QoS Adaptive Streaming Based On Enhanced MPEG-2 Scaling

摘要

In this paper, we propose a novel framework, called the scalable media objects group (SMOG), for enhancing MPEG media scalability, and simple rate control algorithm utilizing the framework. In our framework, an MPEG-2 video stream is segmented into SMOG, which is a set of fine-grained bit-streams. During a streaming, a subset or the full set of the scalable media objects is dynamically selected based on QoS feedback information and delivered to the client. And then they are reassembled at the client to form an MPEG-2 conformance video bit-stream that can be decoded at any MPEG-2 decoder. Our approach provides the manifold degree of the service quality from a bit-stream pre-coded at high quality, as well as it can be easily extended to support the dynamic QoS adaptation for the continually QoS-varying channels. We shows some experimental results that the SMOG is working with even a simple rate control algorithm, and possibly more effective when it is used in conjunction with traditional TCP-friendly rate control schemes.