A perceptual model of motion quality for rendering with adaptive refresh-rate and resolution

摘要

Limited GPU performance budgets and transmission bandwidths mean thatreal-time rendering often has to compromise on the spatial resolution ortemporal resolution (refresh rate). A common practice is to keep either theresolution or the refresh rate constant and dynamically control the othervariable. But this strategy is non-optimal when the velocity of displayedcontent varies. To find the best trade-off between the spatial resolution andrefresh rate, we propose a perceptual visual model that predicts the quality ofmotion given an object velocity and predictability of motion. The model considerstwo motion artifacts to establish an overall quality score: non-smooth(juddery) motion, and blur. Blur is modeled as a combined effect of eyemotion, finite refresh rate and display resolution. To fit the free parametersof the proposed visual model, we measured eye movement for predictableand unpredictable motion, and conducted psychophysical experiments tomeasure the quality of motion from 50 Hz to 165 Hz. We demonstrate theutility of the model with our on-the-fly motion-adaptive rendering algorithmthat adjusts the refresh rate of a G-Sync-capable monitor based on a givenrendering budget and observed object motion. Our psychophysical validationexperiments demonstrate that the proposed algorithm performs better thanconstant-refresh-rate solutions, showing that motion-adaptive rendering isan attractive technique for driving variable-refresh-rate displays.