Seven problems that keep MIR from attracting the interest of cognition and neuroscience

  • Authors:
  • Jean-Julien Aucouturier;Emmanuel Bigand

  • Affiliations:
  • IRCAM/UPMC/CNRS STMS UMR 9912, Paris, France;LEAD/CNRS UMR 5022, Dijon, France

  • Venue:
  • Journal of Intelligent Information Systems
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

Despite one and a half decade of research and an impressive body of knowledge on how to represent and process musical audio signals, the discipline of Music Information Retrieval still does not enjoy broad recognition outside of computer science. In music cognition and neuroscience in particular, where MIR's contribution could be most needed, MIR technologies are scarcely ever utilized--when they're not simply brushed aside as irrelevant. This, we contend here, is the result of a series of misunderstandings between the two fields, about deeply different methodologies and assumptions that are not often made explicit. A collaboration between a MIR researcher and a music psychologist, this article attempts to clarify some of these assumptions, and offers some suggestions on how to adapt some of MIR's most emblematic signal processing paradigms, evaluation procedures and application scenarios to the new challenges brought forth by the natural sciences of music.