Gertjan Scheir

A psychophysical model for visual discomfort based on receptive fields

Visual discomfort is predicted from a luminance map with a model based on the receptive field mechanism in the human eye. A centre-surround receptive field is described by a Difference of Gaussians. Eight commercially available office luminaires are assessed for visual discomfort in a paired comparison experiment. The correlation between the subjective data and the receptive field model is optimized for three factors: the centre Gaussian width, the surround Gaussian width and the centre-to-surround weighing factor (WF). A centre and surround visual angle of 0.53 and 2.19 min arc, respectively, and a WF of 0.87 result in a coefficient of determination of 0.77. The model is validated independently with magnitude estimation data obtaining a coefficient of determination of 0.82. Where the standard unified glare rating method fails (coefficient of determination of 0.45), the receptive field model ameliorates predictability for visual discomfort. The model based on receptive fields is promising to replace current standard glare metrics, specifically when non-uniform luminaires are to be evaluated.

Defining the Actual Luminous Surface in the Unified Glare Rating

ABSTRACT A nonuniform luminance distribution produces more discomfort glare than a uniform one of equal average luminance. Because the standard unified glare rating (UGR) generally considers the average luminance level of the total luminaire surface, its applicability for nonuniform light sources is under discussion. With a growing market share of highly nonuniform light emitting diode (LED) luminaires, a valid discomfort glare metric becomes essential. The UGR can elegantly be improved by discriminating between background and luminous part(s) of a luminaire. A luminance threshold identifies the high-luminance pixels in a luminance map as an actual light emitting surface. Two subjective visual experiments, encompassing commercially available recessed office luminaires and custom-designed luminaires, validate the improved UGR method. With a coefficient of determination of 0.45, the standard UGR fails to predict visual glare sensation. Also covering nonuniform light sources, the improved UGR, with a coefficient of determination of up to 0.91, elegantly ameliorates the visual discomfort calculation from a luminance map.

Pupillary light reflex, receptive field mechanism and correction for retinal position for the assessment of visual discomfort

Light sources causing annoyance or pain produce discomfort glare. Traditional glare metrics fail for non-uniform luminaires. As an alternative, visual discomfort is determined by a model incorporating the centre–surround receptive field mechanism, the pupillary light reflex and a correction for retinal position. The pupil area, controlled by the pupillary light reflex, regulates the retinal illuminance. A centre–surround receptive field, described by a difference of Gaussians, represents the visual signal. A correction according to the Guth position index accounts for the reduction in brightness perception when a light source is moved away from the line of sight. The model is analysed with a forced choice paired comparison experiment involving 17 non-uniform rear projected stimuli with different spatial frequencies and luminance steps. A coefficient of determination of 0.68 between the subjective assessment and the model is obtained. A paired comparison office luminaire experiment and a magnitude estimation experiment involving diffusor luminaires validate the model resulting in a coefficient of determination of 0.86 and 0.81, respectively. By including the pupillary light reflex, receptive field mechanism and a correction for retinal position, the more physiologically justified model is a promising alternative to current, often empirical, glare metrics, especially for non-uniform luminaires.