Lawrence Berkeley National Laboratory

Glare models were reviewed with an eye for missing conditions or inconsistencies. We found ambiguities as to when to use small source versus large source models, and as to what constitutes a glare source in a complex scene. We also found surprisingly little information validating the assumed independence of the factors driving glare. A barrier to progress in glare research is the lack of a standardized dependent measure of glare. We inverted the glare models to predict luminance, and compared model predictions against the 1949 Luckiesh & Guth data that form the basis of many of them. The models perform surprisingly poorly, particularly with regards to the luminance-size relationship and additivity. Evaluating glare in complex scenes may require fundamental changes to form of the glare models.