A Toolbox for Inspection of Music Plagiarism
20th European Signal Processing Conference (EUSIPCO 2012) Bucharest, Romania, August 27 - 31, 2012 AUDIO FORENSICS MEETS MUSIC INFORMATION RETRIEVAL - A TOOLBOX FOR INSPECTION OF MUSIC PLAGIARISM Christian Dittmar, Kay F. Hildebrand*, Daniel Gaertner, Manuel Winges, Florian Muller*,¨ Patrick Aichroth Fraunhofer IDMT, Department Metadata, Ilmenau, Germany *University of Munster,¨ Department of Information Systems, Munster,¨ Germany ABSTRACT the qualified cases, and to lower the costs associated with settling disputes. This paper presents a toolbox that has been developed in or- Typically, when music plagiarism cases are brought to der to facilitate the inspection of suspected music plagiarism court, independent music experts, often musicologists, are as- cases. The basic concept is the use of techniques from Music ked to analyze the similarities between two songs, and the Information Retrieval for semi-automatic inspection of origi- judges rely on their opinion. We believe that, in order to sup- nal and suspect song. Basic types of music plagiarism are port such analysis, specialized software can be provided to discussed. Several signal processing approaches suitable to analyze musical features of the suspected music recordings. reveal these types are introduced. They are intended to be Similarities can be identified by applying well-described pat- used under supervision of a human expert. Evaluation of the tern matching algorithms from the Music Information Retrie- proposed methods in a non-supervised scenario is not within val (MIR) literature. Moreover, such software can display the scope of this paper. similarities in a way that experts can not only use to evaluate Index Terms— Music Information Retrieval, Music Pla- their importance, but also to visualize and explain it to an un- giarism, Audio Forensics trained audience.
[Show full text]