Determination of the Whiteness of Paper Substrates Containing Fluorescent Whitening Agents Under CIE D50

Danny C. Rich, Veronika Lovell, and Robert Marcus, Sun Chemical Corporation

Changes to the ISO 3664 standard for the visual and instrumental assessment of print on paper substrates has been causing problems for converters who try to use these standards to assess the quality of print relative to a digital proof.  One possible cause was thought to be a result of the use of CIE D65 illuminant and CIE 10° observer data by the paper industry and the use of CIE D50 illuminant and CIE 2° observer data by the graphics industry.   This study investigated whether the CIE metrics for paper appearance and whiteness are applicable to the substrates used in modern graphic reproduction.  Both visual and instrumental data given in this report indicate that the correlation between the two measurement conditions and the CIE recommendations are moderate to poor and that there needs to be a new index developed for use in grading and specifying production stocks that will match proofing stocks.  It is shown that differences between the print and proof may not be due to the lighting.

A set of 30 white paper specimens were selected, 15 with some level of fluorescent OBA added and 15 without any fluorescent additives. A visual experiment was designed in which observers judged the level of whiteness in a rank order experiment, thus developing the basis for assessing the metric of whiteness.  The specimens were presented to twenty observers and the rank orders of each observer were compared and then averaged.  The agreement between the scale of the observers and the CIE Whiteness index, computed as per ASTM E313 was 0.87 for the papers without OBA and 0.82 for the papers with OBA.  In contrast, the ISO Brightness index gave correlations of 0.93 for the non-OBA containing substrates and 0.87 for the OBA containing substrates. The Ganz equations, adopted by the CIE and recommended in ASTM E313, contain 5 adjustable parameters.  A modern spectrocolorimeter with an ISO 13655 conforming source was used to collect D50 measurement data on the 30 specimens.  The measurements were submitted to a multiple linear regression, just as Ganz had performed, to develop new estimates for the coefficients for CIE D50 and the 1931 CIE standard observer rather than CIE D65 and the 1964 CIE supplementary observer.  The two data sets were analyzed separately and in combination.  When analyzed separately, both data sets, those with fluorescence and those without fluorescence, produced the same set of whiteness index coefficients.  The resulting equations reported correlations of 0.96 for the non-fluorescent papers and 0.94 for the papers with the fluorescent additive.  It can be concluded that the CIE equations do not produce the best agreement to visual judgments under CIE D50 as required by ISO compliant graphic reproduction systems.