............................................................................................................................................................... iv Introduction.......................................................................................................................................................... iv 1 Scope ...............................................................................................................................................................1 2 Theory..............................................................................................................................................................1 3 Experiments .....................................................................................................................................................3 4 Results .............................................................................................................................................................4 5 Discussion........................................................................................................................................................5 6 Conclusions .....................................................................................................................................................7 Figures 1 Spectral transmittance of a red specimen (solid curve) ...................................................................................8 2 Extension of Figure 1, showing first and second derivatives ..........................................................................9 3 Schematic illustration of an ideal linear reflectometer versus two real instrument curves............................10 4 Test sample set of 420 specimens displayed on CIELAB a*, b* plane.........................................................11 5 Test sample set of 420 specimens displayed on CIELAB a*, L* plane ........................................................12 6 Test sample set of 420 specimens displayed on CIELAB b*, L* plane ........................................................13 7 Test samples of 718 ink-on-film specimens displayed on CIE a*, b* plane .................................................14 8 Test samples of 718 ink-on-film specimens displayed on CIE C*, L* plane................................................15 9 Plot of model parameter B3 versus wavelength showing the linear change in .............................................16 bandwidth for the ChromaSensor CS-5 correlated to a Spectraflash 500 10 Histogram of the 420 color differences between the ChromaSensor CS-3 ...................................................17 and the Spectraflash 500 before application of the correlation model 11 Histogram of the 420 color differences between the ChromaSensor CS-3 ...................................................18 and the Spectraflash 500 after application of the correlation model 12 Histogram of the 718 color differences between the 0E:45E and the di:8E ...................................................19 before application of the correlation model 13 Histogram of the 718 color differences between the 0E:45E and the di:8E ...................................................20 after application of the correlation model Tables 1 Regressing coefficients for CS-5 regressed onto Spectraflash 500, specular component included...............21 2 Model fitting results.......................................................................................................................................22 3 Statistical analysis of the histograms shown in Figures 9 and 10..................................................................24 4 Results of the average inter-instrument model in terms of CIELAB )E test instruments ............................24 versus Spectraflash 500, with the specular component included for the 420 total samples References ............................................................................................................................................................25 Biographical Information on the Author ..........................................................................................................27 Improving the inter-instrument agreement of spectrocolorimeters iv © NPES 2004 – All rights reserved Abstract The first edition of ANSI CGATS.5 contained an appendix that discussed a method for improving the agreement between two instruments. The technology described was taken from a paper given at the first Oxford Conference, in which Alan Robertson gave a presentation in which he proposed using spectral differences to predict systematic errors in spectrophotometric measurements. Subsequent to that, Berns and Petersen published a paper in Color Research & Application that modified the Robertson method slightly by using numerical derivatives and derived a model for improving inter-instrument agreement. That model has found limited success in practical application. This technical report reviews the progress that has been made since the original release of CGATS.5-1993 and demonstrates that improvements in inter-instrument agreement are now possible at a level that equals or exceeds that of national standardizing laboratories. In addition, it will be shown that it may also be possible to bring together the readings taken with diffuse illumination and directional viewing with those collected with directional illumination and directional viewing measurement geometries.