Preferred Color Gamut Boundaries for Wide-Gamut and Multi-Primary Displays

Preferred chroma enhancement and its dependence on hue are studied in a twopart experiment using a wide-gamut multi-primary display. Earlier research showed a clear dependence on hue but was limited by the gamut of the display it employed; the present work builds on this while removing the gamut constraints. In the first part of the present experiment, a tuning task was used to refine the preference for chroma boost starting with standard-gamut (Rec. 709) images. The overall median preferred boost is roughly 20%, but it is not uniform over hues: the preferred boost for orange, yellow, green, and cyan colors is greater than that for blue, magenta, and red colors. Dependence on image content and observer is noted, though a content-independent chroma boost created by aggregating preference over many images performs well, on average. An adjustment parameter for overall chroma which incorporates the hue dependence averaged over image content should be sufficient to handle the vast majority of inter-observer variance in preference. In the second part of the experiment, various chroma boost algorithms were evaluated through a paired comparison task. The prescribed hue-dependent chroma boost is preferred over all other variations, and all hue-preserving chroma boost variations are preferred over both colorimetrically accurate and naive same-drive-signal renderings. The results may be applied to display design to select gamut boundaries that maximize satisfaction over the observer population.