透過您的圖書館登入
IP:18.221.146.223
  • 學位論文

使用音樂聲音及樂譜圖檔對齊技術的線上音樂課程資料庫

A Web-based Music Lecture Database Framework with Music Audio to Score Image Alignment Technology

指導教授 : 陳恆佑
共同指導教授 : 張克寧

摘要


本論文提出一個為呈現線上音樂課程而設計的架構。此架構提供音樂課程的編輯及存取的功能。我們提出以動態規畫法(Dynamic Programming)為基礎的演算法來找出樂譜圖檔上的位置與其對應的演奏聲音的時間的關係,進而提供樂譜圖檔與演奏聲音的同步呈現。我們也開發一個音高辨識(pitch extraction)技術,將聲音的波形訊號轉換為音高代碼。然而,對於多音及多種樂器的音樂,目前仍然會存在許多音高辨識的錯誤。自行開發的樂譜辦識技術(Optical Music Recognition)則用來盡可能地抽出樂譜圖檔上每個音符及它們的音高。我們提出的排比演算法(alignment algorithm)在存在音高辨識錯誤及樂譜辨識錯誤的情況下,仍能盡可能地找出兩者間的關係。 教師們可以在已對齊好的樂譜圖檔及演奏聲音上附加許多種的教材。有了樂譜圖檔和聲音的同步呈現後,學習者可以在聆聽樂曲時,同步讀譜及由教師們提供的附加資訊。 我們建立一個實驗系統來評估對於同步準確度的主觀感受。以小節為單位請使用者就主觀的感受評定是否同步準確。實驗結果顯示覺得完全正確及有可接受的誤差的小節數占95.63%。我們的實驗對象包含音樂背景及工程背景的人士。對於我們系統成效的問卷都給予正面的意見。

並列摘要


This thesis presents a framework for authoring, storing, retrieving, and presenting music lectures on the Web. For a synchronized presentation between score and recorded performance audio, we propose a dynamic programming algorithm for Score-to-Wave alignment to explore the relations between positions on music score images and the time of corresponding performance recording. We propose a pitch extraction technique to transcribe performance audio into the symbolic domain. Transcription errors, however, are inevitable, especially for polyphonic music or multi-instruments music because of the complex harmonics of different instruments. We also developed a score image recognition engine to extract music notes from score images. After the audio and image are transcribed to notes, the proposed alignment algorithm is applied on them even if both of them contain errors. The aligned music score images and performance recordings can be attached to many kinds of teaching materials. With a synchronized presentation, learners can read music scores and get instructional information when listening to certain sections of music pieces. An evaluation was made to obtain user feedback based on their subjective experience when using our system. The percentage of bars which were considered as aligned perfectly or within acceptable limits was 95.63%. The questionnaire in the evaluation system also reported positive opinions from both engineers and musicians.

參考文獻


? Digital Music Library
[Dunn99] J. W. Dunn and C. A. Mayer, “VARIATIONS: A digital music library system at Indiana University”. In Proceedings of the Fourth ACM Conference on Digital Libraries, 1999. pp.12-19
[Dunn06] J. W. Dunn, D. Byrd, M. notes, J. Riley, and R. Scherle, “VARIATIONS2: Retrieving and using music in an academic settings,” Special Issue: Music Information retrieval, Communications of the ACM, 2006, pp.53-58.
[Kurth07] F.Kurth, M. Muller, C. Fremerey, Y. Change, and M. Clausen. “Automated Synchronization of Scanned Sheet Music with Audio Recordings.” In Proceedings of The International Society for Music Information Retrieval, Vienna, Austria, September 2007, pp.261-266.
? Music Score Alignment

延伸閱讀