Dissecting cloud game streaming platforms regarding the impacts of video encoding and networking constraints on QoE

Autor: Frédérique Humbert, Charline Taibi, Christoph Neumann, Charles Salmon-Legagneur, Franck Aumont
Rok vydání: 2021
Předmět:
Zdroj: GameSys@MMSys
DOI: 10.1145/3458335.3460963
Popis: Cloud gaming -- sometimes also referred to as game streaming -- is the concept of executing games on remote servers -- typically within a cloud -- capturing the video output normally displayed onto a screen and sending the result as a video stream to end-user devices. While many game streaming solutions have existed for many years now with first proof of concept shown in 2000, many of these solutions still seem to suffer from technical or quality limitations. Recently, several new actors entered the cloud gaming market with the promise to solve these limitations. This paper provides a systematic benchmark of two game streaming services -- Google Stadia and Nvidia GeForce Now -- under a variety of networking conditions, while subjectively evaluating the resulting QoE. We further expose and analyse the network adaptation and video encoding choices of these two platforms. We present a few key takeaways and provide recommendations regarding video encoding choices, buffering and bitrate adaptation mechanisms.
Databáze: OpenAIRE