Abstract

This study investigates methods to improve perceptual color matching between CCFL-backlight and LED-backlight displays evaluated against color-managed prints in professional soft-proofing workflows, through two sequential experiments. In the first experiment, a psychophysical paired-comparison study involving seven expert observers and 11 test images compared each backlight technology, hardware-calibrated to identical colorimetric targets, against reference prints on color similarity, saturation appearance, and overall preference. Results showed the CCFL display was consistently rated as perceptually closer to the prints and preferred over the LED display, despite the two displays being colorimetrically near-identical, indicating that the discrepancy originates from spectral rather than colorimetric differences. Building on this finding, the second experiment derived observer correction matrices for the LED display from a large-scale psychophysical matching study involving 45 observers, who interactively adjusted seven color patches to visually match printed references. K-means clustering reduced the 45 individual matrices into three representative observer classes. In a forced-choice validation, individual observer-specific matrices and the Cluster 1 representative matrix consistently outperformed the CIE 1931 2° Standard Observer baseline. Quantitative validation against ISO 14861:2015 criteria corroborated these results, with the corrected LED profile achieving statistically significant reductions in average and maximum ΔE00 and meeting ISO compliance thresholds. Together, these findings demonstrate that perceptually significant differences exist between CCFL and LED backlight displays even under identical colorimetric calibration, and that embedding observer correction matrices within standard ICC v4 profile architecture offers an implementable solution for improving LED soft-proofing fidelity without requiring migration to iccMAX.