CRI: CRD: VIDEO TRACES II - Designing, Creating, and Disseminating Video Traces for Networking Research

视频编码和网络的许多新发展正在显著地改变多媒体网络，包括：(I)通过IP网络的HDTV分辨率视频传输，(Ii)随着H.264/AVC(高级视频编码)编解码器的可伸缩性扩展而引入的全新的可伸缩编码范例，(Iii)新的容错技术，以及(Iv)UMA(通用媒体接入)范例，其允许在内容表征和视频传输的适应性方面具有前所未有的灵活性。这些新的编码技术为新的视频传输机制开辟了广阔的研究和开发空间，并可能在现有机制的基础上取得突破性进展。为了能够探索这一研究空间，我们(A)设计新的视频编码范例的踪迹特征，(B)为这些新的编解码器开发创建和传播广泛的公开可用的视频踪迹库，(C)对踪迹中捕获的视频流量和质量以及视频内容特征进行统计分析，以及(D)制定踪迹库的使用指南。该社区资源开发项目的智力优点是设计了踪迹表征，捕获视频的流量、质量和内容特征，直到使用视频踪迹中的新的可伸缩性、容错性和UMA范例编码的HDTV分辨率。提供这些痕迹以及对流量、质量和内容特征的统计分析将为网络、系统和视频研究提供信息并推动其发展。更广泛的影响是以免费和公开的形式提供这些复杂的痕迹和分析结果。特别是，该视频跟踪库使网络研究团体能够跟上视频编码领域的发展步伐，并提供反映将主导未来互联网和无线网络的视频编码的属性的跟踪。视频跟踪库和附带的跟踪使用指南/教程使没有深厚视频编码专业知识和/或没有最新设备的网络研究人员能够对各种关键多媒体网络项目进行及时、高质量的研究，从而扩大研究基础。

A number of new developments in video coding and networking are dramatically changing multimedia networking, including: (i) HDTV resolution video transmission over IP networks, (ii) fundamentally new scalable coding paradigms introduced with the scalability extensions of the H.264/AVC (Advanced Video Coding) codec, (iii) novel error resilience techniques, and (iv) the UMA (Universal Media Access) paradigm, which allows for unprecedented flexibility in content characterization and adaptivity in video transport. These new coding techniques open a vast research and development space for novel video transport mechanisms and make possible groundbreaking advances over existing mechanisms. To enable the exploration of this research space we (a) design trace characterizations of the new video coding paradigms, (b) create and disseminate an extensive publicly available library of video traces for these new codec developments, (c) conduct statistical analyses of the video traffic and quality and the video content features captured in the traces and (d) formulate usage guidelines for the trace library.The intellectual merit of this community resource development project is the design of trace characterizations capturing the traffic, quality, and content features of video up to HDTV resolution coded with the new scalability, error-resilience, and UMA paradigms in video traces. The provision of both these traces and of the statistical analysis of the traffic, quality, and content characteristics will inform and advance networking, systems and video research.The broader impact is provision of these sophisticated traces and analytical results in free and publicly available form. In particular, this video trace library allows the networking research community to keep pace with the developments in the video coding area and provides traces that reflect the properties of the video codings that will dominate the future Internet and wireless networks. The video trace library and the accompanying trace usage guidelines/tutorials enable and educate networking researchers without deep video coding expertise and/or without up-to-date equipment to conduct timely, high-quality research on a wide range of critical multimedia networking projects and hence to expand the research base.