System Modeling and Software Implementation of MPEG-4 Video Encoder Literature Survey For EE382C Embedded Software Systems

MPEG-4 standard provides support for content-based interactivity, high compression, and/or universal accessibility and portability of audio and video content. Due to its content-based representation nature (except the simple profile used for wireless video communication) and flexible configuration structure, any MPEG-4 hardware implementation is likely to be very application specific. Therefore, software-based implementation seems to be a natural and viable option. In addition, a software-based approach allows flexibility and portability, which are extremely desirable features for MPEG-4 based interactive multimedia systems. In this survey, we will review existing system-level modeling and software-based implementation approaches for real-time MPEG-4 codecs. It is the inherent parallelism and flexible configuration structure with the MPEG-4 encoder that motivate us to model the encoder using Process Networks (PN)-the concurrent computation model, and to implement a scalable software-based encoder under the framework proposed by Allen and Evans. We also intend to adopt the dynamic shape-adaptive data partitioning in each Video Object Plane (VOP) encoder to maximize the parallelism by dividing the tasks and balancing the load on each process node. Therefore, our implementation will utilize two levels of parallelism of the MPEG-4 encoder: control parallelism (VOP level) and data parallelism (Macro-Block level).