How Well Can People Predict Subtractive Mixing?

This study is a preliminary investigation towards the design of effective colour spaces and colour tools to allow users to quickly and accurately select a given colour in a digital-display environment. It has been shown that many non-expert users find the RGB colour space to be non-intuitive. The choice of colour space on various visual tasks has also been shown to be an important factor and that experts show greater precision in colour-matching experiments than non-expert observers. We propose that non-experts find manipulation and selection in an RGB colour space to be difficult because they do not possess an appropriate internal model for additive colour mixing. On the other hand, observers from a young age may develop a useful internal model of subtractive colour mixing processes as they experiment with inks and paints. The purpose of this work is to determine whether it is indeed the case that observers possess more useful internal models for subtractive colour mixing than for additive colour mixing. The work reported in this study describes only an assessment of subtractive colour mixing. Three experiments are described whereby expert and naïve observers select matches from a library of colours for individual samples or for imagined subtractive mixtures of paint samples. Qualitative and quantitative analyses are presented to measure the ability of observers to make predictions of subtractive mixing processes. When mixing the subtractive primaries (e.g. cyan with magenta) the performance of expert and naïve observers are the same but in general the expert performance far exceeds that of the naïve observer for other colours.