An HTTP/2-Based Adaptive Streaming Framework for 360° Virtual Reality Videos

Abstract

Virtual Reality (VR) devices are becoming accessible to a large public, which is going to increase the demand for 360°VR videos. VR videos are often characterized by a poor quality of experience, due to the high bandwidth required to stream the 360°video. To overcome this issue, we spatially divide the VR video into tiles, so that each temporal segment is composed of several spatial tiles. Only the tiles belonging to the viewport, the region of the video watched by the user, are streamed at the highest quality. The other tiles are instead streamed at a lower quality. We also propose an algorithm to predict the future viewport position and minimize quality transitions during viewport changes. The video is delivered using the server push feature of the HTTP/2 protocol. Instead of retrieving each tile individually, the client issues a single push request to the server, so that all the required tiles are automatically pushed back to back. This approach allows to increase the achieved throughput, especially in mobile, high RTT networks. In this paper, we detail the proposed framework and present a prototype developed to test its performance using real-world 4G bandwidth traces. Particularly, our approach can save bandwidth up to 35% without severely impacting the quality viewed by the user, when compared to a traditional non-tiled VR streaming solution. Moreover, in high RTT conditions, our HTTP/2 approach can reach 3 times the throughput of tiled streaming over HTTP/1.1, and consistently reduce freeze time. These results represent a major improvement for the e cient delivery of 360°VR videos over the Internet.

Spherical Clustering of Users Navigating 360° Content

“…Moreover, in the K-means clustering, the number of clusters K is imposed as the value achieved by the Louvain method (labelled "K-means 1"), as well as the K value obtained from our proposed clustering (labelled "K-means 2"). The proposed implementations have been made publicly available 4 Table 1. Clustering analysis of users in three selected frames from Rollercoaster (first half) and Timelapse (second half).…”

“…Understanding how users explore the VR content is important in order to optimize content creation [1] and distribution [2][3][4][5][6], develop user-centric services [7,8], and even for medical applications that use VR to study psychiatric disorders [9]. In the last few years, many studies have appeared collecting and analysing the navigation patterns of users watching VR content [6,8,[10][11][12][13][14][15][16].…”

“…Clusters should be disjoint for most content-delivery applications. For example, if clusters are used for prediction, each user must belong only to one cluster 4. https://github.com/LASP-UCL/spherical-clustering-in-VR-content.…”

Network Support for AR/VR and Immersive Video Application: A Survey

“…In (Petrangeli et al, 2017b), they presented a novel framework for the efficient streaming of VR videos over the Internet, which aims to reduce the high bandwidth requirements and storage costs of current VR streaming solutions. The video is spatially divided in tiles using H.265, and only tiles belonging to the user viewport are streamed at the highest quality.…”

Pano

“…Unfortunately, the current approaches fall short of achieving this goal. Most solutions (e.g., [26,32,34,50,52,59,68]) follow the viewport-driven approach, where only the viewport (the region facing the viewer) is streamed in high quality, but this approach has several limitations. First, a viewport (∼110°in width [63]) is still much larger than a laptop screen (∼48°in width) as perceived by users, so to stream a viewport region would still need at least twice the bandwidth of streaming a screen-size video at the same quality [28].…”