Abstract
We investigated chromatic induction, which has been attributed to either von Kries adaptation or later-stage post-receptoral adaptation, or a combination of both. Observers' heads were placed inside a large opaque hemisphere, the interior of which consisted of a 3 degree neutral disk surrounded by a chromatic ganzfeld that filled the remainder of the visual field. This apparatus, as opposed to a monitor display, not only allows us to fill the observer's entire visual field, but also to obtain a highly saturated surround of high luminance close to the chromaticity of a spectral color. Both the saturation and luminance of the disk and ganzfeld can be separately adjusted. After one minute of dark adaptation, 20 observers adapted for 45-60 seconds to a colored surround and then adjusted the disk color to make it appear neutral. Conditions included six red surround saturation levels, within a range of 0.736-0.868 in L/(L+M) coordinates, all at a luminance of 90 cd/m2 and four disk values, including two decremental targets (40, 60 cd/m2), one equiluminant, and one incremental target (100 cd/m2). More surround color was needed to null the induced color on a darker target than on a brighter one. This is inconsistent with Kirschmann's third law. However, consistent with Kirschmann's fourth law, the L/(L+M) coordinate of a neutralized target increases non-linearly with that of the surround. This non-linearity, in conflict with Weber cone contrast prediction, becomes more evident at higher surround saturations (L/(L+M) >0.76). According to von Kries adaptation, the chromaticity of all targets which look neutral in a given surround should be the same at all luminance levels. Our data show clear violations of this prediction. Observers found the task difficult, with a majority reporting both red and green close to the neutral setting.
Meeting abstract presented at VSS 2017