Abstract
The system gamma is the product of the encoding and decoding gammas. For cinema displays viewed in a dark environment the system gamma is 1.5, but for office displays, typically viewed in a lighter environment, the system gamma is nearly linear at 1.1. Lightness perception can be approximated as a gamma function of onscreen luminance (Stevens, 1975): for dark backgrounds, lightness perception is modeled with an exponent of less than one, for bright backgrounds the exponent may be greater than one (Whittle, 1994). Thus lightness- and system- gammas are inversely correlated. A recent study (Liu and Fairchild, 2007) demonstrated that for high dynamic range (HDR) images subjects often prefer a system gamma of less than one, presumably reflecting the greater need for histogram equalization in HDR images that are often dominated by dark, low contrast regions when linearly scaled for presentation on a LDR display. We investigate the interaction between lightness perception and histogram equalization by asking subjects to rate the quality of HDR images for various levels of system gamma. We find that preferred gamma is correlated with the dynamic range of the scene. Moreover, the highest image quality scores are found for images with a preferred gamma of one, suggesting a preference for linearity (where possible). Subject's preference for system gamma can be predicted by searching for the image with the flattest lightness distribution, where lightness is a pixel-wise gamma function of luminance that varies with the mean luminance of the stimulus and background.