Online Music-Recognition and Searching (OMRAS) embraces a number of interrelated problems in computational musicology, computer science and library science. The overall goal of this cross-disciplinary research is to fill a gap in the provision of online facilities for musical collections: the inability to search the content of the collections for ‘music’ itself.
OMRAS is a system for the efficient content-based searching and retrieval of musical information from online databases stored in a variety of formats ranging from encoded score-files (including files from commercial music-notation programs) to digital audio. The system will be easily operated by users with a wide range of musical ability and understanding and will not require sophisticated musical knowledge or the memorising of codes. It will be controlled by a simple-to-use graphical ‘musical’ interface, both for search queries and for the presentation of results.
Musical databases range from a modest collection of score-files stored on a single hard disk for an individual musical research project, to the ‘collection’ of the countless MIDI- or audio-files accessible via the Internet. This existing collection of online information is not currently exploitable for research or recreation (in the sense of being searchable or retrievable) other than by title or other descriptive metadata, where such is provided. The OMRAS project seeks to offer an easy-to-use solution to the problems of content-based music search and retrieval, useable on any computer platform via the World-Wide-Web.
By using both familiar and newly-developed musical displays for the construction of search-queries and for showing the results of searches, the system will avoid the problem of requiring users to learn complex encoding methods and interpretative skills in carrying out musical searches. Thus the research aims to enable users to access and exploit musical information in new ways based on familiar concepts, and will provide search-tools which are independent of considerations of language (other than the common ‘language’ of music), and through WWW-implementation are not affected by geographical location, disability, etc.
The three-year project for which funding has been granted by the International Digital Libraries programmes run by JISC (UK) and NSF (US) is a collaboration between two highly-respected partner institutions, the Center for Intelligent Information Retrieval at the University of Massachusetts (US) and King’s College London (UK). The partnership will build on the expertise and experience of both teams, and the project will use the results of previous research in both institutions and elsewhere as well as involving a significant amount of new research.
You can see an announcement ( UK alternative 'mirror' site) of the six US/UK projects funded under the joint JISC/NSF Digital Libraries scheme in the June 1999 issue of D-Lib online magazine ( UK mirror).
A list of OMRAS publications is available. We hope to provide downloadable versions of most of them and other goodies before long.