臺灣博碩士論文加值系統

MPEG-1/Audio標準規格提供了對高品質數位音樂的編解碼方法，其中又以MPEG-1/Audio第三層編碼法（MP3）在相同位元傳率下的效果最佳，其複雜度亦最高。
本論文主要探討MP3的編碼法，並且期望在一16位元定點運算之數位訊號處理器（DSP）上實現MP3編碼器。文中介紹MP3音訊編碼標準的原理與實現方法，並且說明在實作上為配合DSP有限的資源所做的種種改善，包含了定點數的運算與減少記憶體空間使用之方法。

The MPEG-1/Audio standard provides a method for the compression of high-quality digital audio, especially the MPEG-1/Audio layer3. It performed the best quality of the three layers but also needs the most computational complexity.
This thesis is to investigate the MP3 encoding algorithm and expects to implement MP3 encoder on a 16-bit, fixed-point digital signal processor (DSP). This thesis introduces the principles of MP3 encoding standard and the implementation of MP3 encoder. Because of the limited resource of the DSP, this thesis also describes some modified methods for the DSP, including the fixed-point operation and memory space reducing.

第一章 簡介 1
1.1.0 研究背景與動機 1
1.2.0 章節安排 2
第二章 MP3編碼原理 3
2.1.0 濾波器排（The Filter Bank） 4
2.1.1 分析子頻帶濾波器（Analysis Subband Filter） 4
2.1.2 改良式離散餘弦轉換（MDCT） 7
2.2.0 聲響心理模型（The Psychoacoustic Model） 9
2.2.1 靜音門檻曲線（The Absolute Threshold of Hearing） 10
2.2.2 臨界頻帶（Critical Bands） 11
2.2.3 頻域上的遮蔽效應（Frequency masking） 13
2.2.4 時軸上的遮蔽效應（Temporal Masking） 15
2.2.5 Perceptual Entropy 17
2.2.6 第二聲響心理模型 19
2.3.0 位元分配及量化（Bit Allocation and Quantization） 21
2.3.1 位元分配(Bit allocation) 21
2.3.2 非均勻量化(Nonuniform quantization) 23
2.3.3 霍夫曼編碼(Huffman Encoding) 25
2.4.0 位元串格式（Bitstream Formatting） 27
2.4.1 位元儲藏處(bit reservoir) 28
2.4.2 MPEG-1/Audio的檔頭格式 29
第三章 MP3編碼實作 32
3.1.0 定點數運算 33
3.2.0 分析子頻帶濾波器 35
3.3.0 改良式餘弦轉換 39
3.4.0 第二聲響心理模型 41
3.5.0 量化和位元分配 46
3.5.1 Iteration Loop 46
3.5.2 外部迴圈 47
3.5.3 內部迴圈 50
第四章 結果與未來展望 54
參考資料 57

[1] ISO/IEC JTC1/SC29/WG11 MPEG, IS11172-3 “Information Technology - Coding of Moving Pictures and Associated Audio for Digital Storage Media at up to About 1.5Mbit/s, Part 3: Audio” 1992.
[2] ISO/IEC 13818-7 “Information Technology — Generic Coding of Moving Pictures and Associated Audio Information, Part 7: Advanced Audio Coding” 1997.
[3] Davis Pan, “A Tutorial on MPEG/Audio Compression,” IEEE Multimedia, Vol. 2, No. 2, Summer 1995.
[4] Ted Painter, “Perceptual Coding of Digital Audio,” Proceedings of IEEE, Vol. 88, No. 4, April 2000.
[5] Charles D.Murphy ,and K.Anandakumar , ”Real-Time MPEG-1 Audio Coding and Decoding on a DSP Chip,“ IEEE transactions on Consumer Electronics ,Vol. 43, No.1 ,FEBRUARY 1997.
[6] Noll, P. ,”MPEG Digtal Audio Coding,” IEEE Signal Processing Magazine , Volume: 14 Issue: 5 , Sept. 1997
[7] TMS320VC5416 Fixed-Point Digital Signal Processor Data Manual, Ti, 2001.
[8] Hung-Chih Lai, “Real-Time Implementation of MPEG-1 Layer 3 Audio Decoder on a DSP Chip,” National Chiao Tung University, June 2001.
[9] E. Zwicker and H. Fastl, Psychoacoustics Facts and Models. Berlin, Germany: Springer-Verlag, 1990.
[10] E. Terhardt, “Calculating virtual pitch,” Hearing Res., vol. 1, pp.155—182, 1979.
[11] Lin Pei-Ying, “Audio Coding Using Multi-Subband Hybrid Trees,” National Chiao Tung University, June 1998.
[12] M. Schroeder, B. S. Atal, and J. L. Hall, “Optimizing digital speech coders by exploiting masking properties of the human ear,” J. Acoust. Soc. Amer., pp. 1647—1652, Dec. 1979.
[13] J. Johnston, “Estimation of perceptual entropy using noise masking criteria,” in Proc. ICASSP-88, May 1988, pp. 2524—2527.
[14] TMS320C54XX Evaluation Module Technical Reference, Ti, 2001.