Two Classification Methods for Grouping Common Environmental Sounds in Terms of Perceived Pleasantness
Abstract
Real-world sounds are perceived as more than just a collection of acoustic attributes. They contain both acoustic and semantic attributes, which together influence perception. Acoustic measurements are usually well-defined and broadly agreed upon, but semantic information can be difficult to operationalize. Semantic attributes contain information related to context and listener experience, attributes that are subjective and prone to individual differences. The current study had 2 goals: 1) characterization of a set of sounds in terms of their perceived pleasantness and 2) comparison of 2 classification methods, experimenter-defined and objectively defined (i.e., data-defined) classification. Fourteen listeners rated 36 common environmental sounds on a scale of 1 to 7 for perceived pleasantness. Overall, the set trended toward the unpleasant end of the ratings scale, and impulse sounds were rated as more pleasant than continuous sounds. Further, it appears that natural sounds were rated as more pleasant than man-made sounds, and this is consistent with both experimenter- and objectively defined classification approaches, suggesting a latent categorical structure for pleasantness.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 2016
- Accession Number
- AD1029204
Entities
People
- Brandon S. Perelman
- Jeremy R. Gaston
- Kelly Dickerson
- Laura Sherry
Organizations
- United States Army Research Laboratory