MPEG video coding with temporal scalability
Huifang Sun, Wilson Kwok · 2002
MPEG video coding standards have found extensive applications in the areas of video storage and transmission. The scalability features of MPEG-2 allow for layered coding, so that a base layer bitstream can be generated to provide a basic quality or basic spatial/temporal resolution output for some applications which are constrained by limited channel rate or display environment. An additional enhancement layer bitstream can be generated to augment the basic layer service to provide higher quality or resolution. We present the basic concept of MPEG-2 temporal scalability and provide some simulation results to indicate the typical figures for image quality loss due to the overhead imposed by temporal scalability. An optimum joint rate control algorithm which allows for dynamic allocation of bits between the base layer and the enhancement layer is proposed and simulated. This joint rate control algorithm codes the two layers together with a fixed total bit rate, while each individual layer is adaptively assigned bits according to scene content to achieve near-equal quality in the two layers. In this scheme, the two individual layers are variable bit rate processes which add together to form a constant bit rate process.