QoE Assessment of Cloud-Based Social Extended Reality Applications Over Heterogeneous Access Networks

The next generation of immersive applications, such as eXtended reality (XR), will likely be cloud-based and streamed over mobile networks using myriad technologies such as WiFi and 6th-generation mobile networks. Mobile networks promise ubiquitous connectivity but are prone to stochastic network conditions that may be detrimental to end users' quality of experience (QoE). The impact of network conditions on QoE has been studied extensively by industry and academia regarding various multimedia services such as audio, video, and gaming. However, the impact of network conditions on users' QoE for XR-based social applications has yet to be thoroughly investigated. This paper presents novel results assessing the impact of network conditions <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$(\boldsymbol{N}=\mathbf{20})$</tex> involving factors such as round trip time (RTT), jitter (RJ), and packet losses (PL) on users' QoE via realistic subjective tests <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$(\boldsymbol{N}=\mathbf{28})$</tex> regarding social XR application. Our results show that social XR applications require stringent QoS conditions. In particular, our results show that increasing RTT values do not significantly affect users' QoE up to 77ms. Combined PL and RTT cases cause significant QoE degradation from 77ms onward with greater than 2% PL. Most importantly, results show that a very small jitter value with one standard deviation beyond 52 milliseconds can lead to significant QoE degradation. Further, jitter values beyond three standard deviations for 27ms RTT and beyond should be avoided.