1. NCCU Academic Hub
  2. Publications
  3. College of Science
  4. Theses
  5. Item

Publications-Theses

Article View/Open

Publication Export

Google ScholarTM

NCCU Library

Citation Infomation

Related Publications in TAIR

題名 小波理論於曲風辨識上之應用
The Application of Wavelet Transform on Automatically Musical Genre Classification
作者 陳彥名
貢獻者 曾正男
陳彥名
關鍵詞 小波轉換
線性判別分析
離散餘弦轉換
決策樹
曲風辨識
日期 2015
上傳時間 1-Oct-2015 14:17:32 (UTC+8)
摘要 隨著科技的進步，網際網路已充斥在我們的生活之中。音樂也不再以硬體儲存的方式流傳（例如ＣＤ、黑膠唱片），而是轉變為數位音樂的方式，透由網路平台散播。許多數位音樂串流服務平台網站也如雨後春筍般誕生，例如iTunes、Spotify、Musicovery。加上文化水平的提升，音樂已是現代人生活之中，不可或缺的一部分。世界上的音樂難以計數，如何將音樂分門別類做好管理乃為現代商業應用的一個重要課題。因此，音樂曲風自動化辨識的技術確實為一個實用且難以迴避的課題。
     
     過去在曲風自動化辨識已有許多研究，但內容不外乎音訊處理、頻譜轉換、特徵擷取、特徵降維、監督式學習機。在相同的模式下提出各種改良，或是全新的特徵擷取…諸如此類，而辨識率也達到了七成以上。本篇論文採用不同於以往的做法，將訊號進行頻譜轉換後層層降維，所得之訊號搭配LDA與決策樹進行辨識，最後去比較與分析離散餘弦轉換與小波轉換在辨識率上的優劣。我們發現搭配小波轉換與混合LDA及決策樹的方法，可以將音樂曲風之分辨率達到八成五以上。
目錄
     口試委員會審定書.................................................................................................................. i
     致謝.......................................................................................................................................... ii
     中文摘要.................................................................................................................................. iii
     Abstract .................................................................................................................................... iv
     目錄.......................................................................................................................................... vi
     表目錄...................................................................................................................................... viii
     圖目錄...................................................................................................................................... ix
     第一章緒論....................................................................................................................... 1
     第一節研究背景與動機..................................................................................... 1
     第二節研究目的................................................................................................. 2
     第三節研究架構................................................................................................. 3
     第二章文獻探討.............................................................................................................. 4
     第一節前言......................................................................................................... 4
     第二節預處理..................................................................................................... 5
     第三節音樂特徵擷取......................................................................................... 7
     一、梅爾倒頻譜係數（Mel Frequency Cepstral Coefficients, MFCC）
     ................................................................................................................. 8
     二、雷尼熵值（Renyi Entropy, RE） .................................................. 9
     三、頻譜質心（Spectral Centroid, SC）.............................................. 9
     四、強度與音色（Intensity and Timbre） ........................................... 9
     第四節建置分類器............................................................................................. 11
     一、支持向量機（Support Vector Machines, SVM） ......................... 11
     二、最近鄰居法（k-Nearest Neighbors algorithm, k-NN）................ 12
     三、高斯混合模型（Gaussian Mixture Models, GMM）................... 13
     第三章降維方法.............................................................................................................. 14
     第一節小樣本分析............................................................................................. 14
     第二節音訊分析與k-means 演算法................................................................. 16
     第三節頻譜與降維............................................................................................. 17
     第四節線性判別分析......................................................................................... 21
     一、監督式維度縮減（Supervised Dimension Reduction）............... 21
     二、LDA 公式推導............................................................................... 22
     三、LDA 實驗結果............................................................................... 28
     第四章實驗方法.............................................................................................................. 30
     第一節挑選實驗音樂樣本................................................................................. 31
     第二節音訊處理................................................................................................. 33
     第三節維度縮減................................................................................................. 34
     第四節隨機七三分配......................................................................................... 34
     第五節線性判別分析之降維與預測................................................................. 35
     第六節離散小波轉換......................................................................................... 36
     第七節系統決策樹............................................................................................. 42
     第八節混合系統................................................................................................. 44
     一、Classical - Classical ........................................................................ 45
     二、Classical - Electron ......................................................................... 45
     三、Classical - Rock .............................................................................. 45
     四、Classical - Pop................................................................................. 46
     五、Classical - Vocal Pop ...................................................................... 46
     六、其餘情境......................................................................................... 48
     第九節混合系統的最終決策............................................................................. 50
     第五章結論與未來展望............................................................................................... 54
     參考文獻.................................................................................................................................. 55
參考文獻 [1]
     D.Pye
     Content-Based Methods for the Management of Digital Music
     in Proc IEEE Conf. Acoustics, Speech, Signal Processing (ICASSP), pp. 2437-2400, 2000
     
     [2]
     Dhanalakshmi, P. , Palanivel, S. , Ramalingam, V.
     Classification of audio signals using SVM and RBFNN.
     Expert Systems with Applications, 36(3), 6069-6075. doi: DOI 10.1016/j.eswa.2008.06.126, 2009
     
     [3]
     Xu, C. S. , Maddage, N. C. , Shao, X.
     Automatic music classification and summarization.
     Ieee Transactions on Speech and Audio Processing, 13(3), 441-450. doi: Doi 10.1109/Tsa.2004.840939, 2005
     
     [4]
     Ajmera, J. , McCowan, I. , Bourlard, H.
     Speech/music segmentation using entropy and dynamism features in a HMM classification framework.
     Speech Communication, 40(3), 351-363. doi: Doi 10.1016/S0167-6393(02)00087-0, 2003
     
     [5]
     Shao, B. , Wang, D. D. , Li, T. , Ogihara, M.
     Music Recommendation Based on Acoustic Features and User Access Patterns.
     Ieee Transactions on Audio Speech and Language Processing, 17(8), 1602-1611. doi: Doi 10.1109/Tasl.2009.2020893, 2009
      109(9):553--572, 1938.
     
     [6]
     Grey, J. M., Gordon, J. W.
     Perceptual effects of spectral modifications on musical timbres.
     Journal of the Acoustical Society of America 63 (5), 1493–1500, doi:10.1121/1.381843, 1978.
     
     [7]
     Duo-Fu Bao
     Supervised and Unsupervised Music Genre Classification
     NTUT Institute of Computer and Communication, 2008
     
     [8]
     Shih-kai Chen​
     Methodology of stage lighting control based on music emotion feeling.
     Department of Industrial Design, National Cheng Kung University 2015
     
     [9]
     Cortes, C. Vapnik, V.
     Support-vector networks.
     Machine Learning 20 (3): 273. doi:10.1007/BF00994018, 1995
     
     [10]
     Drucker, Harris; Burges, Christopher J. C.; Kaufman, Linda; Smola, Alexander J.; and Vapnik, Vladimir N.
     Support Vector Regression Machines.
     in Advances in Neural Information Processing Systems 9, NIPS 1996, 155–161, MIT Press. 1997
     
     [11]
     Altman, N. S.
     An introduction to kernel and nearest-neighbor nonparametric regression.
     he American Statistician 46 (3): 175–185. doi:10.1080/00031305.1992.10475879. 1992
     
     [12]
     Lie, L. , Liu, D. , Zhang, H. J.
     Automatic mood detection and tracking of music audio signals.
     Ieee Transactions on Audio Speech and Language Processing,9014(1), 5-18. doi: Doi 10.1109/Tsa.2005.860344. 2006
     
     [13]
     Baum, L. E.; Petrie, T.
     Statistical Inference for Probabilistic Functions of Finite State Markov Chains.
     The Annals of Mathematical Statistics 37 (6): 1554–1563. doi:10.1214/aoms/1177699147. Retrieved 28 November 2011. 1966
     
     [14]
     Nock, R. and Nielsen, F.
     On Weighting Clustering.
     IEEE Trans. on Pattern Analysis and Machine Intelligence, 28 (8), 1–13, 2006
     
     [15]
     Steinhaus, H.
     Sur la division des corps matériels en parties.
     Bull. Acad. Polon. Sci. 4 (12): 801–804. MR 0090073. Zbl 0079.16403 (French). 1957
     
     [16]
     MacQueen, J. B.
     Some Methods for classification and Analysis of Multivariate Observations.
     1, Proceedings of 5th Berkeley Symposium on Mathematical Statistics and Probability. University of California Press. 1967: pp. 281–297. 2009
     
     [17]
     Lloyd, S. P.
     Least square quantization in PCM.
     ell Telephone Laboratories Paper. 1957. Published in journal much later: Lloyd., S. P. Least squares quantization in PCM. IEEE Transactions on Information Theory. 1982, 28 (2): 129–137. doi:10.1109/TIT.1982.1056489. 2009
     
     [18]
     Jake Vanderplas
     Comparison of Manifold Learning methods
     http://scikit-learn.org/stable/auto\\_examples/manifold/plot\\_compare\\_methods.html
     
     [19]
     E-Course of NUTH
     https://ecourse.nutn.edu.tw/
     
     [20]
     曾正男
     一套提升凌波函數逼近能力與平滑度的方法
     國立中央大學 民國85年
     
     [21]
     Decision Trees Analysis
     http://www.mindtools.com/dectree.html?
     
     [22]
     格式工廠
     http://www.azofreeware.com/2008/10/formatfactory-155.html
描述 碩士
國立政治大學
應用數學研究所
101751014
資料來源 http://thesis.lib.nccu.edu.tw/record/#G0101751014
資料類型 thesis
dc.contributor.advisor 曾正男zh_TW
dc.contributor.author (Authors) 陳彥名zh_TW
dc.creator (作者) 陳彥名zh_TW
dc.date (日期) 2015en_US
dc.date.accessioned 1-Oct-2015 14:17:32 (UTC+8)-
dc.date.available 1-Oct-2015 14:17:32 (UTC+8)-
dc.date.issued (上傳時間) 1-Oct-2015 14:17:32 (UTC+8)-
dc.identifier (Other Identifiers) G0101751014en_US
dc.identifier.uri (URI) http://nccur.lib.nccu.edu.tw/handle/140.119/78751-
dc.description (描述) 碩士zh_TW
dc.description (描述) 國立政治大學zh_TW
dc.description (描述) 應用數學研究所zh_TW
dc.description (描述) 101751014zh_TW
dc.description.abstract (摘要) 隨著科技的進步，網際網路已充斥在我們的生活之中。音樂也不再以硬體儲存的方式流傳（例如ＣＤ、黑膠唱片），而是轉變為數位音樂的方式，透由網路平台散播。許多數位音樂串流服務平台網站也如雨後春筍般誕生，例如iTunes、Spotify、Musicovery。加上文化水平的提升，音樂已是現代人生活之中，不可或缺的一部分。世界上的音樂難以計數，如何將音樂分門別類做好管理乃為現代商業應用的一個重要課題。因此，音樂曲風自動化辨識的技術確實為一個實用且難以迴避的課題。
     
     過去在曲風自動化辨識已有許多研究，但內容不外乎音訊處理、頻譜轉換、特徵擷取、特徵降維、監督式學習機。在相同的模式下提出各種改良，或是全新的特徵擷取…諸如此類，而辨識率也達到了七成以上。本篇論文採用不同於以往的做法，將訊號進行頻譜轉換後層層降維，所得之訊號搭配LDA與決策樹進行辨識，最後去比較與分析離散餘弦轉換與小波轉換在辨識率上的優劣。我們發現搭配小波轉換與混合LDA及決策樹的方法，可以將音樂曲風之分辨率達到八成五以上。		zh_TW
dc.description.abstract (摘要) 目錄
     口試委員會審定書.................................................................................................................. i
     致謝.......................................................................................................................................... ii
     中文摘要.................................................................................................................................. iii
     Abstract .................................................................................................................................... iv
     目錄.......................................................................................................................................... vi
     表目錄...................................................................................................................................... viii
     圖目錄...................................................................................................................................... ix
     第一章緒論....................................................................................................................... 1
     第一節研究背景與動機..................................................................................... 1
     第二節研究目的................................................................................................. 2
     第三節研究架構................................................................................................. 3
     第二章文獻探討.............................................................................................................. 4
     第一節前言......................................................................................................... 4
     第二節預處理..................................................................................................... 5
     第三節音樂特徵擷取......................................................................................... 7
     一、梅爾倒頻譜係數（Mel Frequency Cepstral Coefficients, MFCC）
     ................................................................................................................. 8
     二、雷尼熵值（Renyi Entropy, RE） .................................................. 9
     三、頻譜質心（Spectral Centroid, SC）.............................................. 9
     四、強度與音色（Intensity and Timbre） ........................................... 9
     第四節建置分類器............................................................................................. 11
     一、支持向量機（Support Vector Machines, SVM） ......................... 11
     二、最近鄰居法（k-Nearest Neighbors algorithm, k-NN）................ 12
     三、高斯混合模型（Gaussian Mixture Models, GMM）................... 13
     第三章降維方法.............................................................................................................. 14
     第一節小樣本分析............................................................................................. 14
     第二節音訊分析與k-means 演算法................................................................. 16
     第三節頻譜與降維............................................................................................. 17
     第四節線性判別分析......................................................................................... 21
     一、監督式維度縮減（Supervised Dimension Reduction）............... 21
     二、LDA 公式推導............................................................................... 22
     三、LDA 實驗結果............................................................................... 28
     第四章實驗方法.............................................................................................................. 30
     第一節挑選實驗音樂樣本................................................................................. 31
     第二節音訊處理................................................................................................. 33
     第三節維度縮減................................................................................................. 34
     第四節隨機七三分配......................................................................................... 34
     第五節線性判別分析之降維與預測................................................................. 35
     第六節離散小波轉換......................................................................................... 36
     第七節系統決策樹............................................................................................. 42
     第八節混合系統................................................................................................. 44
     一、Classical - Classical ........................................................................ 45
     二、Classical - Electron ......................................................................... 45
     三、Classical - Rock .............................................................................. 45
     四、Classical - Pop................................................................................. 46
     五、Classical - Vocal Pop ...................................................................... 46
     六、其餘情境......................................................................................... 48
     第九節混合系統的最終決策............................................................................. 50
     第五章結論與未來展望............................................................................................... 54
     參考文獻.................................................................................................................................. 55		-
dc.description.tableofcontents 目錄
     口試委員會審定書.................................................................................................................. i
     致謝.......................................................................................................................................... ii
     中文摘要.................................................................................................................................. iii
     Abstract .................................................................................................................................... iv
     目錄.......................................................................................................................................... vi
     表目錄...................................................................................................................................... viii
     圖目錄...................................................................................................................................... ix
     第一章緒論....................................................................................................................... 1
     第一節研究背景與動機..................................................................................... 1
     第二節研究目的................................................................................................. 2
     第三節研究架構................................................................................................. 3
     第二章文獻探討.............................................................................................................. 4
     第一節前言......................................................................................................... 4
     第二節預處理..................................................................................................... 5
     第三節音樂特徵擷取......................................................................................... 7
     一、梅爾倒頻譜係數（Mel Frequency Cepstral Coefficients, MFCC）
     ................................................................................................................. 8
     二、雷尼熵值（Renyi Entropy, RE） .................................................. 9
     三、頻譜質心（Spectral Centroid, SC）.............................................. 9
     四、強度與音色（Intensity and Timbre） ........................................... 9
     第四節建置分類器............................................................................................. 11
     一、支持向量機（Support Vector Machines, SVM） ......................... 11
     二、最近鄰居法（k-Nearest Neighbors algorithm, k-NN）................ 12
     三、高斯混合模型（Gaussian Mixture Models, GMM）................... 13
     第三章降維方法.............................................................................................................. 14
     第一節小樣本分析............................................................................................. 14
     第二節音訊分析與k-means 演算法................................................................. 16
     第三節頻譜與降維............................................................................................. 17
     第四節線性判別分析......................................................................................... 21
     一、監督式維度縮減（Supervised Dimension Reduction）............... 21
     二、LDA 公式推導............................................................................... 22
     三、LDA 實驗結果............................................................................... 28
     第四章實驗方法.............................................................................................................. 30
     第一節挑選實驗音樂樣本................................................................................. 31
     第二節音訊處理................................................................................................. 33
     第三節維度縮減................................................................................................. 34
     第四節隨機七三分配......................................................................................... 34
     第五節線性判別分析之降維與預測................................................................. 35
     第六節離散小波轉換......................................................................................... 36
     第七節系統決策樹............................................................................................. 42
     第八節混合系統................................................................................................. 44
     一、Classical - Classical ........................................................................ 45
     二、Classical - Electron ......................................................................... 45
     三、Classical - Rock .............................................................................. 45
     四、Classical - Pop................................................................................. 46
     五、Classical - Vocal Pop ...................................................................... 46
     六、其餘情境......................................................................................... 48
     第九節混合系統的最終決策............................................................................. 50
     第五章結論與未來展望............................................................................................... 54
     參考文獻.................................................................................................................................. 55		zh_TW
dc.source.uri (資料來源) http://thesis.lib.nccu.edu.tw/record/#G0101751014en_US
dc.subject (關鍵詞) 小波轉換zh_TW
dc.subject (關鍵詞) 線性判別分析zh_TW
dc.subject (關鍵詞) 離散餘弦轉換zh_TW
dc.subject (關鍵詞) 決策樹zh_TW
dc.subject (關鍵詞) 曲風辨識zh_TW
dc.title (題名) 小波理論於曲風辨識上之應用zh_TW
dc.title (題名) The Application of Wavelet Transform on Automatically Musical Genre Classificationen_US
dc.type (資料類型) thesisen
dc.relation.reference (參考文獻) [1]
     D.Pye
     Content-Based Methods for the Management of Digital Music
     in Proc IEEE Conf. Acoustics, Speech, Signal Processing (ICASSP), pp. 2437-2400, 2000
     
     [2]
     Dhanalakshmi, P. , Palanivel, S. , Ramalingam, V.
     Classification of audio signals using SVM and RBFNN.
     Expert Systems with Applications, 36(3), 6069-6075. doi: DOI 10.1016/j.eswa.2008.06.126, 2009
     
     [3]
     Xu, C. S. , Maddage, N. C. , Shao, X.
     Automatic music classification and summarization.
     Ieee Transactions on Speech and Audio Processing, 13(3), 441-450. doi: Doi 10.1109/Tsa.2004.840939, 2005
     
     [4]
     Ajmera, J. , McCowan, I. , Bourlard, H.
     Speech/music segmentation using entropy and dynamism features in a HMM classification framework.
     Speech Communication, 40(3), 351-363. doi: Doi 10.1016/S0167-6393(02)00087-0, 2003
     
     [5]
     Shao, B. , Wang, D. D. , Li, T. , Ogihara, M.
     Music Recommendation Based on Acoustic Features and User Access Patterns.
     Ieee Transactions on Audio Speech and Language Processing, 17(8), 1602-1611. doi: Doi 10.1109/Tasl.2009.2020893, 2009
      109(9):553--572, 1938.
     
     [6]
     Grey, J. M., Gordon, J. W.
     Perceptual effects of spectral modifications on musical timbres.
     Journal of the Acoustical Society of America 63 (5), 1493–1500, doi:10.1121/1.381843, 1978.
     
     [7]
     Duo-Fu Bao
     Supervised and Unsupervised Music Genre Classification
     NTUT Institute of Computer and Communication, 2008
     
     [8]
     Shih-kai Chen​
     Methodology of stage lighting control based on music emotion feeling.
     Department of Industrial Design, National Cheng Kung University 2015
     
     [9]
     Cortes, C. Vapnik, V.
     Support-vector networks.
     Machine Learning 20 (3): 273. doi:10.1007/BF00994018, 1995
     
     [10]
     Drucker, Harris; Burges, Christopher J. C.; Kaufman, Linda; Smola, Alexander J.; and Vapnik, Vladimir N.
     Support Vector Regression Machines.
     in Advances in Neural Information Processing Systems 9, NIPS 1996, 155–161, MIT Press. 1997
     
     [11]
     Altman, N. S.
     An introduction to kernel and nearest-neighbor nonparametric regression.
     he American Statistician 46 (3): 175–185. doi:10.1080/00031305.1992.10475879. 1992
     
     [12]
     Lie, L. , Liu, D. , Zhang, H. J.
     Automatic mood detection and tracking of music audio signals.
     Ieee Transactions on Audio Speech and Language Processing,9014(1), 5-18. doi: Doi 10.1109/Tsa.2005.860344. 2006
     
     [13]
     Baum, L. E.; Petrie, T.
     Statistical Inference for Probabilistic Functions of Finite State Markov Chains.
     The Annals of Mathematical Statistics 37 (6): 1554–1563. doi:10.1214/aoms/1177699147. Retrieved 28 November 2011. 1966
     
     [14]
     Nock, R. and Nielsen, F.
     On Weighting Clustering.
     IEEE Trans. on Pattern Analysis and Machine Intelligence, 28 (8), 1–13, 2006
     
     [15]
     Steinhaus, H.
     Sur la division des corps matériels en parties.
     Bull. Acad. Polon. Sci. 4 (12): 801–804. MR 0090073. Zbl 0079.16403 (French). 1957
     
     [16]
     MacQueen, J. B.
     Some Methods for classification and Analysis of Multivariate Observations.
     1, Proceedings of 5th Berkeley Symposium on Mathematical Statistics and Probability. University of California Press. 1967: pp. 281–297. 2009
     
     [17]
     Lloyd, S. P.
     Least square quantization in PCM.
     ell Telephone Laboratories Paper. 1957. Published in journal much later: Lloyd., S. P. Least squares quantization in PCM. IEEE Transactions on Information Theory. 1982, 28 (2): 129–137. doi:10.1109/TIT.1982.1056489. 2009
     
     [18]
     Jake Vanderplas
     Comparison of Manifold Learning methods
     http://scikit-learn.org/stable/auto\\_examples/manifold/plot\\_compare\\_methods.html
     
     [19]
     E-Course of NUTH
     https://ecourse.nutn.edu.tw/
     
     [20]
     曾正男
     一套提升凌波函數逼近能力與平滑度的方法
     國立中央大學 民國85年
     
     [21]
     Decision Trees Analysis
     http://www.mindtools.com/dectree.html?
     
     [22]
     格式工廠
     http://www.azofreeware.com/2008/10/formatfactory-155.html		zh_TW