Towards a discomfort glare metric for non-uniform luminaires

Discomfort glare is defined by the International Commission on Illumination (CIE) as: “glare that causes discomfort without necessarily impairing the vision of objects” and by the Illuminating Engineering Society of North America (IES) as “a sensation of annoyance or pain caused by high luminance in the field of view”. The quantification of discomfort glare has been researched ever since the beginning of the previous century, but even today, the physiological and psychological mechanisms are not fully understood. 
Discomfort glare can have several causes. Both an erratic luminance distribution and a high luminance relative to the average luminance level of the field of view can cause annoyance. The traditional glare metrics often only include the average luminance level calculated from the far field luminous intensity distribution. Any non-uniformity in luminance distribution is ignored and luminaires are approximated by a point source. Therefore, discomfort glare from relative uniform light sources is adequately described by the UGR whereas discomfort glare from non-uniform light sources is not. Since a non-uniform luminaire produces more discomfort glare than a uniform one of equal average luminance, the applicability of traditional glare metrics for non-uniform light sources is under discussion. LED luminaires often exhibit a non-uniform luminance distribution with areas of high luminance alternating with relatively dark zones. A discussion arises if the low luminance area should be considered for the total luminous surface. The non-uniformity of a luminaire can be accurately described by its luminance map. Since nowadays LED luminaires dominate the market for interior lighting, a better assessment of discomfort glare based on luminance maps becomes essential.
As an attempt to tackle these problems, the CIE developed a method for small, large and complex sources but discussion about the validity remains, as stated by a CIE reporter (CIE, 2009):
‘Your reporter has referred to UGR as a “leaky boat”. Is it time, therefore, to develop a new system rather than to try further patch the “leaks”?’
Although some physiological mechanisms involved in glare perception are known, sometimes already for decades, traditional glare formulae are merely phenomenological and lack any physiological or psychological justification. As an alternative, by incorporating the receptive field mechanism, the pupillary light reflex and a correction for retinal position, a more physiologically justified discomfort glare model is developed. The receptive field neural pathways, practically acting like an edge filter, have been studied already from the 1930’s on. The pupillary light reflex controls the retinal illuminance as part of the adaptation process. The Guth position index describes the relation between the light source retinal position and the apparent brightness: the more the image is located away from the line of sight, the less bright it will appear.
This thesis includes two parts. In the first part, the traditional UGR method is explained. The appropriateness of the UGR calculation for uniform and non-uniform luminaires based on luminance maps is compared. Ultimately, a proposition to improve the UGR method is outlined. The actual luminous surface of a luminaire is defined by a luminance boundary discriminating between background and luminous part(s). In the second part, a more physiologically justified discomfort glare model is developed including the receptive field mechanism, the pupillary light reflex and a retinal position correction. This model is a promising alternative to current, merely empirical glare metrics, especially when non-uniform luminaires are to be evaluated.