學術產出-Theses
Article View/Open
Publication Export
-
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) 2015 en_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) G0101751014 en_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) 101751014 zh_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/#G0101751014 en_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 Classification en_US dc.type (資料類型) thesis en 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