A Parallelism Encoding Framework for the Temporal Scalability of H.264/AVC Scalable Extension

Autor: Hong-Ming Chen, Sung-Wen Wang, Ja-Ling Wu, Shu-Sian Yang
Rok vydání: 2007
Předmět:
Zdroj: Ninth IEEE International Symposium on Multimedia Workshops (ISMW 2007).
DOI: 10.1109/ism.workshops.2007.72
Popis: The encoder complexity of H.264/MPEG-4 AVC scalable extension is much higher than that of original H.264/MPEG-4 AVC encoder. To make Scalable Video Coding techniques popular and practical, the ways for speeding encoding stage up becomes an important issue. Along with the increasing of clock frequency for a single processor, the benefit gained from sequential optimization technique could be further improved. Nevertheless, tradition sequential processing may not enhance the performance directly from increasing the number of cores. That is because the benefit of multi- core system can only be obtained by effective parallel processing. To deal with the prescribed high complexity, we propose a parallel encoding framework for the temporal scalability of H.264/MPEG-4 AVC scalable extension. In our experiments, the execution time can be reduced to 58.54% of that of the normal sequential processing under a quad-core processor. As compared with the most famous parallel encoding method, wavefront encoder, the proposed algorithm performs 5.9% better in execution time.
Databáze: OpenAIRE