Real time labeling of affect in music using the affectbutton

  • Authors:
  • Joost Broekens;Anne Pronker;Marian Neuteboom

  • Affiliations:
  • TU Delft, Delft, Netherlands;Stedelijk Gymnasium Leiden, Leiden, Netherlands;Stedelijk Gymnasium Leiden, Leiden, Netherlands

  • Venue:
  • Proceedings of the 3rd international workshop on Affective interaction in natural environments
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

Valid, reliable and quick measurement of emotion and affect is an important challenge for the use of emotion and affect in human-technology interaction. Emotion and affect can be measured in two different ways: explicit, the user is asked for feedback, and implicit, signals from the users are automatically translated to affective and emotional meaning (affect recognition). Here we focus on explicit affective feedback. More specifically, we focus on the evaluation of an affect measurement tool called the AffectButton. Previous evaluation studies [2] showed that the AffectButton enables users to give affective feedback in a low-effort, reliable and valid way. In this paper we report a study involving real-time affective labeling of movie music by primarily high school students, i.e., a realistic domain with mainstream users. Our results show that (a) users (n=21) are able to use the AffectButton in real time while listening to the music; (b) the labeling very-well follows the changes in the music and gives insight into the different affective dimensions of the music, and; (c) objective music properties correlate to these affective dimensions replicating findings of others. This provides evidence that the AffectButton is a viable affect measurement tool usable by non-expert users in real-time realistic domains.