Readme.dat for Colour Proof

What follows is a sampling of results from North America's most recent, independent evaluation of the colour proofing market. The Association of Graphic Solution Providers (IPA) charged five technical experts with investigating six categories vital to any successful proofing integration. With an advance copy of the report PrintAction provides a window into one of the industry's most technologically challenging issues, but it is recommended the full report be purchased to truly understand the technical parameters and vendor concerns surrounding these results.

The IPA roundup team looked at colour-image quality by visual match; colour-image quality by colourimetric measurement; spot-colour support; RIP performance in terms of speed, overprint and colour management; PDF/X ability; and finally the team compared features of the systems in terms of system price and per proof consumable costs. Those readers more proficient in the technical aspects of colour proofing, or for those who want to perform their own tests, should refer to the sidebar IPA Guidelines - again, a complete explanation can be found in the report.

Many of the results are described in terms of DeltaE (DE) which, when loosely defined, is used as a standard to measure the difference between two colours where a DE value of zero represents a perfect match. The greater the DeltaE value, the poorer the match. As a general rule, a DE greater than 1.0 is said to represent a noticeable visual difference in the two colours being compared - in this case, within an image. Therefore, a colour proofing system would ideally register results somewhere between DE 0 and DE 1.0.

Colourimetric

Clearly, the best systems should exhibit a low mean DE and also a low DE maximum (Figure 1 shows results for each supplier, excluding soft-proofing systems for the colourimetric test). The average DE of all the suppliers was 2.67. It must be noted that the judges estimated that their results had an overall measurement uncertainty of DE 1.0 for the reported average scores and DE 3.0 for the reported maximum scores because of potential issues arising from repeatability, inter-instrument, UV influence, backing, IT8 layout and colour stability. While only two systems, GMG's ColorProof (Epson 7600) and CGS' Oris Color Tuner, rated below the ideal target of DE 1.0, the top ten systems fall within the uncertainty interval of this ideal target.

The testing revealed which vendor systems faced colour issues, including yellow solids for Dupont (CromaPro/Epson 10600), yellow overprints for KPG (Matchprint Digital/Creo Spectrum), magenta and cyan solids for Fujifilm (Proof Director/FinalProof), yellow overprints for Blanchard (Vortex/Epson 7600), magenta overprints for Epson (Chromaticity/ColorBurst/Epson4000), cyan and cyan overprint for PerfectProof (ProofMaster Folio/Canon W2200), solid ink highlights for EFI (Colorpoof /Canon 8200), some overprints for EFI (Colorpoof /HP Designjet 130), and shadows for KPG (Matchprint Professional/CLC 1180).

Visual

For the visual component of the test, nine additional viewers experienced in colour evaluation helped to come up with the final results (Figure 2). The judges were asked to evaluate the proof using seven criterions to arrive at a final score based on a scale of 1 to 10, with 1 representing serious imbalances of colour and 10 being a virtually perfect proof. The criteria consisted of gray balance, saturation, contrast and weight, flesh tones, accuracy of vignettes and visible artifacts.

The judges expected that some systems would excel in particular areas, such as gray balance, but fare poorly in others, such as flesh tones. While this did happen in some cases, the results showed a distinct pattern for all seven categories of colour judging to cluster closely along a common trend. The expert team laid this out in a chart (not shown here) to arrive at this conclusion.

Interestingly, Oris Color Tuner found itself on the wrong end of this visual test despite strong colourimetric results. In fact, the IPA team made an attempt to see if any correlation could be found between visual score and Colourimetric rating. They concluded that the visual results correlated moderately, but not universally, with the Colourimetric results. Indeed looking back at Figure 1 it is clearly that that ratings for DE max did not always correlate with that of DE average. Since DeltaE is a measure of error, an inverse relation would be expected between DE and visual scores. In other words, low DE should correlate with high visual scoring and high DE should correlate with low visual scoring, which can be determined by comparing Figure 1 and Figure 2.

Spot

Of the colours it chose to test, Purple 526 showed the best DE at 4.27 and orange proved to be the most problematic with Orange 151 having a DE of 16.99 and Orange Hex a rating average DE of 21.04. This shows the Colourimetric challenge of various spot colours that, for the most part, are not within the gamut of typical proofing systems.