Titlebar

Bibliographische Daten exportieren
Literatur vom gleichen Autor
plus im Publikationsserver
plus bei Google Scholar

Lesezeichen anlegen
 

In the Blink of an Ear : A Critical Review of Very Short Musical Elements (Plinks)

Titelangaben

Verfügbarkeit überprüfen

Thiesen, Felix C. ; Kopiez, Reinhard ; Reuter, Christoph ; Czedik-Eysenberg, Isabella ; Schlemmer, Kathrin:
In the Blink of an Ear : A Critical Review of Very Short Musical Elements (Plinks).
In: Proceedings of the 14th International Conference on Music Perception and Cognition, July 5-9, 2016, San Francisco, California. - San Francisco, CA, USA : Causal Productions, 2016. - S. 147-150
ISBN 1 876346 65 5

Volltext

Link zum Volltext (externe URL): http://icmpc.org/icmpc14/files/ICMPC14_Proceedings...

Kurzfassung/Abstract

At the turn of the millennium, several much-noticed studies on the rapid assessment of musical genres by very short excerpts ("plinks") have been published. While at first the research mostly focused on clips of popular pieces of music, the recognition of elements in the millisecond range has since been attracting growing attention in general. Although previous studies point to surprisingly short durations of exposure for above-chance genre recognition (mostly between 125 and 250 ms), interpretations of the results are ambiguous due to blurring factors. For example, varying duration thresholds for genre recognition could have been caused by coarse-meshed iterations, unstable audio-quality of sources used, unsatisfactory sample sizes, and insufficient theoretical foundation. In a pilot study, based on a selection of original stimuli used in related precedent studies, we aim to (a) identify essential spectral low-level audio features of the stimuli; (b) determine spectral cues which are obsolete for rapid assessement by means of a correlation analysis between audio features and recognition rates; (c) test for the reliability of genre recognition performance; and (d) compare the selected excerpt to other sections of the particular song (spectral matching). Spectral analyses of audio features are based on the MIR toolbox and on the software package dBSONIC. Replication of genre recognition is based on online-experiments, and the comparison between sound excerpts and the full song is being conducted by a researcher-developed software application. The work is currently under progress but first results indicate that recognition rate strongly depends on the more or less random selection of excerpts in previous studies, which might not always be representative of an entire song. In future studies, researchers should exert more control over the sound material: for instance, by selecting multiple excerpts from one song which should be controlled for the spectral fit to the whole piece.

Weitere Angaben

Publikationsform:Aufsatz in einem Buch
Institutionen der Universität:Philosophisch-Pädagogische Fakultät > Musikwissenschaft > Professur für Musikwissenschaft
Titel an der KU entstanden:Ja
Eingestellt am:26. Jul 2016 14:14
Letzte Änderung:26. Jul 2016 14:14
URL zu dieser Anzeige:http://edoc.ku-eichstaett.de/18335/