B2.1 A mesopic experiment series at automotive visual conditions Nils Haferkemper, Anja Frohnapfel, Galina Paramei, Tran Quoc Khanh Keywords: mesopic vision, contrast threshold, spectral sensitivity 1 Introduction In the last few years, there has been a lot research action around the mesopic adaptation level and the necessity of this research for driver’s safety through lighting optimisation is beyond question. With LED Front Lighting Systems, a third light source enters the streets and in terms of contrast sensitivity in night time driving it isn’t sufficient to perform a “simple” achromatic contrast sensitivity determination but a research on the spectral characteristics of the chromatic contrast detection threshold based luminous efficiency function. One reason to analyse the spectral contrast sensitivity function at mesopic adaptation levels is the actual efficiency discussion going through the media. The aim was to find out, how much energy in which spectral region is really needed to achieve a specific contrast sensation. If there is a model that describes the human contrast sensitivity function for different adaptation levels, one can design i.e. street lighting, which achieves an optimised contrast sensitivity providing more safety for traffic participants and higher energy efficiency. Until now there are just few publications known [1] in which context researchers dealt with that kind of question, hence another purpose of the experiment series is to revise and validate their results. What a spectral analysis of human recognition offers, is the possibility to compute the quantitative parameters from a model by using real spectral power distributions (SPDs) of the light source and not just a chromaticity coordinate or a colour temperature. 2 Test design The approach of the pre-test, which took place at the Lighting Laboratory in Darmstadt, was primarily to get in touch with testing procedures in the mesopic range before starting the real test. Because we encountered a lot of problems