Video Service Quality Estimation in Significance Aware Transport Networks

As video distribution over optical transport networks becomes more and more important, considerable focus is put on providing the appropriate quality for this transport service. A typical trade-off seen in this context is between sophisticated compression techniques used in order to save required transmission bandwidth, and the sensitivity of service quality with respect to transport network induced information (packet) loss. At the application layer, sensitivity to losses due to transport level impairments is content dependent, thus introducing a gap between typically used quality of service (QoS) parameters and the actual perceived quality (quality of experience - QoE). Extending a proposal to enhance transport level quality insurance mechanisms to be aware of the QoE significance of transmitted data, this paper presents a proposal to use significance information for QoE monitoring within the transport network as well. Two measurement techniques are described for delivering QoE indicators. Based on analytical expressions for the influence of packet loss on the MPEG-2 video decoding a Quality of Experience metric is derived which is applicable in significance and non-significance aware video transmission environments. The influence of different video encoding parameters on the overall achievable video quality and the expected gain in a significance aware environment are evaluated.