Performance analysis of frame partitioning in parallel HEVC encoders

详细信息    查看全文

• 作者：H. Migallón ; P. Piñol ; O. López-Granado ; V. Galiano…
• 关键词：Tiles ; HEVC ; Video coding ; Parallel algorithms ; Multicore ; Performance
• 刊名：The Journal of Supercomputing
• 出版年：2017
• 出版时间：January 2017
• 年：2017
• 卷：73
• 期：1
• 页码：543-556
• 全文大小：
• 刊物类别：Computer Science
• 刊物主题：Programming Languages, Compilers, Interpreters; Processor Architectures; Computer Science, general;
• 出版者：Springer US
• ISSN：1573-0484
• 卷排序：73

文摘

The new video coding standard HEVC includes two concepts that allow to partition a frame into regions that can be independently encoded and decoded. These two concepts are named “Tiles” and “Slices”. In this paper, we present and analyze optimized parallel versions of the HEVC encoder based on tile and slice partitioning. We have evaluated the benefits and drawbacks of both approaches in terms of computational times and rate distortion performance. The results show that both approaches obtain good speed-ups, being the parallel version based on tiles the one that obtains the best trade-off between speed-up achieved (up to 9.3\(\times \)) and rate distortion performance loss (1.6% BD rate for AI mode and 2.2% for LB mode on average).