Download notess-minibayeva-jcdl2002.pdf

Contents : Variations2: Toward Visual Interfaces for Digital Music Libraries Mark Notess University Information Technology Services Cook Music Library Indiana University Bloomington IN 47405 USA mnotess@indiana.edu Natalia Minibayeva University Information Technology Services Cook Music Library Indiana University Bloomington IN 47405 USA nminibay@indiana.edu ABSTRACT In this paper we discuss unique challenges of visualizing musical data address limitations of existing music interfaces and data structures and propose some approaches to music visualization based on the data model developed by Variations2 digital music library project. Keywords Digital library music information visualization. INTRODUCTION Visualization of music bibliographic data encounters difficulties on two levels. First today s music bibliographic data is typically force-fit into a catalog structure that obscures most of its richness and utility. Second and certainly as a result of the first obstacle useful visualization tools for music data inquiry and presentation simply do not exist. Variations2 the Indiana University Digital Music Library Project 12 is making a significant effort to address these problems by providing a metadata record structure to capture the richness and relationships inherent in music bibliographic data. Variations2 is also examining the visualization opportunities provided by this new catalog structure. This paper describes the limitations of existing data structures and technologies and offers a glimpse of work in progress on the related issues of catalog structure and visualization. logical (logical information required for music realization) analytical (concepts involved in musical analysis) Each domain captures certain music properties yet excludes others. Second any given musical work is typically manifested in various performances and printed editions. These various instantiations of the same work often differ drastically from one another as well as from the more abstract concept of the musical work in terms of their instrumentation structure duration and interpretation. Thus certain attributes can be shared among the instantiations of a work and some can differ. This presents a challenge in terms of linking instantiations to works which is necessary for efficient and effective organization retrieval and manipulation of music information. Finally instantiations of a musical work exist in a variety of formats (score recording video text). Therefore in addition to the differences resulting from various interpretations of the musical work there are also variations imposed by the different formats. Music representation and organization present a number of challenging problems that affect the ways in which visual music interfaces are constructed. First of all as defined by the Standard Music Description Language (SMDL) 8 music exists in several physical and logical dimensions: Without consideration of these issues specific to the music domain it would be difficult if not impossible to design an effective visual music interface. The needs of music users can be satisfied only if musical objects are adequately described represented and organized. This paper focuses on visual interfaces related to music information retrieval which is concerned mostly with bibliographic music information rather than actual music representation schemas (which present their own fascinating issues). visual/graphic (music notation) CATALOG RECORD STRUCTURE LIMITATIONS performance/gestural performance) UNIQUENESS OF MUSICAL DATA (actual sounding and The existing text-based catalog record structures such as the USMARC 7 provide a foundation for capturing some of the required descriptive music information but also present some significant limitations. The biggest problem is the flat nature of the record structure which does not reflect the complexity of the relationships existing among various musical objects. For example when a music library user is looking for some recordings and scores of a specific musical work s/he has to perform several complex searches to collect all the needed records (e.g. Mozart AND Sonatas AND K332 AND Score etc.). This is because MARC records mostly capture container-level information and do not provide a robust mechanism for linking various instantiations of the same musical work. This presents difficulties in searching for music information as search queries are often very complex and require special user training. axes are as follows: x-date of composition y-composer zsolo instrument. INQUIRY AND PRESENTATION TOOL LIMITATIONS Visual interfaces to music bibliographic data are not broadly used today. To explore what could be done with existing tools and data we carried out some experiments which we describe in this section. Variations and IUCAT The original Variations project 4 11 for which Variations2 is named pr

INFO HASH : e0f1159ebd161b5d807478edec60e84bcd1a6d6a

Download now

File Size: 415.3 kb