A basic multimedia quality model

This paper describes two experiments designed to develop a basic multimedia predictive quality metric. In Experiment 1, two head and shoulder audio-video sequences were used for test material. Experiment 2 used one of the head and shoulder sequences from Experiment 1 together with a different, high-motion sequence. In both experiments, subjects assessed the audio quality first, followed by the video quality and finally a third test evaluated multimedia quality. The results of these studies found that human subjects integrate audio and video quality together using a multiplicative rule. A regression analysis using the subjective quality test data from each experiment found that: 1) for head and shoulder content, both modalities contribute significantly to the predictive power of the resultant model, although audio quality is weighted slightly higher than video quality and 2) for high-motion content, video quality is weighted significantly higher than audio quality.