A novel approach is used to characterize and compare the performance of night vision systems in conditions more
representative of night operation in terms of spectral content. Its main advantage compared to standard testing
methodologies is that it provides a fast and efficient way for untrained observers to compare night vision system
performances with realistic illumination spectra. The testing methodology relies on a custom tumbling-E target and on a
new LED-based illumination source that better emulates night sky spectral irradiances from deep overcast starlight to
quarter-moon conditions. In this paper, we describe the setup and we demonstrate that the novel approach can be an
efficient method to characterize among others night vision goggles (NVG) performances with a small error on the photogenerated
electrons compared to the STANAG 4351 procedure.