Delay Sensitivity Classification of Cloud Gaming Content

Saeed Shafiee Sabet, Steven Schmidt, Saman Zadtootaghaj, Carsten Griwodz, Sebastian Möller

In: MMVE '20: Proceedings of the 12th ACM International Workshop on Immersive Mixed and Virtual Environment Systems. ACM International Workshop on Immersive Mixed and Virtual Environment Systems (MMVE-20) located at MMSys '20: 11th ACM Multimedia Systems Conference June 8-8 Istanbul Turkey Pages 25-30 ISBN 978-1-4503-7947-2 ACM New York, NY 2020.


Cloud Gaming is an emerging service that catches growing interest in the research community as well as industry. Cloud Gaming require a highly reliable and low latency network to achieve a satisfying Quality of Experience (QoE) for its users. Using a cloud gaming service with high latency would harm the interaction of the user with the game, leading to a decrease in playing performance and, thus players frustrations. However, the negative effect of delay on gaming QoE depends strongly on the game content. At a certain level of delay, a slow-paced card game is typically not as delay sensitive as a shooting game. For optimal resource allocation and quality estimation, it is highly important for cloud providers, game developers, and network planners to consider the impact of the game content. This paper contributes to a better understanding of the delay impact on QoE for cloud gaming applications by identifying game characteristics influencing the delay perception of the users. In addition, an expert evaluation methodology to quantify these characteristics as well as a delay sensitivity classification based on a decision tree are presented. The results indicated an excellent level of agreement, which demonstrates the reliability of the proposed method. Additionally, the decision tree reached an accuracy of 90% on determining the delay sensitivity classes which were derived from a large dataset of subjective input quality ratings during a series of experiments.

German Research Center for Artificial Intelligence
Deutsches Forschungszentrum für Künstliche Intelligenz