Billger, M. et al. AN ASSESSMENT METHOD FOR EVALUATING COLOUR RENDERING PROPERTIES OF LIGHT SOURCES AN ASSESSMENT METHOD FOR EVALUATING COLOUR RENDERING PROPERTIES OF LIGHT SOURCES Monica Billger 1 , Björn Löfving 2 , Jörgen Thaung 2 1 Chalmers Architecture, Gothenburg, SWEDEN, 2 University of Gothenburg, Ophthalmology, Gothenburg, SWEDEN monica.billger@chalmers.se Abstract We present an assessment method to evaluate how the spectral power distribution of light influences the perception of colours and small colour differences. Included in the method are several ways to measure properties of reflected light with a spectroradiometer, visual assessments of perceived colours, and a visual performance test that aims at identifying an individual minimum colour contrast visibility threshold. In a first application of the assessment method, we used a colour contrast test chart printed on paper and studied it under the illumination of three LED light sources, halogen light and daylight. We concluded that the visual performance test was effective for reviling visibility thresholds for colour contrasts and that the method has potential for understanding the influence of spectral power distribution on perceived colour and visual performance. The methodology can be useful for optimizing illumination for individual needs. Keywords: colour contrast, colour rendering, visual performance, assessment methods 1 Introduction The spectral power distribution (SPD) of lamps through history has been determined by the available technology at any given time. In recent years, with the development of light-emitting diode (LED) based light sources, there are unique possibilities to customize the SPD to better suit the needs in different contexts (Wei et al 2014). Today, in 2015, the most commonly used principle in white LED technology is to convert the power from a short wavelength LED into longer wavelengths with phosphors, to cover most of the visible light spectrum. At present, focus is to develop LED light sources that produce as high luminous efficiency as possible, to defeat rival technology. However, there are other aspects on light except luminous efficiency that deserve attention, for example colour rendering, spectral homogeneity and naturalness. Colour rendering is important for many daily tasks like judging skin tones, quality of food etc., but also for detecting objects on a background with similar colours. Different kinds of metrics and assessment methods are used in colour and lighting research to describe how different light sources affect colour appearance (CIE 2004, Smet et al 2011, Hårleman et al 2007). Many studies aim at defining preference, attractiveness and naturalness of colour appearance under illumination of different light sources (Smet et al 2011, Islam et al 2013). Another kind of test is based on judging the visual clarity or feeling of contrast of colours under a specific illumination (Kenijiro 2007). These methods don’t take visual performance into account i.e. when observers are forced to perform a task other than assessments. The most common performance tests are colour vision tests and common for these tests are that they are all designed to detect colour vision deficiencies. Our ability to detect and discriminate different colours across the entire visual spectrum is not necessarily covered by these tests. Few tests on colour rendering are based on visual performance and to define how illumination can support human vision. One example is Mahler et al 2009, who have mapped colour rendering of LED light sources. In their study, a sorting test was applied where the task was to arrange coloured capsules in a colour circle under illumination from different light sources. They describe for example how RGB LED distorted colour appearance so that patches were falsely saturated, i.e. more colourful but less efficient to discriminate between.