DocumentCode :
3368281
Title :
Hardware architecture design for H.264/AVC intra frame coder
Author :
Huang, Yu-Wen ; Hsieh, Bing-Yu ; Chen, Tung-Chien ; Chen, Liang-Gee
Author_Institution :
Dept. of Electr. Eng., Nat. Taiwan Univ., Taipei, Taiwan
Volume :
2
fYear :
2004
fDate :
23-26 May 2004
Abstract :
In this paper, we contributed a VLSI architecture design for H.264/AVC intra frame coder. First, analysis of coding algorithm is provided by using a RISC model to obtain the proper degrees of parallelism under SDTV specification. Second, a two-stage macroblock pipelining is proposed to double the processing capability and hardware utilization. Third, Hadamard-based mode decision is modified as DCT-based version to reduce the 40% of memory access. To sum up, our system architecture achieves 215 times of speed compared with RISC-based software implementation in terms of processing cycles. In addition, we also made a lot of efforts on developing area-speed efficient modules. Reconfigurable intra predictor generator can support all kinds of prediction modes. Parallel multi-transform has four times throughput of the serial one with little area overhead. CAVLC engine can efficiently provide coding information for the bitstream packer. A prototype chip was fabricated with TSMC 0.25 μm CMOS technology and is capable of encoding 720×480 4:2:0 30 Hz video in real time at the working frequency of 54 MHz. The transistor count is 429 K, and the core size is only 1.855×1.885 mm2.
Keywords :
CMOS integrated circuits; Hadamard transforms; VLSI; discrete cosine transforms; encoding; integrated circuit design; pipeline processing; reduced instruction set computing; video coding; 0.25 micron; 30 Hz; 54 MHz; CAVLC engine; DCT based version; H.264/AVC intra frame coder; Hadamard based mode decision; RISC based software; SDTV specification; TSMC CMOS technology; VLSI architecture design; bitstream packer; coding algorithm; encoding; hardware architecture design; hardware utilization; parallel multitransform; prediction modes; processing cycles; prototype chip; reconfigurable intra predictor generator; system architecture; two stage macroblock pipelining; Algorithm design and analysis; Automatic voltage control; CMOS technology; Computer architecture; Engines; Hardware; Pipeline processing; Reduced instruction set computing; Throughput; Very large scale integration;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Circuits and Systems, 2004. ISCAS '04. Proceedings of the 2004 International Symposium on
Print_ISBN :
0-7803-8251-X
Type :
conf
DOI :
10.1109/ISCAS.2004.1329260
Filename :
1329260
Link To Document :
بازگشت