Logo image
Design and Implementation of a Cloud-based Full-HD H.264/SVC Live Streaming Broadcast Platform
Thesis

Design and Implementation of a Cloud-based Full-HD H.264/SVC Live Streaming Broadcast Platform

Hsu, Wei-Ting
Masters, 國立清華大學, 通訊工程研究所
2010

Abstract

即時影音串流 可調式編碼 雲端運算 Live Video Streaming Scalable Video Coding Cloud Computing
With the progression in technology, there are more and more requirements of real-time video streaming of different devices nowadays. Users may use different devices to watch the same content. For example, someone may seat at home and watch a MLB live show by a high-definition TV, but someone may seat at bus and watch the same MLB show by his own smart phone. Obviously, we should push the HD quality streaming to the HD TV, but push 360p or lower quality to smart phone. Therefore, we should supply the same content with different qualities at a time to serve different users in different situation. Scalable Video Coding (SVC) is an extension compression profile of ISO/IEC 14496 MPEG4 Part.10 H.264;it can supply different qualities by using different decoding parameters according to different situations. Our goal is to exert this feature of SVC to build such a real-time streaming broadcast platform which can supply different qualities of video streaming at a time. First of all, we divide a SVC streaming to many subsets streaming and forward them independently. This is a multi-tree relay structure. The advantage of this structure is we can decide how many relay tree we receive from depends on which quality we want to watch. Second, we do the SVC encoding and decoding in our platform so that streaming providers can push its raw video streaming to our platform easily and audiences can watch the streaming directly. Third, we use the concept of distributed computing to achieve full-HD 1080p SVC real-time en/decoding so that this platform can support full-HD live streaming broadcast. Eventually, we use Cloud computing to build our platform as a result of us want to implement a low cost and high scalability service. We expect this platform can broadcast the live contents with different qualities easily.

Metrics

1 Record Views

Details

Logo image