Real-Time Estimation of User-Level QoS in Audio-Video IP Transmission by Using Temporal and Spatial Quality

This paper deals with audio-video transmission over IP networks and proposes a method of estimating perceptual QoS (i.e., user-level QoS) in real time in terms of the interval scale, which is referred to as the psychological scale. The proposed scheme utilizes both temporal and no reference spatial (picture's) quality, which are application-level QoS, and estimates the user-level QoS from the application-level QoS. To evaluate the effectiveness of the proposed scheme, we compare the proposed scheme with a scheme which uses only temporal quality for the estimation and a scheme with both temporal and Full Reference spatial quality. We made a simple experiment of audio-video IP transmission; we measured the application-level QoS directly and assessed the user-level QoS by the method of successive categories. Applying the principal component analysis and multiple regression analysis to the experimental results, we obtained multiple regression lines (i.e., equations for the estimation). As a result, we see that the user-level QoS can be estimated with high accuracy by using both temporal quality and spatial quality.