360° Video Transmission Based On Tile And Bitrate Adaptation

With the advent of the 5 G era,the proportion of video content in the Internet continues to increase,and emerging media such as virtual reality(VR)and 3600 video have also entered people’s lives.Because 360°video requires high performance such as network bandwidth and delay,it is a huge challenge to transmit high-quality video content in complex networks.In order to reduce problems such as stuttering during video playback,the bitrate-based adaptive algorithm can effectively improve the quality of video transmission,and has been successfully applied to ordinary video,but the application of 360° video is still in the development and exploration stage.For ordinary video and 360°video,this paper studies high-quality video transmission from the perspective of video dynamic frame rate.A video frame rate and its bitrate adaptive algorithm is proposed for ordinary video,which is a real-time exchange model of bandwidth and delay,which can be used to combat network fluctuations.The simulation results show that the algorithm proposed in this paper has good performance in download bitrate and QoE(Quality of Experience)under different network conditions.Secondly,based on the SDK(Software Development Kit)provided by the open source player VLC multimedia player,this paper implements an end-to-end local area network player that supports video frame rate and bit rate adaptation-libvlc.Aiming at the ERP(Equi-Rectangular Projection)projection format of 360°video,this paper proposes a 3DoF(3-Degree of Freedom)for the selected user viewport of the user viewport.Tile selection algorithm and spatial weight greedy algorithm for tile bitrate allocation.The tile selection algorithm of the user’s viewport uses simple geometric relationships to determine the tiles covered by the FoV(Field of View),and the spatial weight greedy algorithm uses the human eye’s sensitivity to video content to allocate limited bandwidth to more important in the area.Theoretical analysis shows that the bandwidth utilization of the space weight greedy algorithm is close to that of the MCKP(Multiple-Choice Knapsack Problem)algorithm optimized for bandwidth.In terms of objective evaluation indicators,WS-PSNR,S-PSNR and WS-SSIM of the spatial weight greedy algorithm all perform better than MCKP.