Software developed by an academic at Goldsmiths, University of London could spell the end for future melody plagiarism.
Dr Daniel Müllensiefen, from the Department of Psychology and formerly working in Computing, has co-published research on how to predict court decisions on music plagiarism using cognitive similarity algorithms. The study has recently been published by the European specialist journal Musicae Scientiae and results were presented publicly for the first time at the international conference of the European Society for the Cognitive Sciences in Music (ESCOM) in Finland in August.
Daniel worked alongside Marc Pendzich, an expert on cover versions and music re-mixes from the Institute of Musicology University of Hamburg, on the software which is based on modelling court decision for cases of alleged melodic plagiarism employing a number of similarity algorithms.
The two researchers used court cases from the US as a testbed for their software and 90 per cent of the court decisions were predicted correctly by the newly developed algorithms.
Tune plagiarism in pop music is a common and often feverishly debated phenomenon, so controversial due to the vast amounts of money involved in today's pop music industry.
Artists as high profile as Madonna, George Harrison and the Bee Gees have all been involved in music plagiarism cases.
The similarity between melodies is assumed to be a very important factor in a court's decision about whether a new tune is an illegitimate version of a pre-existing melody.
Under the current system, the jury is advised by expert witnesses to come to a decision – something both Daniel and Marc have indeed done – but they admit that one of the long term effects of their work could substantially alter the need for a jury and expert witnesses.
"The most provocative question you could ask is whether this software could replace a jury and expert witnesses in court," Daniel said.
"Also, on a very popular level you could claim that the software can detect melodic plagiarism in popular music automatically. Thus, in principle we could develop this into a business where songwriters and music publishers submit songs and we test against a database whether there are any highly similar pre-existing melodies in it."
Currently these developments are hypothetical due to the sample of cases it has been tested on being so small (20 cases), but Daniel and Marc are working on a follow-up study to include more US cases and to test whether the prediction accuracy holds also true for British and German plagiarism suits.