BUILDING COLOR MANAGEMENT MODULES USING LINEAR OPTIMIZATION II - PREPRESS SYSTEM FOR OFFSET PRINTING

A spectral model was derived to predict the spectral reflectance facto r of colors formed using a color proofing system simulating offset pri nting. A first-order model was based on the spectral Neugebauer equati on modified by the Yule-Nielsen correction in which n was assumed to V ary as a function of wavelength. The n(lambda) and effective dot areas were optimized using primary (cyan, magenta, yellow, and black) halft one tints. Systematic errors were observed. The systematic error behav ed in a similar fashion to the phenomenon of ink trapping. Because ink trapping, ink spread, ink mixture and variance of mechanical dot gain were negligible for this proofing system, this is an optical effect t o be referred to as ''optical trapping.'' An interaction model was der ived that compensated for optical trapping. Adding the optical trappin g effect to the first-order model significantly improved model predict ion to an average Delta E-ab(), of 2.2 with a maximum of 5.5. A simpl e black printer model was derived for an inversion of the forward mode l that aimed to provide a similar black amount with a conventional col or-separation method and colorimetric match applying a concept of unde r-color removal (UCR) in original density space with tone reproduction curves of a gray scale. Using the Simplex method, the modified spectr al Neugebauer model with the black printer model was inverted to build a backward model. Influences of the dot area transform function obtai ned from the backward model were compared with those from a convention al method for an evaluation of similarity. A desktop drum scanner was colorimetrically characterized using a spectral reconstruction model f or a renective photographic material to build a concatenated device pr ofile in which digital counts of a scanned photographic reflection pri nt were the input and those of the proofing system were the output. Pe rformances of the concatenated device profile were evaluated for pract ical use. The average Delta E-ab(), error from the profile was 2.1 in cluding colors outside of the proofing system's color gamut.