MacDonald & Morovic (1995)

• paper cast removal - a*b* of substrate are subtracted from all colours.
• lightness mapping - using a linear technique.
• gamut clipping - two approaches were tried: orthogonal (Figure 1a) and cusp clipping (referred to as chord clipping in the paper) (Figure 1b).
• gamut compression - towards L*=50 on lightness axis (i.e. similar to Laihanenís second method (1987)), whereby the compression was the same across the whole space (in this case 20 per cent) rather than being determined along individual lines of compression.

The results of psychophysical evaluation suggested that a different kind of GMA was preferred for each painting whereby the preference seemed to have depended on the percentage of out-of-gamut pixels in a particular painting. The painting which had 68 per cent of its pixels out-of-gamut was best reproduced by the reproductions having gamut compression, paper cast removal and lightness mapping applied to them whereas the painting with virtually all pixels in the gamut of the reproduction device was best reproduced without alteration of the original data, or with only paper cast removal. The third painting, having 30 per cent out of gamut pixels, was best reproduced by the images having lightness mapping and paper cast removal. The results also suggest that there is no statistically significant difference between the two gamut clipping techniques used. For a more detailed description of the experiment see MacDonald et al. (1995).