Music information retrieval

from Wikipedia, the free encyclopedia
Scientific disciplines involved in Music Information Retrieval

Music Information Retrieval [ mjuːzɪkˌɪnfɚˈmeɪʃən ɹɪˈtɹiːvəl ] ( MIR , dt. About music information retrieval) describes an interdisciplinary research direction for the development of methods for the extraction or aggregation of information for data processing. As a sub-area of information retrieval , it includes strategies for searching and finding information in digital environments. The typical practical application for MIR is DJing : the software looks for pairs from thousands of apparently unrelated music tracks that can be easily faded into one another.

Interdisciplinary research approaches

  • digital signal processing

A fundamental approach is to extract features from audio files using methods of digital signal processing . Patterns are obtained from the sampled signals, which can be represented mathematically in vectors (so-called feature vectors ) and matrices.

Areas of responsibility

Use case description
Music identification Identification of CDs, provision of information regarding unknown songs
Plagiarism search
Copyright monitoring
Version identification Remixes, live or studio recording, cover songs.
Melody search Finding pieces of music that contain a certain melody
Artist search Finding pieces of music by a specific artist
Sounds like... Finding a piece of music based on a given example
Automatic music transcription computer-based transmission of audio in MIDI , MusicXML or similar

Web links

Individual evidence

  1. Roland Stigge: Automatic Music Transcription (AMT). (PDF; 585 kB) (No longer available online.) June 16, 2003, archived from the original on January 18, 2016 ; Retrieved September 20, 2011 . Info: The archive link was inserted automatically and has not yet been checked. Please check the original and archive link according to the instructions and then remove this notice. @1@ 2Template: Webachiv / IABot / www.rolandstigge.de