A perceptual model of motion qualityfor rendering with adaptive refresh-rate and resolutionSupplementary material

In the main paper we argue that the difference in perceived quality can be expressed as the difference in CSF-normalized energy of two stimuli (Sections 5.3 and 5.4). In this section we demonstrate how this approach can be explained by detection models. In Equation 13 in the main paper we normalized amplitudes of the Fourier coefficientsm(ω;σ ) by the CSF to account for the visibility thresholds: m̃(ω,σ ) = CSF(ω)m(ω;σ ) , (1)