Sustainable Video Streaming using Acceptability and Annoyance Paradigm

The energy consumption of information and communications technologies is increasing rapidly, contributing to the climate crisis. Video streaming technologies are particularly impactful, accounting for over half of this energy use. Recent advancements in video streaming technologies have improved user experience by delivering higher quality videos but at the expense of significantly increased data volumes. In pursuit of a more sustainable approach to video streaming, we investigate greener methods ways of integrating Quality of Experience (QoE) measures into video streaming pipelines. QoE defines to users' satisfaction while consuming video content. User satisfaction with video content is largely influenced by their expectations. Depending on the context, users may have varying expectations regarding video quality. To address this, the acceptability & annoyance paradigm is proposed to measure user satisfaction as a function of both user expectations and video quality. By understanding user expectations and satisfaction together, video streaming services can reduce energy costs associated with processing, storing, and transmitting video content while maintaining the same level of QoE. In this study, we focus on the acceptability and annoyance of short-form vertical user-generated video content commonly found on online social media platforms. We analyze the relationship between video quality and acceptability & annoyance, demonstrating how this paradigm can be leveraged to achieve significant bitrate savings while preserving user satisfaction.