Human Perceptions of Daylight Composition in Architecture: a 1 Preliminary Study to Compare Quantitative Contrast Measures 2 with Subjective User Assessments in Hdr Renderings

Humans perceive daylit architecture as a rich and dynamic luminous composition and yet existing performance metrics most often evaluate natural light for its ability to provide adequate illumination to a two-dimensional task surface while avoiding glarebased visual discomfort. This rather limited taskdriven approach places a disproportionate emphasis on surface illumination and glare-based discomfort and ignores the likelihood that contrast can provide a positive visual impact on our impression of space. Existing studies on perceptual daylight performance have linked subjective ratings of contrast and brightness to simple global measures in digital images, yet the conclusions are varied and lack robust consensus. These ‘global’ contrast measures do not account for the composition of luminance values within a scene and while more robust methods have been developed in computational graphics, vision research, and psychology, they have not been applied to studies in qualitative lighting research. As daylightdriven visual effects are not only dependent on composition, but are heavily influenced by dynamic sky conditions, this paper will introduce an experimental method for comparing subjective ratings of daylight composition in architecture against existing global and local contrast metrics under a range of annual moments. This preliminary study will test the effects of sun position and spatial composition on subjective ratings for contrast, uniformity, complexity, variation, stimulation, and excitement. It will then identify which quantitative measures (local and/or global contrast metrics) correlate most strongly to these ratings.