MP3(MPEG-1/Audio Layer 3)和MPEG-4 先進音訊編碼(AAC)是由ISO/IEC MPEG (Moving Picture Expert Group)所制定的多媒體音訊壓縮標準，提供了非常優秀的壓縮效率。MP3是目前所使用最普遍的壓縮技術，而MPEG-4 AAC (Advanced Audio Coding)的基本架構和MPEG-2 AAC相同，只是多加入了一些工具來提高編碼的效率，以本論文會先介紹MPEG-1 Layer 3以及MPEG-2 AAC，最後再針對MPEG-4 AAC中的Low Complexity Profile實作來做說明。

越來越多音訊壓縮的解碼器在嵌入式系統中被廣泛的使用，大部分在這類電子產品中，會使用一個數位訊號處理器來加速解碼器中的運算。SCREAM DSP-16是一個由成功大學音樂與多媒體系統實驗室修改Analog Device DSP 2191指令集所開發的數位訊號處理器，本論文即提出了在SCREAM DSP-16上面實作MP3與MPEG-4 AAC解碼器的方法。

在DSP的實現部份，會遇到最多的問題就是在於計算複雜度以及如何使用有限的記憶體空間，在本論文中會針對計算複雜度高的部分使用組語來加速，以及利用SCREAM DSP-16所提供的加速法，和記憶體空間減少的方法來撰寫本論文。最後經由我們的加速以及系統的實現，MP3以及MPEG-4 AAC解碼器均可以在SCREAM DSP-16上達到即時解碼的效果。

MP3(MPEG-1/Audio Layer 3)and MPEG-4 Advanced Audio Coding (AAC) is the multimedia audio compression standard defined by ISO/IEC MPEG (Moving Picture Expert Group), and it provides an outstanding coding efficiency. Now, MP3 is the most popular audio compression codec around the world.The basic structure of MPEG-4 AAC is the same as that of MPEG-2 AAC, but it has many new tools to improve its coding efficiency. For example, this thesis will introduce MPEG-1 Layer 3 and MPEG-2 AAC, then selects the implementation of MPEG-4 AAC Low Complexity Profile.

There is wide demand for audio decoder built around embedded systems. In the most case, this kind of consumer electronics use a Digital Signal Processor to speed up the computation essential to decoder. SCREAM DSP-16 is a DSP based on Analog Device DSP 2191 instructions, and is developed by Scream LAB of National Cheng-Kung University. In this thesis, several techniques are proposed to implement MP3 decoder and MPEG-4 AAC decoder on
SCREAM DSP-16.

Computation complexity and how to use limited memory space are the most difficult problem we meet. In this thesis, we rewrite our source code in assembly language to speed up and use some technique to improve memory utilization. Finally, the speed of the MP3 decoder and MPEG-4 AAC decoder on our SCREAM DSP-16 implementation can achieve real time operation.

[1] ISO/IEC 11172-3: 1993, Information technology – Coding of moving pictures and associated audio for digital storage media at up to about 1.5 Mbit/s – Part 3:Audio.

[2] K. Brandenburg and G. Stoll, “ISO-MPEG-1 Audio:a generic standard for coding of high quality digitalaudio”, In N. Gilchrist and Ch. Grewin, editors, Collected Papers on Digial Audio Bit-Rate Reduction,pages 31 – 42. AES, 1996.

[3] Ted Painter, Student Member IEEE, and Andreas Spanias, Senior Member IEEE, “Perceptual Coding of Digital Audio”, Deapartment of Electrical Engineering, Telecommunications Research Center Arizona State University,

[4] ISO/IEC 13818–7: 1997, Information technology – Generic coding of moving pictures and associated audio information.

[5] M. Bosi, K. Brandenburg, Sch. Quackenbush, L. Fielder, K. Akagiri, H. Fuchs, M. Dietz, J. Herre, G. Davidson, and Yoshiaki Oikawa, “ISO/IEC MPEG-2 Advanced Audio Coding.”, in Proc. of the 101st AES Convention, pp. 789-814, 1996.

[6] ISO/IEC 14496–3: 2005, Information technology – Coding of audio-visual objects – Part 3 : Audio.

[7] Jurgen Herre, Donald Schulz, ”Extending the MPEG-4 AAC Codec by Perceptual Noise Substitution”, Presented at the 104th Convention 1998 May 16-19 Amsterdam, AES

[8] Donald Schulz, ”Improving Audio Codec by Noise Substitution”, Technische Hochschule, Fachgebiet Digitaltechnik, D-64283 Darmstadt, Germany, preprint 4720

[9] Naoki Iwakami and Takehiro Moriya, “Transform-Domain Weighted Interleave Vector Quantization”, Presented at the 101th Convention 1996 November 8-11 Los Angeles, California, AES, pp. 593-597

[10] Zhong-Ho Chen,“Portable DSP Development Environment For Embedded Real-Time System”, master thesis, National Cheng-Kung University, Taiwan, June, 2005.

[11] Kent Salomonsen, Sten Sogaard, Eddie Proft Larsen, “Design and Implementaion of MPEG/Audio Layer III Bitstream Processor”, Aalborg University

[12] Jiirgen Helve and James D. Johnston, “Enhancing the Performance of Perceptual Audio Coders by Using Temporal Noise Shaping (TNS)”, in Proc. of the 101st AES Convention, November 1996.

[13] John P. Pricen, Alan Bernard Bradley, “Analysis/Synthesis Filter Bank Design Basedo n Time Domain Aliasing Cancellation”, IEEE Trans on ASSP, Oct 1986.

[14] B. Edler, “Coding of Audio Signals with Overlapping Block Transform and Adaptive Window Functions”, Frequenz, vol. 43, pp. 252-256, 1989.

[15] Analog Device Ltd., ADSP-219x DSP Hardware Reference.

[16] Analog Device Ltd., ADSP-219x DSP Instruction Set Reference.

[17] Analog Devices : Embedded Processing & DSP: Third Party Developers page for company - Epigon Media Technologies Pvt. Ltd.,
http://dspcollaborative.analog.com/developers/DSP_CompanyListing.asp?c_id=30

[18] Adaptive Digital's DSP Algorithms and Solutions,
http://www.adaptivedigital.com/products/specs/aac_dec.htm

[19] Konstantinos Konstantinides, Member IEEE, “Fast Subband Filtering in MPEG Audio Coding”, pp. 26-28

[20] Vladimir Britanak, K. R. Rao, “An Efficient Implementation of the Forward and Inverse MDCT in MPEG Audio Coding”, IEEE signal processing letters, vol. 8, No. 2, February, 2001.

[21] A. V. Oppenheim, R. W. Schafer, J. R. Buck, Discrete-Time Signal Processing, Prentice Hall, Second Edition.

[22] AudioCoding.com, http://www.audiocoding.com/

[23] Analog Device Ltd., VisualDSP++ 3.5 C Compiler and Library Manual for ADSP-219x DSPs.

[24] Harmonic Structure Quad Tree Audio Codec,
http://scream.csie.ncku.edu.tw/~bff/HSQT.htm

[Figure 2.2] Human Hearing Range
http://sound.westhost.com/articles/fadb.htm
[Figure 2.3] Frequency Masking Effect
http://cs.uccs.edu/~cs525/audio/audio.html
[Figure 2.4] Temperal Masking Effect
參考資料中[3]論文中的圖片