TY - GEN
T1 - A Highly Scalable Hardware HEVC Encoder Based on FPGA
AU - Xu, Guohao
AU - He, Chenlong
AU - Yi, Shiyan
AU - Huang, Leilei
AU - Zeng, Xiaoyang
AU - Fan, Yibo
N1 - Publisher Copyright:
© 2024 IEEE.
PY - 2024
Y1 - 2024
N2 - The proposal of HEVC greatly reduces the bandwidth of video transmission and video storage. Oriented to different scenarios, the requirements of resources, throughput, and compression rate of hardware encoder are different. Therefore, we design a highly scalable HEVC encoder consisting of top-level and module-level and these scalable features enable the encoder to be flexibly changed to fit different devices and scenarios. Furthermore, we propose ping-pong address mapping and shadow registers to improve the overall system's throughput. Finally, we implement the two extreme profiles: best compression performance and best hardware performance. Experiments show that these two profiles achieve 1080p@60/1080p@30 fps with 128K and 405K LUTs in Xilinx ZCU102 and Xilinx Alveo U250, respectively. The BD-Rate compared to x265 achieves - 9.13%/14.45% respectively. To the best of our knowledge, this paper is the first to present a scalable architecture.
AB - The proposal of HEVC greatly reduces the bandwidth of video transmission and video storage. Oriented to different scenarios, the requirements of resources, throughput, and compression rate of hardware encoder are different. Therefore, we design a highly scalable HEVC encoder consisting of top-level and module-level and these scalable features enable the encoder to be flexibly changed to fit different devices and scenarios. Furthermore, we propose ping-pong address mapping and shadow registers to improve the overall system's throughput. Finally, we implement the two extreme profiles: best compression performance and best hardware performance. Experiments show that these two profiles achieve 1080p@60/1080p@30 fps with 128K and 405K LUTs in Xilinx ZCU102 and Xilinx Alveo U250, respectively. The BD-Rate compared to x265 achieves - 9.13%/14.45% respectively. To the best of our knowledge, this paper is the first to present a scalable architecture.
KW - FPGA
KW - HEVC
KW - pingpong address mapping
KW - scalable hardware encoder
UR - https://www.scopus.com/pages/publications/85218133304
U2 - 10.1109/ICSICT62049.2024.10831802
DO - 10.1109/ICSICT62049.2024.10831802
M3 - 会议稿件
AN - SCOPUS:85218133304
T3 - 2024 IEEE 17th International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2024
BT - 2024 IEEE 17th International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2024
A2 - Ye, Fan
A2 - Zhu, Xiaona
A2 - Tang, Ting Ao
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 17th IEEE International Conference on Solid-State and Integrated Circuit Technology, ICSICT 2024
Y2 - 22 October 2024 through 25 October 2024
ER -