臺灣博碩士論文加值系統

在這篇論文中，我將會詳細地介紹三種不一樣的作音樂相似度分析的方式。接著利用自己選出來的50首歌(粗略分為五個種類，每一個種類有十首歌)，依序用這三種方式來對它們做相似度的計算。比較計算出來的結果，我們可以得知哪一種方法最具正確性、最有效，可以將不同種類的音樂分得比較清楚。
因為一個音樂的檔案可以用多種不同的角度來看它，我們必須先決定要比較的是音樂在哪一種特徵上的相似度。在這篇論文中，我選擇「音色」作為我的分類標準。在第二章中，我會對音樂的「音色」這個特徵，做比較詳細的描述。其他的特徵諸如旋律、節奏等等也會稍微提到。
在第三章中，將會說明在語音辨識中被廣泛應用的「梅爾-倒頻譜係數」。另外一種類似「梅爾-倒頻譜係數」的方式--「宋(量測聲音大小的單位)-巴克表示法」也會被一起提到。他們是第四章所要說明的三種方法的基礎。
第四章提到三種比較音樂相似度的方式，也就是頻譜長條圖、週期性長條圖、以及抖動模式。第五章會把這三種方式應用到器樂上，藉以得知哪一種方式在分辨不同的樂器上有最好的效果。第六章則是結論，以及未來可以繼續進行研究的方向。

In this thesis, I introduce three different methods for music similarity measuring in detail. And then I apply these three methods to 50 songs I chose (frankly viewed as 5 groups, each one has 10 songs). By comparing these results, we can understand which method is the most effective and most accurate.
Because we can view an audio data in many different views of point, we have to choose a property on which we focus before starting the similarity measures. In this thesis, I choose the property “timbre” as my yardstick. In chapter 2, I describe the property “timbre” in detail. Other properties such as melody, rhythm, and genre are referred to. Based on these properties, the very first step of measuring similarity of music is constructed.
In chapter 3, some backgrounds of building these similarity measure methods are discussed in detail, e.g. Mel-Frequency Cepstrum Coefficient and Sone-Bark Representation. In chapter 4, three methods for measuring the music similarity are particularly described. They are spectrum histogram (SH), periodicity histogram (PH), and fluctuation pattern (FP).
In Chapter 5, SH, PH, and FP will be applied to instrumental music. We can know the best method to distinguish different instruments. Chapter 6 is about conclusion and future works.

CONTENTS

口試委員會審定書 #
誌謝 i
中文摘要 iii
ABSTRACT v
CONTENTS vii
LIST OF FIGURES xi
LIST OF TABLES xiii
Chapter 1 Introduction 1
Chapter 2 Timbre, Rhythm, Melody, and Genre of Music/Audio 3
2.1 Introduction 3
2.2 Some Insights of Timbre 3
2.2.1 Introduction 3
2.2.2 The Characterization and Features of Sounds 5
2.2.3 John Grey’s Timbre Space 6
2.2.4 Tools of Analysis 7
2.2.5 Experiments 9
2.2.6 Quantitative Analysis 13
2.2.7 Conclusion 15
2.3 Melody 16
2.4 Rhythm 18
2.5 Genre 21
2.6 Conclusion 24
Chapter 3 Mel-Frequency Cepstrum Coefficients (MFCCs) and Sone/Bark Representation 27
3.1 Introduction 27
3.2 Mel-Frequency Cepstrum Coefficient (MFCC) 27
3.2.1 Mel Scale 27
3.2.2 Mel-scale Filter Bank 29
3.2.3 Cepstrum 29
3.2.4 Mel Frequency Cepstrum Coefficient 31
3.3 Sone/Bark Representation 32
3.3.1 Psychoacoustic Preprocessing 32
3.3.2 Plots of Different Time/Frequency/Loudness Representations [4] 35
3.4 Conclusion 36
Chapter 4 Spectrum Histogram (SH), Periodicity Histogram (PH), and Fluctuation Pattern (FP) 39
4.1 Introduction 39
4.2 Spectrum Histogram (SH) 40
4.3 Periodicity Histogram (PH) 43
4.4 Fluctuation Pattern (FP) 45
4.5 An Example about Similarity Measures 49
4.6 Conclusion 50
Chapter 5 Comparison of Instrumental Music Using Spectrum Histogram, Periodicity Histogram, and Fluctuation Pattern 53
5.1 Introduction 53
5.2 Example 1: 50 10-sec Files Divided into 5 Groups 54
5.3 Example 2: 30 10-sec Files Divided into 3 Groups 55
5.4 Example 3: 30 20-sec Files Divided into 3 Groups 56
5.5 Example 4: 30 10-sec Files (Manually Selected) Divided into 3 Groups 57
5.6 Conclusion 58
Chapter 6 Conclusion and Future Works 61
6.1 Conclusion 61
6.2 Future Works 62
REFERENCE 63

[1]E. Pampalk, “A Matlab Toolbox to Compute Similarity from Audio,” Proceedings of the ISMIR International Conference on Music Information Retrieval (ISMIR''04), Barcelona, Spain, 2004.
[2]E. Pampalk, S. Dixon, and G. Widmer, “Exploring Music Collections by Browsing Different Views,” Proceedings of the ISMIR International Conference on Music Information Retrieval (ISMIR''03), pp 201-208, Baltimore, MD, October 26-30, 2003.
[3]E. Pampalk, A. Rauber, and D. Merkl, “Content-based Organization and Visualization of Music Archives,” In the Proceedings of the ACM Multimedia 2002, Juan les Pins, France, December 1-6, 2002
[4]E. Pampalk, S. Dixon, and G.Widmer, “On the evaluation of perceptual similarity measures for music,” in Proc of DAFx, 2003.
[5]J.-J. Aucouturier and F. Pachet, “Improving Timbre Similarity: How high is the sky?,” Journal of Negative Results in Speech and Audio Sciences, 1(1), 2004.
[6]E. D. Scheirer, “Tempo and beat analysis of acoustic musical signals,” JASA, vol. 103, no. 1, 1998.
[7]J. J. Højvang, C. Mads G, and J. S Holdt, “A Framework for Analysis of Music Similarity Measures,” Proceedings of the 15th European Signal Processing Conference, pp. 926-930, 2007.
[8]P. Prandoni, http://lcavwww.epfl.ch/~prandoni/documents/timbre1.pdf, 31 July 1994.
[9]DeLone et al. (Eds.) (1975). Aspects of Twentieth-Century Music, chap. 4, p.270-301. Englewood Cliffs, New Jersey: Prentice-Hall
[10]Holt, Fabian (2007). Genre in Popular Music. Chicago: University of Chicago Press.
[11]Negus, Keith (1999). Music Genres and Corporate Cultures. New York: Routledge.
[12]E. D. Scheirer, “Tempo and beat analysis of acoustic musical signals,” JASA, vol. 103, no. 1, 1998.