AR-DAVID is a video quality dataset that contains display distortions (colour fringes, dithering, blur, ...) as seen on an optical see-through augmented reality (OST-AR) display. The AR display was simulated on a haploscope with beam splitters. The quality was measured for two luminance levels of the environment light (10 and 100 cd/m^2) and three different background patterns (simulated environments).

The dataset contains scaled quality scores in just-objectionable-difference (JOD) units and "raw" pairwise comparison data.

The videos are stored as h264 .mp4 files.

More details can be found in README.md.