Handling Repeats and Jumps in Score-performance Synchronization

Given a score representation and a recorded performance of the same piece of music, the task of score-performance synchronization is to temporally align musical sections such as bars specified by the score to temporal sections in the performance. Most of the previous approaches assume that the score and the performance to be synchronized globally agree with regard to the overall musical structure. In practice, however, this assumption is often violated. For example, a performer may deviate from the score by ignoring a repeat or introducing an additional repeat that is not written in the score. In this paper, we introduce a synchronization approach that can cope with such structural differences. As main technical contribution, we describe a novel variant of dynamic time warping (DTW), referred to as JumpDTW, which allows for handling jumps and repeats in the alignment. Our approach is evaluated for the practically relevant case of synchronizing score data obtained from scanned sheet music via optical music recognition to corresponding audio recordings. Our experiments based on Beethoven piano sonatas show that JumpDTW can robustly identify and handle most of the occurring jumps and repeats leading to an overall alignment accuracy of over 99% on the bar-level.

[1]  Mark B. Sandler,et al.  Polyphonic Score Retrieval Using Polyphonic Audio Queries: A Harmonic Modeling Approach , 2003, ISMIR.

[2]  Daniel P. W. Ellis,et al.  Ground-truth transcriptions of real music from force-aligned MIDI syntheses , 2003, ISMIR.

[3]  Ning Hu,et al.  Polyphonic Audio Matching for Score Following and Intelligent Audio Editors , 2003, ICMC.

[4]  Christopher Raphael,et al.  A Hybrid Graphical Model for Aligning Polyphonic Audio with Musical Scores , 2004, ISMIR.

[5]  Yo Tomita,et al.  Towards an Intelligent Score Following System: Handling of Mistakes and Jumps Encountered During Piano Practicing , 2004, CMMR.

[6]  William P. Birmingham,et al.  Modeling Form for On-line Following of Musical Performances , 2005, AAAI.

[7]  Masataka Goto,et al.  A chorus section detection method for musical audio signals and its application to a music listening station , 2006, IEEE Transactions on Audio, Speech, and Language Processing.

[8]  Christopher Raphael,et al.  Music score alignment and computer accompaniment , 2006, CACM.

[9]  Philip Chan,et al.  Toward accurate dynamic time warping in linear time and space , 2007, Intell. Data Anal..

[10]  Meinard Müller,et al.  Information retrieval for music and motion , 2007 .

[11]  SalvadorStan,et al.  Toward accurate dynamic time warping in linear time and space , 2007 .

[12]  Gerhard Widmer,et al.  Automatic Page Turning for Musicians via Real-Time Machine Listening , 2008, ECAI.

[13]  Falk Scholer,et al.  Searching Musical Audio Using Symbolic Queries , 2008, IEEE Transactions on Audio, Speech, and Language Processing.

[14]  Xavier Serra,et al.  Chroma Binary Similarity and Local Alignment Applied to Cover Song Identification , 2008, IEEE Transactions on Audio, Speech, and Language Processing.

[15]  Mark B. Sandler,et al.  Structural Segmentation of Musical Audio by Constrained Clustering , 2008, IEEE Transactions on Audio, Speech, and Language Processing.

[16]  Meinard Müller,et al.  Multimodal presentation and browsing of music , 2008, ICMI '08.

[17]  Meinard Müller,et al.  Path-constrained partial music synchronization , 2008, 2008 IEEE International Conference on Acoustics, Speech and Signal Processing.

[18]  Meinard Müller,et al.  Sheet Music-Audio Identification , 2009, ISMIR.