Evaluating Human Visual Preference and Performance in an Office Environment Using Luminance-based Metrics

There are not adequate human factors research studies available that examine luminance-based measures as they relate to human visual preference and acceptance in spaces with daylight. The objective of this research is to study several luminance-based metrics to support improved integrated lighting design recommendations, computational analysis methods and control technologies. Therefore, this dissertation executed a two-day pilot study (n=18) and a six-month repeated-measures experiment (n=45) in mock office spaces under naturally occurring daylight conditions in Boise, Idaho. Recent developments in High Dynamic Range digital photography permit investigation of high-resolution luminance-based metrics. Once these metrics and associated recommended criteria are established, they can support computational daylighting design analysis and integrated luminous environmental control systems to improve occupant satisfaction and increase energy savings over traditional illuminance-based methods. This dissertation also examined human visual performance differences in scenes rated as visually comfortable and those having "just uncomfortable glare." This was done to determine if, and to what extent, visual performance decrements exist under uncomfortable conditions.;Luminance-based metrics proved more capable than illuminance-based metrics at fitting the range of subjective responses on visual comfort items. Standard deviation of window luminance produced the highest adjusted squared correlation coefficient of any single metric with subjective responses (adjr 2=0.38, F1,860=536, p-value<0.01). Additionally, metrics based upon luminance within the 40° horizontal band of vision performed strongly. A bounded-borderline between comfort and discomfort is proposed as preliminary criteria for several of the highest-ranked metrics. Illuminance-based metrics, traditional luminance ratios and the recently proposed Daylight Glare Probability were less able to fit subjective responses. The strongest multiple regression model was for the "too dim-too bright" rating of the window wall (adj R2=0.49, F3,688=222, p-value < 0.01) and was built upon three variables: standard deviation of window luminance, 50th percentile luminance value from the view window and the percent of the 40° horizontal band > 2000 cd/m2.;Additionally, a significant decrement (0.51-1.65%) in visual performance was found for one of three objective performance tests in glaring scenes. A significant "seasonal" effect was found for measured sensitivity to brightness between summer and fall using a controlled repeated-measures test. Finally, several human acceptance-based approaches to improve luminous environmental control systems are proposed.