標題: | 基於機器學習的鋼琴演奏技巧評量系統 Machine Learning for Automatic Evaluation of Basic Piano Performance |
作者: | 王予婷 張文輝 電信工程研究所 |
關鍵字: | 機器學習;決策樹迴歸;支持向量機;相關係數;鋼琴觸奏法;樂器數位介面;Machine Learning;decision tree;SVM;correlation;Articulation;MIDI |
公開日期: | 2016 |
摘要: | 由於電子音樂科技的快速發展,透過網路分享MIDI演奏音樂資料的需求日益劇增。為了改善初學者的樂器演奏能力,國內外學者相繼提出緊密結合電腦輔助與多媒體技術的互動式音樂教學平台,然而這些應用程式大都是透過虛擬鋼琴提供基礎的輔導功能,對於學生演奏技巧的提升相當有限。為了提升初學者的學習成效,本論文開發出一套鋼琴演奏熟練度的自動化評量系統。本論文分為兩個部份,第一部份利用MIDI特徵紀錄值進行三次雲規曲線擬合取得演奏趨勢的特徵參數,並使用決策樹演算法,針對鋼琴音階指法進行熟練度進行評分,並考量不同大調左右手的音階指法差異性。實驗結果顯示決策樹評分機制與不同專家間評分的相關係數為0.7625,呈現高度相關。第二部份則是利用支持向量機演算法,針對三種鋼琴觸奏法做分類判別。實驗結果顯示其和各專家的相關係數皆大約落在0.9,亦呈現高度相關。由此可知本系統在鋼琴演奏技巧的熟練度評量具有良好的一致性與可靠性。 In recent years, the low-priced MIDI piano has become one of the most popular musical instruments. However, expert pianists are not usually available when the piano beginners are in need of corrective actions for proficiency improvement. To construct a support system for self-learning at the beginning stage, we proposed an automatic evaluation system of basic piano performance. This paper is divided into two parts. In the first part, we extracted five MIDI-based parameters and applied the cubic spline curve fitting technique to extract the feature parameters representing the tendencies of the proficiency for scale performance. Then we employed the decision tree algorithm to predict the proficiency scores of testing data. Experimental results show that the correlation coefficient between the evaluation scores given by automatic evaluation and three expert pianists is 0.7625. In the second part, we employed the SVM algorithm to classify three different articulations for piano performance. Experimental results show that the correlation coefficients between the automatic classification and three expert pianists are around 0.9. Results as above indicate the effectiveness of the proposed methods. |
URI: | http://etd.lib.nctu.edu.tw/cdrfb3/record/nctu/#GT070460206 http://hdl.handle.net/11536/139745 |
顯示於類別: | 畢業論文 |