Toward Adaptive Volumetric Video Streaming: A Joint Network-Viewport Adaptation Framework

Hao Chen, Bowei Xu, Shaowei Wang, Xun Cao, Zhan Ma

Abstract: Volumetric video offers users a splendid immersive experience by allowing interaction with six degrees of freedom (6-DoF). In practice, visibility-aware streaming has become the prevailing paradigm, in which visible parts of a volumetric video are prioritized to stream and display for users. However, the key challenge of this paradigm resides in balancing two fundamentally contradictory adaptabilities in 6-DoF video: basic network adaptability, and crucial viewport adaptability. Existing schemes still struggle to juggle these two adaptabilities, ultimately deteriorating the user's quality of experience (QoE). In this article, we propose a novel dual-adaptive streaming framework for volumetric video, which is adaptive to network conditions and the user's viewports jointly. We first introduce a flexible, scalable video content representation to support incremental quality upgrades by leveraging the inherent characteristics of typical tree-based volumetric video codecs. Then, a two-stage quality controller is devised to select temporal chunks and spatial tiles for downloads, effectively ensuring a high-quality viewing experience within the field of view (FoV) while minimizing rebuffering occurrences. Experimental results demonstrate that the proposed framework outperforms state-of-the-art schemes with significant improvements in QoE across a wide range of network conditions and user viewing behaviors. The dual-adaptive framework presented in this article is expected to serve as a fundamental cornerstone for the design and implementation of practical volumetric video streaming solutions.