Quality of Experience in Multimedia Streaming: Objective Metrics, Subjective Assessment, and Trends for Research and Practice

Abstract

Quality of Experience (QoE) has become the dominant success criterion for modern multimedia streaming, yet the literature is fragmented across network QoS studies, perceptual quality models, subjective user experiments, and data-driven prediction. This survey consolidates these strands and explains how to translate QoE concepts into measurable, comparable, and actionable engineering signals. Specifically, we (i) propose a taxonomy linking network-, application-, and perceptual-level metrics to user-perceived outcomes; (ii) synthesize evidence on how objective indicators relate to subjective ratings (MOS/ACR) in adaptive streaming scenarios (startup delay, rebuffering, and quality switching); (iii) compare widely used models and standards (e.g., ITU-T P.1203/P.1204 and VMAF) with their assumptions and limitations; and (iv) summarize datasets, protocols, and practical guidance for reproducible evaluation and machine learning-based QoE prediction. Finally, we highlight open gaps—live-streaming QoE, cross-device perception, and model generalization under distribution shift—and outline research directions for 5G/6G, edge computing, and immersive media.