Cutting Across Visual MPEG Standards for Video Scalability

T. Shanableh (UAE)


Scalable video coding, MPEG-A and video transcoding..


2. Illustration of the proposed solutionThis paper proposes to replace the Intraframes of MPEG 2 with JPEG2k coded bitstreams governed by the syntax of MPEG-2 spatial scalability. The paper elaborates upon the proposed coding architectures and compatibility with published MPEG standards. The proposed coding architecture is compared against counterpart video transcoders, scalable and single layer codecs. It is shown that the proposed spatial encoder is either comparable or superior to all the aforementioned codecs whilst enjoying an extensible architecture capable of serving restricted end systems and hosting additional scalable layers such as SNR and temporal scalabilities. The basic concept of the proposed solution is to replace the periodic I-frames of an MPEG-2 coder with JPEG2K embedded images. Recall that according to the MPEG-2 video specification, the base layer format does not have to be MPEG-2 compatible thus providing interoperability between standards [5]. Additionally, JPEG2K generates embedded bitstreams with various spatial and/or SNR resolutions. This can be thought of as a feature that complements MPEG-2 scalability. By using MPEG-2 spatial scalability headers, I-frames can be coded using spatial prediction from the lower layer whilst setting the prediction errors to nil i.e. using skipped macroblocks. Hence only the headers of the spatial I frames are transmitted in the MPEG-2 video stream. The base layer on the other hand uses JPEG2K to encode the content of I-frames. Since a typical MPEG-2 coded video has a GOP structure of N=12 and M=3 (N being the GOP size and M the sub GOP size) it follows that roughly 2

Important Links:

Go Back