Light source classification and colour change modelling for understanding and predicting pigments discolouration

In this study, we introduce a simple but effective model, combined with a classification system for light sources based on their spectral profile approximated by the relative power distribution in the red, green and blue regions, which successfully fits experimental data of pigment colour changes in an oil binder under various lighting conditions. This model is experimentally validated by artificial ageing tests on two sets of model samples made of historical pigments (strontium yellow and Prussian blue mixed with lead white) using three white light sources (two WLEDs and a xenon light source). The model accurately fits colour shifts by analysing changes in International Commission on Illumination L*a*b* values. The results highlight that integrating the spectral classification of the light sources with the colour change model provides deeper insights into the causes of paint colour changes than traditional methods that rely solely on total irradiance. This combined approach enables precise prediction of long-term colour changes by exploiting the spectral characteristics of light sources.