Report on generating a colour circle for testing in screenprinting and inkjet

Transcription

1 Report on generating a colour circle for testing in screenprinting and inkjet In order to test the colour capabilities of a range of inkjet devices and to compare them with screenprint hues, a colour circle was generated in Photoshop. Hue Saturation Brightness (HSB) colour percentages were modified, and working in a CIE L*a*b* colour space; the a* component (green-red axis) and the b* component (blue-yellow axis). HSB was useful in this instance as a means of generating a 36 colour circle: by mathematically dividing the circle into equal segments and modifying the hues using saturation and lightness to obtain black and white. Two colour circles were generated (figure 6): twenty-four segments of equal degrees form the outer ring of the colour wheel. Two colour wheels were devised: 1% hue colour to white and 1% hue colour towards black. Fig.6 Two colour circles generated in Hue Saturation and Brightness. For each circle, the segments were divided into 1% increments, the first working from 1% (saturation and brightness) using the saturation slider to add white by 1% increments to 1% near white; then from 1% (saturation and brightness), using the brightness slider to make of 1% increments to near black (figure 7). 1

2 Fig. 7 HSB slider in Photoshop Inkjet printing the colour circles The inkjet tests were undertaken using the pigment based HP Z31 1 and the dye based HP 13nr 2 inkjet printers. As the Z31 printer also had an extended inkset, which included red, green, blue, and grey, it was useful to compare the printable gamut between these two machines. The circles were printed onto A3 sized sheets onto Hahnemühle PhotoRag fine art paper. The circles were printed from the LAB colour mode, using application managed colour in Photoshop, and the settings best and heavyweight coated fine art paper, were used in the printer driver. Based on initial trials during the Digital Portfolio, we found that just by changing the output profile, resulted in very different colour prints. Therefore, I was also interested in investigating further how the colour information was managed and sent to the printer and, working on the Z31, two output profiles were used and compared. In the first case, it was decided to keep the same application managed colour, and in the second a paper profile was selected to match the heavyweight coated fine art paper in the printer driver. 1 The inkset in the HP Z31 contains: cyan, magenta, yellow, matte black, light magenta, red, green, blue, light grey, medium grey, (photo black, gloss these would not be used in this print mode). 2 The inkset in the HP 31nr contains: cyan, magenta, yellow, black, light cyan, light magenta 2

3 4.2.2 Preparing and making the colour separations for screenprint Unmixed hue screenprint inks of: process cyan, process magenta, process yellow, cadmium yellow, cadmium red, cadmium scarlet, ultramarine, leaf green, phthalocyanine green, black, purple (figure 8) were screenprinted onto a bright white card. Once dry they were measured using a Gretag Macbeth spectrophotometer, using an illumination D5 and at 2. Both L*a*b* values and spectral data were recorded. The measurements of each hue can be used to compare colour differences, create colour profiles and generate multi-colour separations in preparation for printing. Fig. 8 Primary hue inks screenprinted onto white card. Top left: cyan, ultramarine, cadmium red, process magenta, process yellow, purple. Bottom left: phthalocyanine green, black, cadmium scarlet, cadmium yellow, leaf green. The Eye-One Share software 3, which is the software interface for the spectrophotometer, enabled colour patches to be visualised in relation to other colours. The measured data was compared and exported, from which colour palettes could be generated, or as spectral charts as demonstrated below (figure 9). 3 The Eye One spectrophotometer and visualizing software from GretagMacbeth is a hand held measuring device that can measure spectral information from emissive sources such as a CRT/LCD display or reflective sources, such as print, paint or fabric samples 3

4 process cyan ultramarine cadmium yellow cadmium red phthalo green leaf green process yellow process black cadmium scarlet process magenta purple.1 38nm 4nm 42nm 44nm 46nm 48nm 5nm 52nm 54nm 56nm 58nm 6nm 62nm 64nm 66nm 68nm 7nm 72nm Fig. 9 The spectral data of the screenprinted colours Fig. 1 L*a*b* colour measurements, from the screenprinted ink patches, are added to generate a custom CMYK profile. Following Lammen s detailed technical notes on how to create profiles in Photoshop 4, the measured screenprint hues were added to Ink Colors (figure 4 Based on laboratory notes written by Johan Lammens and Pau Soler of HP Labs, Barcelona, 6 th March 1999, Quick ICC profiling with Photoshop 5 4

5 1) in Photoshop, and a new custom CMYK profile was generated. The following settings were used: Open File Colour Settings CMYK Setup select Built-in CMYK model Select Ink Colors Custom, check the LAB coordinates box, uncheck Estimate Overprints box and input the 9 values in LAB (Figure 2). For the section on dot gain, Lammens suggests 15% for glossy papers and 2% for coated papers. For separation options he suggests Separation type: GCR; Black Generation: Medium; Black Ink Limit: 1%; Total ink Limit 24% for glossy bi-level, 2% for native coated bilevel 4% if the ink limits are applied through the printer; UCA amount ; save as xxxx s new profile. This profile was named Screenprint Colours. This profile could then be used to preview how the colour circle would appear when screenprinted and the potential limits in the screenprint colour gamut. In order to make a multi-colour stochastic separation suitable for screenprinting, the LAB colour file had to be converted to a CMYK file. It was found that when converting from LAB to a generic CMYK space, there was a significant reduction of the colour gamut resulting in a de-saturation in colour. Two alternative CMYK profiles were tried, which included the new Screenprint Colours profile and a commercial HP profile for a fine art coated paper. On initial inspection from the monitor screen, the HP profile demonstrated the least colour difference and better information in the green areas, whilst the Screenprint Colours profile, which was based on just nine measured hues, offered a good alternative, especially in the reds, magenta and orange region. This simple exercise demonstrated that there was not one simple approach to colour conversion. The Screenprint Colours profile was used to convert the colours from LAB to CMYK mode. PhotoSpot, by Second Glance Software Inc., a colour separation plug-in to Photoshop, is a useful software for working with bespoke colours and separating the image according to these colours. It was important to bear in mind that the conversion from an RGB or LAB colour mode to CMYK mode would result in a reduced coloured image. Therefore it was useful to convert the image using the custom Screenprint Colours 5

6 profile. Using PhotoSpot, two different colour separations were generated, along with an eight-colour and a nine-colour separation. The colours used for the two screenprint versions were: Screenprint (1): cyan, magenta, yellow, cadmium red, cadmium scarlet, ultramarine, leaf green, black. Screenprint (2): cyan, magenta, yellow, cadmium scarlet, ultramarine, leaf green, phthalo green, black. Fig. 11 The original HSB circles in LAB mode, in comparison to the same circles converted to a generic CMYK colour mode. The circles on the left are the unconverted colours in photoshop and the circles on the right demonstrate how the colour values appear when converted to CMYK colour mode. 6

7 Fig. 12 These two circles are converted using different CMYK profiles, the first is based on a coated fine art paper, which offers the largest CMYK gamut and compared to the 11 measured values from the screenprint hues Results The following figures (14a-e) illustrate measured L*a*b* data from the twenty-four hue patches. These appear on the outside perimeter of each colour circle. Figure 13 shows a colour circle that has been rotated so that the yellow appears at the top and has been flipped horizontally. This is so that a direct comparison can be made with the measured data, the graphs, and the standard CIE L*a*b* colour circle: yellow at the top, blue at the bottom, red to the right and green to the left. 7

8 Fig. 13 Colour circle rotated and flipped to the a* b* axis of the measured charts. Fig. 14a Screenprint hues Fig.14b HP Designjet 13nr. 12 Fig. 14c Screenprinted colours from colour separation Fig. 14d Z31 using an output paper profile. Fig. 14e Z31 output same as source. 8

9 yellow Screen-ink hues green red Screenprint with colour separation HP 31 HP Z31 HP Z31 with profile -8-1 blue Fig. 15 Comparing two profiles that were used to convert to a CMYK colour mode. Note the loss of saturation in the green section of the circle on the right. By over-layering each plot in figures through 14a to 14e, a comparison can be made, demonstrating a larger gamut of the screen ink hues, as represented by the turquoise line (figure 15). The curve of the turquoise line around the bottom of the circle demonstrates more blue and purplish blues in the screen ink gamut, with the Z31 showing more of an inclination towards the greens, due to the inclusion of the green cartridge Conclusion There was a significant difference between the two circles printed on the Z31 using different output profiles. Even though the source image was the same, by changing the output profile, one could dramatically change the outcomes of the printed colour. Both circles utilised the application driven colour setting by driving colour data from the application instead of using the printer default software. The addition of the extra red and green inks in the Z31 demonstrated an advantage over the 13nr six-colour printer and a larger printable colour gamut. 9

10 The measured screenprinted hues (figure 15) showed the largest gamut of all, especially in the blue region. Even though the same screen inks were used to print the circle, if compared to the screen ink hue colour patches, the screenprinted circle resulted in a more reduced colour gamut. As demonstrated in the figures 11 and 12, by converting from LAB to CMYK, there was a noticeable colour reduction in different sections of the colour wheel, especially in the green areas. This could be improved by using a larger gamut CMYK conversion space, such as a custom designed paper profile. The findings from the experiments demonstrated that the hue screenprint inks, when measured independently, revealed a much larger gamut than the same inks used for screenprinting the colour separations. However, during the process of colour conversion the potential of the colours was much reduced. The same problem could be hypothesised for the inkjet colours that, at the point of converting from one colour mode to another, or from image file to printer, there is an incremental loss of colour information. If for example, a more direct approach to digital colour mixing that was less reliant on colour management could be found, there could there be a greater control over each of the mixing and printing of these colours. 1