The perception of realism is correlated with physical gamuts

The relationship between the perceived realism of natural scenes and the physical gamut of colors remains a topic of debate in vision science. This research explores whether the physical gamut, derived from the CIE 1931 color space, influences judgments of realism in natural settings. By presenting observers with manipulated scenes across the color spectrum, the study found a moderate-to-strong correlation between realism judgments and the theoretical physical gamut.

However, the analysis identified the most detrimental points within the saturated green region of the CIE 1931 xy chromaticity diagram. Removing these points led to a significantly stronger correlation. To address this discrepancy, the researchers modeled a real-world physical gamut based on actual spectral data from natural and synthetic sources, as recorded in USGS and ECOSTRESS databases.

This real-world gamut demonstrated no detrimental green chromaticities, suggesting that the theoretical model's failure may be due to the absence of these colors in natural environments. The study posits that the visual system constructs internal color references through empirical observations of the world, implying that the theoretical model may not accurately reflect the actual gamut of visible light.

Consequently, the real-world gamut showed an even stronger correlation with realism judgments, supporting the hypothesis while highlighting the potential limitations of theoretical approaches. These findings contribute to the ongoing discourse on perceptual realism and emphasize the importance of considering real-world data in understanding the intricacies of visual perception.