Color appearance: Properties of the uniform appearance diagram derived from hue and saturation scaling

Abstract

632We have previously shown that color appearance can be specified very reliably by a procedure of direct scaling of hue and saturation (Gordon, Abramov, & Chan, 1994). In that procedure, participants use percentage scales to describe their sensations of the four unique hue sensations of red, yellow, green, and blue, as well as apparent saturation. Participants' responses to the same stimuli remain stable over long periods of time, do not depend on training or past experience with psychophysical procedures and magnitude estimation, and are independent of the participants' linguistic backgrounds.It has been clear at least since the time of Sir Isaac Newton (see Newton, 1704, p. 90) that color is a response of a sensory system and is not a direct property of the physical stimulus. This means that color appearance can and does change with viewing conditions. It is therefore useful to have a quick and simple procedure for specifying appearance for any given set of conditions, so that any changes in appearance can be examined and compared. Using our scaling procedure, a complete set of color appearance data can usually be obtained in a single session of about 1 h.Here, we describe how the scaling data can be used to generate a color space representative of the specific viewing conditions. A standard chromaticity diagram, such as the CIE 1931 x, y, z diagram and any of its later transformations, is not adequate for describing appearance. These diagrams are based on single-wavelength additivity mixtures that specify metameric matches. In practice, the spectral distribution of any stimulus is simply scaled by the fixed coefficients of the system to determine the placement of that stimulus on the diagram. Nowhere does appearance enter into the calculation. For example, a stimulus that plots near the spectrum locus close to about 580 nm will usually appear yellowish. However, its appearance can be shifted drastically if it is surrounded by either a reddish or a greenish field, and yet it still plots to exactly the same point on a chromaticity diagram. Moreover, by their very nature, these diagrams specify stimuli independently of their luminances, and yet it has long been known that many aspects of color appearance vary with luminance-such as, for example, the Bezold Brücke phenomenon and wavelength discrimination. Also, the problems with chromaticity diagrams are compounded © Color appearance can be specified by a procedure of direct hue and saturation scaling. The scaling data can be represented on a 2-D color space termed a uniform appearance diagram (UAD). The orthogonal and bipolar axes of the UAD are based on the four unique hue sensations, which are blue-yellow and green-red. We have previously shown that the technique is reliable and rapid. We now show that the UAD is sufficiently uniform metrically that it can be used to derive, from a single set of scaling data, a wide range of color functions, such as the spectral loci of the unique hues, wavelength discrimination, and similarities among very different colored ...

Fixation light hue bias revisited: Implications for using adaptive optics to study color vision

“…Mean ratings were computed for each subject and condition after performing an arcsine transform (Abramov et al, 2009) to insure appropriately distributed variance. Ratings were inspected for each subject as a function of stimulus intensity, and while mean hue was generally found to shift with increasing intensity (most notably decreasing in blueness and increasing in yellowness with smaller changes in the red-green balance), any differences in the red vs. white target or beacon absent vs. beacon present conditions were consistent at all intensities examined.…”

“…In addition to computing and plotting mean ratings in each color category, data were also visualized with Uniform Appearance Diagrams (also described by Abramov et al, 2009) by plotting the difference in green and red ratings on the y axis and the difference in yellow and blue ratings on the x axis. Figure 1 shows one example of the variation in color appearance of individual 580 nm foveal small spot stimuli for one subject (subject 1).…”

Sex and vision II: color appearance of monochromatic lights

“…Our magnitude scaling methods, derived directly from Hurvich and Jameson [22], require participants to assign numbers to the sensations elicited by each stimulus. To do this, we use a strict protocol (described below), whose reliability and validity we have explored quite extensively [23-27]. We used two optical systems: about half the participants viewed the stimuli with their natural pupils (Newtonian-View); for the others, the light from a second optical system was focused through the central 2 mm zone of the pupil (Maxwellian-View); in both cases, the illuminance on the retina was the same.…”

Adaptive optics without altering visual perception

“…For this reason we restricted data analysis to stimuli with frequencies of seeing between 20% and 85%. Mean ratings were computed for each subject and condition after performing an arcsine transform to uniformly distribute variance (Abramov, Gordon, & Chan, 2009). Two-tailed z -tests were used to assess the significance of differences in mean hue ratings across conditions, with results verified using non-parametric permutation tests (i.e.…”

“…Data were visualized with Uniform Appearance Diagrams (UAD; Abramov, Gordon, & Chan, 2009). Figure 3 shows an example UAD illustrating the variation in color appearance for repeated 580 nm foveal small spot stimuli for one subject (subject 1).…”