A scalable MPEG-2 encoding system using the multithread architecture

Abstract

MPEG-2 video coding standard is well adopted as the video compression method for various multimedia applications, such as the DVD, video-on-demand, digital video library, etc. MPEG-2 video encoders are now available in a variety of forms using both hardware and software based approaches. Software based approach potentially offers a better picture quality but is computationally quite intensive. With a software based approach, one can either achieve a better visual quality at the same bit rate, or, alternatively, reduce the bit rate for the desired level of picture quality. There exist a number of techniques to speedup the MPEG-2 video encoding using a parallel machine or a network of workstations. These approaches promise good concepts but do not offer practical solutions due to factors such as cost, size, etc. In this work, we propose a new approach with the aim to build a cost-effective and complete practical solution that is not only highly efficient but is also scalable in that it can be used on a single-processor or multiple-processor PC. We develop a multithreaded encoder that uses temporal parallelism with small overhead. Further, we use various optimizations to enhance the coding speed. The proposed encoder achieves faster than real-time and half of real-time encoding rate for CIF (352 x 288) and CCIR601 (720 x 576) video sequences, respectively.

Date of Award	2000
Original language	English
Awarding Institution	The Hong Kong University of Science and Technology