A Digital Library Framework for Heterogeneous Music Collections - from Document Acquisition to Cross-Modal Interaction

Abstract

In the digital age, increasing amounts of digitized music materials result in the need for automated management processes. Especially for libraries and other content-providing organizations holding large and due to ongoing digitization efforts steadily increasing amounts of digital copies of music materials, there is a high demand on automatisms to cope with the vast number of documents and streamline their processing.<br /> In this thesis, a digital library system for managing heterogeneous music collections is developed. The aforementioned heterogeneity refers to various document types and formats, as well as to different modalities, e.g., compact disc-audio recordings, scans of sheet music, and lyrics. The system offers a full-fledged, widely automated document processing chain: digitization, indexing, annotation, linking, access, and presentation. The system is implemented as a generic and modular music repository based on an extensible service-oriented architecture. As a particular benefit of the approach pursued in this thesis, the various documents, representing different aspects of a piece of music, are jointly considered in all stages of the document processing chain. Concerning retrieval functionalities, incorporated state-of-the-art music information retrieval techniques and adequately designed user interface components allow for integrated, synchronized, and multimodal presentation of documents (WYSIWYH: what you see is what you hear), cross-modal score- or lyrics-based navigation in audio and vice versa, as well as sophisticated cross- and multimodal retrieval. Hence, the type of repository proposed in this thesis might be called a truly cross-modal music digital library system. <br /> This thesis describes a complete framework that the system is based on including business processes and system architecture, exposes applied music information retrieval techniques incorporated in the document processing chain, and illustrates implemented functionalities for user interaction. As a part of the German PROBADO digital library initiative, with a view on practical application and integration into existing business processes, it is described how the framework is put into practice at the Bavarian State Library Munich which houses a music data stock whose volume is of internationally significant rank.