为了保证一定视频质量下编码器输出的码率符合给定的目标码率,提出一种考虑视频内容特性的H.265/HEVC帧层码率分配算法。首先从率失真理论分析的角度推导出影响输出码率的主要因素,然后根据视频编码原理,预测出帧内容复杂度参数,从而建...为了保证一定视频质量下编码器输出的码率符合给定的目标码率,提出一种考虑视频内容特性的H.265/HEVC帧层码率分配算法。首先从率失真理论分析的角度推导出影响输出码率的主要因素,然后根据视频编码原理,预测出帧内容复杂度参数,从而建立一种更有效的帧层码率分配算法。实验结果表明,所提算法可以使帧层目标码率与编码码率保持更好的一致性,且重构视频质量平均提高了0.103 d B。展开更多
Video transcoding is to create multiple representations of a video for content adaptation.It is deemed as a core technique in Adaptive BitRate(ABR)streaming.How to manage video transcoding affects the performance of A...Video transcoding is to create multiple representations of a video for content adaptation.It is deemed as a core technique in Adaptive BitRate(ABR)streaming.How to manage video transcoding affects the performance of ABR streaming in various aspects,including operational cost,streaming delays,Quality of Experience(QoE),etc.Therefore,the problems of implementing video transcoding in ABR streaming must be systematically studied to improve the overall performance of the streaming services.These problems become more worthy of investigation with the emergence of the edge-cloud continuum,which makes the resource allocation for video transcoding more complicated.To this end,this paper provides an investigation of the main technical problems related to video transcoding in ABR streaming,including designing a rate profile for video transcoding,providing resources for video transcoding in clouds,and caching multi-bitrate video contents in networks,etc.We analyze these problems from the perspective of resource allocation in the edge-cloud continuum and cast them into resource and Quality of Service(QoS)optimization problems.The goal is to minimize resource consumption while guaranteeing the QoS for ABR streaming.We also discuss some promising research directions for the ABR streaming services.展开更多
为解决多客户端的带宽资源分配问题,提高用户体验质量(quality of experience,QoE),建立多客户端视频流的体验质量优化框架。针对已有视频流算法在多客户端领域的缺陷,基于模型预测控制算法提出一个多客户端带宽动态调度算法,根据每个...为解决多客户端的带宽资源分配问题,提高用户体验质量(quality of experience,QoE),建立多客户端视频流的体验质量优化框架。针对已有视频流算法在多客户端领域的缺陷,基于模型预测控制算法提出一个多客户端带宽动态调度算法,根据每个客户端的带宽预测情况对它们进行动态资源分配,通过提高带宽利用率进而提升总体用户QoE。在HSDPA网络带宽轨迹上的仿真结果表明,相比各客户端平均带宽分配方式,优化方法在总体用户体验质量上提升42.6%以上,相比最新的Minerva方案提升了7.8%。展开更多
六自由度(Six Degree of Freedom,6DoF)视频系统允许用户从全方位、以任意视角身临其境地体验场景,是沉浸式视频技术的发展方向。根据6DoF视频系统中用户观看位置的变化,对多视点彩色与深度视频的码率分配是高质量场景生成的关键。本文...六自由度(Six Degree of Freedom,6DoF)视频系统允许用户从全方位、以任意视角身临其境地体验场景,是沉浸式视频技术的发展方向。根据6DoF视频系统中用户观看位置的变化,对多视点彩色与深度视频的码率分配是高质量场景生成的关键。本文从虚拟视点的失真出发,提出一种基于虚拟视点质量预测(Quality prediction model of virtual view,QPMVV)模型的视点级码率分配方法。首先理论分析了彩色和深度视频编码失真和虚拟视点失真的关系,然后通过实验统计分析了虚拟视点质量与彩色和深度视频编码量化参数的关系,建立了多视点彩色和深度视频的QPMVV模型,最后推导出多视点彩色和深度视频的相关视点的码率分配比例。实验表明,与平均分配的方法相比,本文码率分配方法能显著提升虚拟视点的主观和客观质量。虚拟视点越偏离中心位置,质量改善越明显。展开更多
文摘为了保证一定视频质量下编码器输出的码率符合给定的目标码率,提出一种考虑视频内容特性的H.265/HEVC帧层码率分配算法。首先从率失真理论分析的角度推导出影响输出码率的主要因素,然后根据视频编码原理,预测出帧内容复杂度参数,从而建立一种更有效的帧层码率分配算法。实验结果表明,所提算法可以使帧层目标码率与编码码率保持更好的一致性,且重构视频质量平均提高了0.103 d B。
基金supported in part by the Natural Science Foundation of Jiangsu Province under Grant BK20200486.
文摘Video transcoding is to create multiple representations of a video for content adaptation.It is deemed as a core technique in Adaptive BitRate(ABR)streaming.How to manage video transcoding affects the performance of ABR streaming in various aspects,including operational cost,streaming delays,Quality of Experience(QoE),etc.Therefore,the problems of implementing video transcoding in ABR streaming must be systematically studied to improve the overall performance of the streaming services.These problems become more worthy of investigation with the emergence of the edge-cloud continuum,which makes the resource allocation for video transcoding more complicated.To this end,this paper provides an investigation of the main technical problems related to video transcoding in ABR streaming,including designing a rate profile for video transcoding,providing resources for video transcoding in clouds,and caching multi-bitrate video contents in networks,etc.We analyze these problems from the perspective of resource allocation in the edge-cloud continuum and cast them into resource and Quality of Service(QoS)optimization problems.The goal is to minimize resource consumption while guaranteeing the QoS for ABR streaming.We also discuss some promising research directions for the ABR streaming services.
文摘为解决多客户端的带宽资源分配问题,提高用户体验质量(quality of experience,QoE),建立多客户端视频流的体验质量优化框架。针对已有视频流算法在多客户端领域的缺陷,基于模型预测控制算法提出一个多客户端带宽动态调度算法,根据每个客户端的带宽预测情况对它们进行动态资源分配,通过提高带宽利用率进而提升总体用户QoE。在HSDPA网络带宽轨迹上的仿真结果表明,相比各客户端平均带宽分配方式,优化方法在总体用户体验质量上提升42.6%以上,相比最新的Minerva方案提升了7.8%。
文摘六自由度(Six Degree of Freedom,6DoF)视频系统允许用户从全方位、以任意视角身临其境地体验场景,是沉浸式视频技术的发展方向。根据6DoF视频系统中用户观看位置的变化,对多视点彩色与深度视频的码率分配是高质量场景生成的关键。本文从虚拟视点的失真出发,提出一种基于虚拟视点质量预测(Quality prediction model of virtual view,QPMVV)模型的视点级码率分配方法。首先理论分析了彩色和深度视频编码失真和虚拟视点失真的关系,然后通过实验统计分析了虚拟视点质量与彩色和深度视频编码量化参数的关系,建立了多视点彩色和深度视频的QPMVV模型,最后推导出多视点彩色和深度视频的相关视点的码率分配比例。实验表明,与平均分配的方法相比,本文码率分配方法能显著提升虚拟视点的主观和客观质量。虚拟视点越偏离中心位置,质量改善越明显。
