Session-level QoE models for temporal quality transitions

Develop session-level QoE models that capture the temporal effects of quality transitions in real-time interactive video streams.

Background

The resource-allocation framework primarily uses spatial QoE indicators such as VMAF and utility functions over quality levels. However, changes in quality over time can affect user experience independently of the instantaneous spatial quality, particularly in interactive video where latency, frame drops, and quality oscillations matter.

The conclusion explicitly identifies models capturing these temporal effects as a remaining open challenge, indicating that the current treatment does not fully characterize the QoE consequences of transitions between quality levels.

References

Key open challenges remain: runtime estimation of the spatial complexity curve from limited observations, session-level QoE models that capture the temporal effects of quality transitions, privacy-preserving mechanisms for sharing spatial complexity information, and incentive structures that discourage applications from misrepresenting their QoE to gain resource advantages.

— Prototyping QoE-Aware Rate Adaptation in Cellular Networks with Commercial Applications  (2609.09490 - Nádas et al., 8 Sep 2026) in Section VI, “Conclusion”