Evaluation of algorithms for color gamut compression was conducted using two different experimental methods. In the first, real prints in a viewing cabinet were compared side- by-side with the original images displayed on a CRT monitor. In the second, virtual prints were simulated on the monitor for comparison with the original images. The results showed that the new topographic algorithm achieved a good performance in both cases. Comparison of the two experiments, however, indicated rather different results for the other three algorithms applied to the same four test images, with the magnitude of differences between the algorithms being smaller in the case of simulated prints.