This research project has two goals. One is to understand the third dimension of colour scales describing the extent of chromatic contents such as saturation, vividness, chromaticness and colourfulness, which are less widely used than the other dimensions, e.g. lightness and hue. With that in mind, the first aim of this work is to derive new models that may serve as an alternative to the third-dimension scale of colour appearance on the basis of colorimetric values. The second goal is to develop important scales, blackness and whiteness. They are widely used because of the popularity of the NCS system.
To achieve the first goal, a psychophysical experiment for scaling 15 attributes (Korean corresponding words of “bright”, “light-heavy”, “active-passive”, “fresh-stale”, “clean-dirty”, “clear”, “boring”, “natural-not natural”, “warm-cool”, “intense-weak”, “saturated”, “vivid-dull”, “distinct-indistinct”, “full-thin” and “striking”) using the NCS colour samples was carried out with Korean observers. Each sample was presented in a viewing cabinet in a darkened room. Naive observers were asked to scale each sample using a categorical judgement method.
From the results, two scales widely used to represent the third dimension were identified: saturation and vividness. The same samples were assessed by British observers using these two scales. There was a great similarity between the results of the British and Korean observers. Subsequently, more samples were included to scale not only the new third dimension scales (saturation and vividness) but also whiteness and blackness scales. In total, 120 samples were scaled for saturation, vividness and whiteness experiments, and 110 samples were scaled for a blackness experiment.
Four sets of models were developed for each of the three colour spaces (CIELAB, CIECAM02 and CAM02-UCS). Type one was based on the ellipsoid equation. Type two was based on the hue-dependent model proposed by Adams (called “the hue-based model”). Each of the above two types was used to fit the present experimental (Cho) data and the NCS data, which were measured using a spectrophotometer. In total, 39 models were developed.
The newly developed models were tested using the Cho and NCS datasets. The models that were based on the present visual data were tested using the NCS data. Similarly, the models developed from the NCS data were tested using the present visual data. The results showed that both types of models predicted visual data well. This means that the two sets of data showed good agreement. It is also proposed that the four scales (saturation, vividness, blackness and whiteness) based on CIECAM02 developed here are highly reliable.
