Abstract
Some objects in the real world themselves emit a light, and we typically have a fairly good idea as to whether a given object is self-luminous or illuminated by a light source. However, it is not well understood how our visual system makes this judgement. This study aimed to identify determinants of luminosity threshold, a luminance level at which the surface begins to appear self-luminous. We specifically tested a hypothesis that our visual system knows a maximum luminance level that a surface can reach under the physical constraint that surface cannot reflect more lights that incident lights and apply this prior to determine the luminosity thresholds. Observers were presented a 2-degree circular test field surrounded by numerous overlapping color circles, and luminosity thresholds were measured as a function of (i) the chromaticity of the test field, (ii) the shape of surrounding color distribution and (iii) the color of illuminant lighting surrounding colors. We found that the luminosity thresholds strongly depended on test chromaticity and peaked around the chromaticity of test illuminants and decreased as the purity of the test chromaticity increased. However, the locus of luminosity thresholds over chromaticities were nearly invariant regardless of the shape of surrounding color distribution and generally well resembled the locus drawn from theoretical upper-limit luminance but also the locus drawn from the upper boundary of real objects. These trends were particularly evident for test illuminants on blue-yellow axis and curiously did not hold under atypical illuminants such as magenta or green. Based on these results, we propose a theory that our visual system empirically internalizes the gamut of surface colors under illuminants typically found in natural environments and a given surface appears self-luminous when its luminance exceeds this heuristic upper-limit luminance.
Competing Interest Statement
The authors have declared no competing interest.