From luminance to brightness: A data-driven approach to support brightness assessments in open plan offices

Abstract

People can instantly distinguish a brighter from a darker environment, but it is still unknown how to estimate brightness from the luminance distribution in complex visual scenes. In this study, we performed a meta-analysis of three experiments in which participants assessed brightness in an open plan office environment. Experiment 1 varied the luminance distribution on the wall, Experiment 2 varied the desk illuminance and Experiment 3 varied the ceiling illumination. Correlating various measures derived from high-resolution luminance images with participants’ brightness ratings, we investigated to what extent brightness could be predicted. In particular, we focused on 19 different luminance distribution characteristics calculated over 11 different areas of the visual field. In line with earlier work, participants could be grouped in two categories, one consisting of participants who substantially and consistently varied in their brightness assessments for the different settings, and the other consisting of participants who responded more evenly, regardless of the setting. Based on the brightness-responsive group of participants, brightness could be best modelled with the logarithm of the median luminance calculated over a 60° horizontal band in the field of view or with the logarithm of the 95th percentile over the median calculated over the 40° horizontal band, explaining, respectively, 38% and 35% of the variance in brightness perception.

The effect of correlated colour temperature and wall luminance on spatial brightness and scene preference in a windowless office setup

“…A pilot experiment by Sullivan et al 19 was also conducted, which indicated that less uniformly lit spaces were evaluated as brighter. A similar result was also found in Tiller et al, 22 whereas the data in de Vries et al 18 suggested the opposite. The contradictory results may be attributed to the lack of an unambiguous definition of uniformity in complex scenes.…”

“…9 Other studies discovered that the average luminance and the standard deviation of the luminance distribution in the horizontal 40° band are good predictors for the acceptability of lighting scenes. 16,17 In a recent paper by de Vries et al , 18 a meta-analysis of three experiments assessing brightness in an open office environment was conducted. In the brightness-responsive group of observers, brightness was best modelled with the logarithm of the median luminance of a 60° horizontal band or the logarithm of the 95th percentile luminance over the median luminance in a 40° horizontal band.…”