MPEG-4為MPEG在1999年為了新一代多媒體需求所提出的影像壓縮標準，為了新的影像傳輸媒介以及應用制定了許多新的功能。XviD則是在目前所有的MPEG-4編碼器當中相當知名的一個，其提供對於MPEG-4的高度支援以及開放源碼的特性使其在不論是商業或是學術研究用途上皆受到相當的歡迎。因此本論文在將程式移植至 BlackFin DSP平台外，亦利用 BlackFin DSP 所擁有的特性將效能做改進。 本論文在 ADDS-BF533-stamp 開發板上，實現 XviD 編碼器的完整功能，並針對動態估計（motion estimation）及離散餘弦轉換（DCT）作最佳化。此平台以 ADSP-BF533 Blackfin 做為處理器，採用 Blackfin 處理器有下列的優點： Blackfin 處理器支援 uClinux 作業系統，程式的移植性高； Blackfin 處理器亦提供許多適合影像運算的指令集，如影像像素運算指令 (Video Pixel Operations)、向量運算 (Vector Operations) 等。此外 BlackFin 處理器具有可平行發出（parallel issue）指令的能力，開發者可以將處理器內部多個MAC及ALU做最有效率的使用。 本論文的研究目的，在於將完整的 XviD 編碼器移植到 BlackFin DSP的平台，並能夠透過使用 Blackfin 處理器的特殊指令以及平行發出指令的功能，充份發揮 Blackfin 處理器的效能，縮短程式執行的時間，以得到最佳的使用效率。

For the new multimedia applications and wireless transmission, MPEG (Motion Picture Experts Group) introduced the MPEG-4 standard in 1999. In these years, the encoders that support MPEG-4 are getting mature. The XviD MPEG-4 codec is one of the most popular encoders. This dissertation presents the implementation of XviD MPEG-4 codec on the ADI BF533 Blackfin DSP platform which has RISC-like ease-of-control architecture and computation power of DSP. Blackfin DSP has dual MACs and dual 40-bit ALUs, which can provide the SIMD (Single Instruction Multiple Data) instructions called vector operations and video pixel operations, and the ability of issuing parallel instructions which give efficient way for computation during the encoding process. Based on the Blackfin DSP architecture, the XviD core library was fully implemented and the performance was increased by using Blackfin special instructions. In this implementation, the DCT and the ME (motion estimation) computation were accelerated by using assembly code instead the original C code. In the performance measurement, the assembly approach speeds up by 36.57% compare to pure C approach.

[1] J. R. Jain and A.K. Jain, “Displacement Measurement and Its Application in Interframe Image Coding,” IEEE Transactions On Communications, Vol. COM-29, Dec. 1981, pp.1799-1808.
[2] T. Koga, K. Iinuma, A. Hirano, Y. Iijima, and T. Ishiguro, “Motion Compensated Interframe Coding for Video Conferencing,” Proc. National Telecommunication Conf., New Orleans, LA, Dec. 1981, pp. G5.3.1-G5.3.5.
[3] S. Zhu and K. K. Ma, “A New Diamond Search Algorithm for Fast Block Matching Motion Estimation,” Proc. of International Conference on Information Communications and Signal Processing, Vol. 1, 1997, pp.292-296.
[4] Wen-Hsiung Chen, C. Harrison Smith, and S. C. Fralick, “A Fast Computational Algorithm for the Discrete Cosine Transform,” IEEE Transactions on Communications, Vol. COM-25, Sep. 1977, No. 9.
[5] Christoph Loeffler, Adriaan Ligtenberg, and George S. Moschytz, “Practical, Fast 1-D DCT Algorithms with 11 Multiplications,” Proc. International Conf. on Acoustics, Speech, and Signal Processing, 1989, pp. 988-991.
[6] Analog Device Inc., ADSP-BF533 Blackfin Processor Hardware Reference, Mar. 2003.
[7] Analog Device Inc., Blackfin Processor Instruction Set Reference, May 2003 [8] Ravi K. Kolagotla, Jose Fridman, Bradley C. Aldrich, Marc M. Hoffman, William C. Anderson, Michael S. Allen, David B. Witt, Randy R. Dunton, and Lawrence A. Booth, Jr., “High Performance Dual-MAC DSP Architecture,” IEEE Signal Processing Magazine, Jul. 2002, pp. 42-53.
[9] ISO/IEC JTC1/SC/WG11, N2502a, “Generic Co of Audio-Visual Objects: Visual 14 496-2,” Atlantic City, Final Draft IS, Dec. 1998. 65
[10] J. L. Mitchell, W. B. Pennebaker, C. E fogg, D. J. LeGall, “MPEG video standard,” Chapman & Hall, 1997.
[11]張皓傑, “MPEG-4視訊編解碼系統之架構設計,” 台灣大學碩士論文, 2001.
[12] Blackfin uClinux網站: www.blackfin.uclinux.org
[13] XviD 官方網站: www.xvid.org
[14] ADI Blackfin processor 網站: www.analog.com/processors/processors/blackfin/