color overprint should hinge on whether any of the four pre-media settings yields the best display-to-print match. If, however, the pre-media setting from the measured CIELAB values is selected as the best displayto-print match, it is an indication that there is no mechanism in today s pre-media software to predict spot color overprint accurately. 3. Results This research builds on the understanding of color management in pictorial color image reproduction and extends the understanding to including color management of spot colors and their overprints. A modified viewing condition was constructed to accommodate display and print comparison under bright surround. Display-to-print color match in the modified viewing condition was validated by means of paired comparison with the use of printed process color and the correct pre-media display settings. The same display-to-print viewing condition and paired comparison technique were used to test the predictability of spot colors and their overprints. 3a. Predictability of individual spot color solids A total of four spot color solids were tested for predictability. Two spot colors, 1788C (red) and 7466C (turquoise), are high chromatic colors; the other two, 493C (pink) and 577C (green), are moderate chromatic colors. Each spot color was configured with four pre-media settings. By means of paired comparison, two of the four display settings were shown to the judge at a time. The judge was asked to pick the display that matches the reference print. If every judge picks a particular display every time, the sample will receive a maximum of 4 points. On the other hand, if no judge picks a particular display, the sample will receive a minimum of one point. Table 4 shows the average scores from paired comparison testing. Table 4. Pre-media settings of individual spot color solids and paired comparison scores Color Definition Correct PMS Color and Correct Substrate CIELAB measure from hardcopy reference Incorrect PMS Color and Correct Substrate Correct PMS Color and Incorrect Substrate Average Scores for Spot Color Solids Sample 493 C 577 C 1788 C 7466 C A 2.8 1.1 3.6 2.0 B 1.2 3.1 2.9 1.1 C 2.3 2.9 1.0 3.4 D 3.7 2.9 2.6 3.4 The paired comparison scores indicate that there was no particular pre-media setting that would yield the best display-to-print match. The correctly named spot color that provides the best display-to-print match only occurred once, i.e., Pantone 1788C. Reasons that contribute to different display-to-print match outcome include (a) the similarity between the printed spot color reference and the colorimetric definition of the digital color swatch, (b) the similarity between the correct (coated) substrate and incorrect (matt) substrate, and (c) the similarity between the colorimetric definition of the correct color and the measured CIELAB values from the spot color print. 3b. Predictability of spot color overprints Four spot color overprints were also tested for color predictability, i.e., overprint solid and overprint tint at 50% tint of the Pantone 1788C-7466C pair and the Pantone 493C-577C pair. Table 5 summarizes the overprint settings and paired comparison results for the Pantone 1788C-7466C color pair only. Table 5. Pre-media overprint solid settings and paired comparison scores for Pantone 1788C-7466C
Submitted to the 35th Int l Research Conference, Sept. 7-10, 2008, Valencia, Spain 7 Overprint Definition T+R (100%) Sample Average Score Oveprint Fill A 1.7 Photoshop Duotone B 1.9 CIELAB measure from hardcopy reference Oveprint Fill with Opacity C 3.7 D 2.7 The paired comparison scores in Table 5 indicate that none of the pre-media settings yields the best displayto-print match of overprint solids. The display sample using the measured CIELAB values provides the best display-to-print color match. Paired comparison scores from the other three spot color overprints indicate the same outcome as the Pantone 1788C-7466C scenario. This is indicative that there is a gap in pre-media software s ability to predict solid color overprint accurately. When conducting the spot color press run, CIELAB values of the resulting spot color overprint solid vary as a function of ink sequence used. As shown in Figure 7, C1 and C2 represent the Pantone 1788C (red) and Pantone 7466C (turquoise) respectively. The spot color overprint solid, C3, is printed with the Pantone 1788C (red) on top of the Pantone 7466C (turquoise). And the spot color overprint solid, C4, is printed the other way around. There are three important findings: (a) the color difference between C3 and C4 is quite large (23 E ab ) and visible; (b) the hue of the overprint, C3, with the Pantone 1788C (red) ink overprint on top of the Pantone 7466C (turquoise) ink, looks bluer; and (c) the hue of the overprint, C4, with the Pantone 7466C (turquoise) ink printing on top of the Pantone 1788C (red) ink, looks redder. The last two findings are rather counter-intuitive, i.e., the last-down ink has less influence on the hue of its overprint. Ink trapping is suspected to be the cause. Figure 7. Colorimetric plots of Pantone 1788C, Pantone 7466C, and their overprints The color difference of spot color overprints due to ink sequences decreases when either moderate-chroma spot color solids are printed (3 E ab ) or when tints of high-chroma spot colors are printed (3.8 E ab ). Nevertheless, ink and press related factors, such the effect of ink sequence on spot color overprint, have not been taken into consideration in current pre-media software. This has limited the ability of pre-media software to predict spot color overprint accurately. 4. Conclusions This research places the predictability aspect of the color management under scrutiny. The display-to-print match succeeded for process color applications. The combination of the monitor profile, the CMYK printer profile, a well-defined viewing condition, and the correct pre-media color settings make display-to-print comparison under a common bright surround valid. The display-to-print match failed completely for spot color overprint applications. Spot color overprints, defined by CIELAB measurement of the hardcopy reference, is chosen as the best match in all four scenarios. None of the pre-media settings renders spot color overprints accurately. Overprint is defined as a Boolean flag in the third edition of the PostScript Language Manual (Adobe, 1999). Page 248 of the Manual states, If the overprint flag is false (default), painting a color in any color space
Submitted to the 35th Int l Research Conference, Sept. 7-10, 2008, Valencia, Spain 8 causes the corresponding areas of unspecified colorants to be erased. If the overprint flag is true and the output device supports overprinting, anything previously painted in other colorants is left undisturbed. The effect produced by such overprinting is not defined by the PostScript language. Adobe Acrobat Reader may be used, via Advanced / Print Production / Overprint Preview, to display the effect of the Overprint Fill of spot colors, but the resulting display is incorrect because there is no device-to-pcs (A-to-B) color conversion for spot color overprints. Soft proofing performance ought to be judged visually. Paired comparison analysis techniques are effective in finding the best visual agreement between print and the display representing correct pre-media settings. The choice of spot color and pre-media settings can influence the outcome of paired comparison. For examples, a spot color is outside of the display gamut; colorimetric values of multiple pre-media color settings are too close to one another. To improve the predictability of spot colors and their overprints in a pre-media environment, device-to-pcs color conversion must exist to include ink- and print-related parameters, e.g., paper color, ink color, ink transparency, ink tack, ink-down sequence, and ink trapping. Since there are more than 1,000 spot colors in the spot color library, it will be difficult, if not impossible, to obtain colorimetric values of all combinations of two-color overprints. While there is a void in the prediction of spot color overprints in current pre-media software package, we encountered a few commercial software packages, e.g., X-Rite ProfileMaker 5 MultiColor Package, AVA CAD/CAM, Artworks Display Ink, that provide solutions for a limited number of spot colors. In general, predictability of spot color overprints with accuracy is likely to remain as a void in shrink-wrapped premedia software for some time to come. Acknowledgments The authors wish to thank Mr. Franz Sigg, RIT School of Print Media, for his effort in modifying the gti viewing booth for display-to-print comparison under bright surround, and the RIT Printing Applications Laboratory for implementing the spot color press run. Literature Cited Adobe Systems Incorporated (1999) PostScript Language Reference Manual (3rd Edition). Reading, MA: Addison-Wesley, ISBN 0-201-37922-8. ISO/DIS 12646 (2006) Graphic technology Displays for colour proofing Characteristics and viewing conditions ISO/WD 3664 (2006) Viewing conditions Graphic technology and photography Katoh, Naoya (1995), Appearance Match between Soft Copy and Hard Copy under Mixed Chromatic Adaptation, 3rd IS&T/SID Color Imaging Conference, Scottsdale, Arizona, p. 22-25, ISBN 0-89208- 188-0. Pantone (n.d.) Pantone Spot Color vs. Process Color. http://www.colorguides.net/pantone_spot_color.html (retrieved on June 18, 2008) Rickmers, A. (1973) Paired Comparison, Course Notes on Statistics, Rochester Institute of Technology, Rochester, NY