System modeling and software-based implementation of MPEG-4 video encoder

In this paper, we model the MPEG-4 video encoder using computational process networks (CPN), which is a determinate and concurrent computation model, and implement a scalable software-based encoder in C++ and portable operating system interface (POSIX) threads under the framework proposed by Allen and Evans (see IEEE Trans. Signal Proc., vol.48, no.3, p.921-6, Mar. 2000) for constructing high-throughput real-time signal processing applications. The performance of the original sequential implementation and our concurrent implementation of the MPEG-4 simple profile video encoder are compared on a single-processor workstation. We find that the concurrent implementation is faster since the benefits from concurrent execution of CPN nodes outweigh the overheads created by C++/Pthread programming. Our approach is scalable to multiprocessor environment, which makes the software-based real-time MPEG-4 video encoder on multiple or multiprocessor workstations feasible.