Abstract
Even under restricted viewing conditions (e.g., monocular, stationary) people usually recognize that an LCD screen emits light instead of reflecting incident light. In previous experiments, we found that colour and texture were driving cues for glow detection with LCD screens. When a translucent, textured paper sample was placed in front of a computer screen, and the CIE xy chromaticity coordinates of the screen were matched to paper samples, participants were unable to differentiate between small patches of an LCD screen and real paper. Here, we hypothesize that the same realism can be achieved by matching the CIE xy chromaticity coordinates of the environment's lighting to the LCD monitor's white point. In a 9AFC task, observers viewed a 3x3 grid of nine 3.2 cm square apertures. Through one randomly chosen aperture, observers viewed a sample of translucent paper on an LCD screen, and through the other eight apertures, they viewed samples of opaque paper. The observer judged which aperture was light-emitting rather than reflective. Conditions 1 and 2 took place under beige ambient lights (CIE x=0.39; y=0.38). In these condition, the (1) luminance or (2) colour (CIE XYZ) of the translucent paper was matched to a randomly chosen paper sample. In condition 3, the lighting in the room was matched to the computer's white point (x=0.31, y=0.32) and the screen showed luminance-calibrated patches. Observers were significantly better at identifying the light-emitting patch in condition 1 than in conditions 2 and 3, but performance in the latter two was still above chance. We conclude that neither color matching screen display under normal room lighting or matching the chromaticity of the environment's lighting to an LCD screen is sufficient to eliminate cues to glow.
Meeting abstract presented at VSS 2018