A Multi-Path-Based Adaptive Scheme for Multi-View Streaming Over HTTP

Abstract

Listen

Translate article

A majority of the existing research on HTTP focuses on improving the quality of experience for single-source videos. Recently, there has been a focus on multi-view streaming applications, where multiple videos are streamed simultaneously. The challenge is to improve the viewing experience of concurrently streamed videos over the bandwidth-constrained network. In this paper, we propose an adaptive scheme for multi-view streaming over HTTP that requests segments of multiple videos concurrently. First, we propose a unified throughput estimation method to calculate the throughput, solving unfair and inefficient throughput utilization problems of existing estimation methods. Second, we design a unified segment scheduling method that guarantees synchronicity among segments of concurrently downloaded video streams. Third, we present a multi-path streaming method that leverages path diversity to improve the user experience of the video streams. Finally, we present a rate adaptation algorithm that dynamically downloads high-quality segments of multiple video streams, while minimizing video quality changes and the risk of playback interruption in single- and multi-path environments.