The article investigates five types of parametric effects in assessing colo
ur difference. The visual assessment experiments were conducted in six phas
es. The viewing parameters studied include the sample separation, sample si
ze, and the background colours. The dataset prepared in this study, include
s a total of 107 cotton-dyed pairs belonging to five colour centers. The av
erage colour difference of the sample pairs was about 5.0 CIELAB units. Gra
y-scale method was used fbr colour-difference assessment with a panel of 10
observers (2 assessments each). The CIELAB colour-difference formula was f
ound to have the best performance over other formulae using the current dat
aset. In general, the parametric effects vary from 8-15%. When,further anal
yzing the colour centers, it was found that much larger parametric effects
of 42% and 26% were associated, respectively, with green sample pairs asses
sed on the green background and blue samples pairs assessed on the blue bac
kground. The green and blue background colours in these two cases are very
close to the green and blue colour centers selected, which clearly indicate
s the crispening effect. (C) 2001 John Wiley & Sons, Inc.