Development of a multichannel spectral simulation tool and experimental validation with different lighting scenarios

The spectral composition of light has been linked to various non-image-forming responses besides visual photoreception. Accordingly, simulation tools must incorporate the spectral composition of light to account for such responses. A simulation tool was developed which uses N-step algorithm and subdivides the (red, green, and blue) RGB bands into multiple channels. This research intends to validate the tool for different lighting scenarios. A physical model was constructed in which the integral irradiance from 380 nm to 780 nm was measured for three scenarios: diffuse daylight, electric light with variable correlated colour temperature and a combination of both. All three scenarios were simulated with 3, 9, 27 and 81 channels. For scenarios with electric light and combination of daylight and electric light, the nine-channel simulation improved the mean absolute percentage error (MAPE) by 13.9% to 33.9% compared to the three-channel simulation. For continuous daylight, there was only a small improvement of 0.4% when increasing from 3 to 27 channels. In comparison to 9 channels, 27 channels slightly improved MAPE in all the scenarios but substantially increased the simulation time. Increasing the number of channels to 81 is likelier to bring a contribution to more complex scenarios than that presented in this study.