EMQF 2014 - Special Session on Experiencing Multimedia Quality Fluctuations

Organizers
Raimund Schatz, FTW, Austria
Tobias Hossfeld, University of Würzburg, Germany
Topics of Interest
Measurement studies and models on typical quality fluctuations (e.g. in wireless networks)
Subjective assessment methods for session-level quality fluctuations
Understanding of human perception of stimulus variations (in general and modality-specific)
Measurement and modeling frameworks for quality variations
Extension of objective metrics and integration of variability-related factors
Inclusion of context-related (e.g. lighting, location) and content-related (e.g. scene changes, program changes) factors causing fluctuations
Evaluation of fluctuation-aware network and application management approaches
Databases featuring representative temporal variations of key variables
Motivation and Objectives
Research on Quality of Experience (QoE) of networked multimedia applications addresses the influence of user-, system- and context-related factors, with the majority of work focusing on static configurations of quality-related parameters and stimuli. Typical examples are studies aiming to quantify the influence of constant bandwidth bottlenecks, packet-loss and delay settings or of a specific encoding or content type on the perceived quality of video, voice, Web and other applications.
However, with the proliferation of mobile wireless use cases and quality management technologies (such as adaptive bitrate adaptation for streaming video), quality as experienced by end users is increasingly becoming subject to temporal variations and fluctuations. As regards the wireless mobile domain, aforementioned trend towards increased dynamics is driven by variations of underlying technical performance parameters like available bandwidth and packet loss as caused by changing cell load, network congestion or even outages. In addition, user mobility does not only lead to quality fluctuations related to handovers, signal strength changes, etc., but also to shifts of contextual influence factors like location, lighting and background noise. Even in static environments, aforementioned context-related factors may cause quality fluctuations. Analogously, content-related factors like scene changes or channel switching can manifest as quality variations.
Similarly, quality management mechanisms and technologies (like MPEG-DASH, HAS and Smooth Streaming in the case of Internet video) do not completely eliminate impairments they target (like rebuffering and freezing), but rather trade them off for other ones (e.g. switches in video encoding quality). The outcomes, i.e. quality fluctuations, are prone to become QoE impairments themselves because of the quality dynamics they tend to add to the experience.
All of the above developments necessitate intensified efforts towards investigating the impact of quality fluctuations on QoE. This does not only relate to improved subjective assessment methods (beyond existing continuous quality evaluation methods like SSCQE), but also to better understanding of stimulus variations on human quality perception in general. Furthermore, improved methods for adequately describing or modeling quality fluctuations (on technology as well as on perception level) and for integrating them in objective quality metrics and measurement techniques are required. Finally, QoE-based network and application management approaches are supposed to particularly benefit from these efforts, as they need to adequately address (and mitigate) temporal variations of quality-related parameters to reach their optimization goals.
Given these challenges, the main goal of this special session is to raise awareness of the QoMEX community for this topic by bundling existing efforts already related to it. In this context, outreach to related scientific communities (networking, speech and psychology) and the realization of synergies between different strands (video, image, audio, etc.) of QoE research constitute important elements in setting up this special session.