利用音色特徵及支持向量機於內涵式音樂分類系統之設計與實現

Content-based Music Genre Classification Using Timbral Feature Vectors and Support Vector Machine

陳世豪(Shih-Hao Chen)

指導教授 ： 柯松源

共同指導教授 ： 陳璽煌(Shi-Huang Chen)

義守大學/電機資訊學院/資訊工程學系/博士(2016年)

摘要

本論文主要研究一種具自動分析之新型音樂曲風類型分類技術，該技術結合小波包轉換(wavelet packet Transform，WPT)、最佳小波包基的選取(best wavelet packet basis via best basis algorithm，BBA)、改良式梅爾倒頻譜係數(modified-MFCC)及分類器(Classifier)運用來有效地改善音樂曲風分類問題。在音樂曲風分類技術上，本文先使用兩種常見的濾波器分析技術如移動平均濾波器(moving average filter)及巴特渥斯低通濾波器(Butterworth low-pass filte)來部份地刪除短時距訊號分割之波動現象，接著使用小波包轉換(WPT)及改良式梅爾倒頻譜係數將音樂資料中的音色特徵如低階特徵及高階特徵擷取出來，然後根據被擷取出的音色特徵將音樂資料分類到各個音樂的類別。ISMIR2004 GENRE database是本實驗中所使用的音樂資料庫，該資料庫為2004 年音樂曲風分類競賽(ISMIR2004 Music Genre Classification Contest)所使用，此資料庫共1458首計六類曲風的音樂檔案可以被用來評估比較各種音樂資料分類演算法的良寙。根據實驗結果顯示，本論文所提之技術對於音樂分類辨識率可達89.7%。

關鍵字

最佳基演算法 ； 小波包轉換 ； 音樂曲風分類 ； 支持向量機 ； 稀疏表示分類器

並列摘要

This dissertation studies a new music genre classification algorithm based on wavelet packet Transform (WPT), best wavelet packet basis via best basis algorithm (BBA), modified Mel-frequency cepstral coefficient (modified-MFCC), and Classifier to accurately classify and increase classification performance. In the proposed music genre classification technique, moving average filter and Butterworth low-pass filter are first applied to partly eliminate the effect of fluctuation in short-term signal. Then WPT and modified-MFCC are used to extract timbre features, such as the low-level feature and high-level feature, to accomplish music genre classification. In our experiments we used a dataset of the ISMIR2004 GENRE classification contest. The dataset consist of 1458 songs in 6 classes, is used to evaluate the performances of music genre classification method against other similar schemes. Experimental results show that the proposed new music genre classification algorithm could achieve the average classification accuracy rate of 89.7%.

並列關鍵字

best basis algorithm ； wavelet packet transform ； music genre classification ； sparse representation based classification

參考文獻

[1] Li, T., and G. Tzanetakis. 2003. “Factors in automatic musical genre classification of audio signals,” Applications of Signal Processing to Audio and Acoustics, 2003 IEEE Workshop on., 19-22 Oct. 2003: 143-146. doi: 10.1109/ASPAA.2003.1285840.

[2] N. Scaringella, G. Zoia, and D. Mlynek. 2006. “Automatic genre classification of music content: A survey,” IEEE Signal Processing Magazine, Vol. 23, no.2, pp. 133–141, Mar. 2006.

[3] G. Tzanetakis and P. Cook. 2002. “Musical genre classification of audio signals,” IEEE Trans. Speech Audio Process., Vol. 10, no. 5, pp. 293–302, Jul. 2002.

[4] Nicolas Scaringella, Giorgio Zoia, and Daniel Mlynek. 2006. “Automatic genre classification of music content: A survey,” IEEE Signal Processing Magazine, Vol. 23, Issue 2, pp. 133-141, 2006.

[5] A. Flexer. 2007. “A closer look on artist filters for musical genre classification,” In Proceedings of the 8th International Conference on Music Information Retrieval (ISMIR'07), Vienna, Austria, 2007.