A Semantic Web Annotation Tool for Web-Based Audio Sequencer

摘要

Music and sound have a rich semantic structure which is so clear to the composer and the listener, but that remains mostly hidden to computing machinery. Nevertheless, in recent years, the introduction of software tools for music production have enabled new opportunities for migrating this knowledge from humans to machines. A new generation of these tools may exploit sound samples and semantic information coupling for the creation not only of a musical, but also of a "semantic" composition. In this paper we describe an ontology driven content annotation framework for a web-based audio editing tool. In a supervised approach, during the editing process, the graphical web interface allows the user to annotate any part of the composition with concepts from publicly available ontologies. As a test case, we developed a collaborative web-based audio sequencer that provides users with the functionality to remix the audio samples from the Freesound website and subsequently annotate them. The annotation tool can load any ontology and thus gives users the opportunity to augment the work with annotations on the structure of the composition, the musical materials, and the creator's reasoning and intentions. We believe this approach will provide several novel ways to make not only the final audio product, but also the creative process, first class citizens of the Semantic Web.