Collaboratively Identifying and Referring to Sounds with Words and Phrases

Abstract

Machine classification of underwater sounds remains an important focus of U.S. Naval research due to physical and environmental factors that increase false alarm rates. Human operators tend to be reliably better at this auditory task than automated methods, but the attentional properties of this cognitive discrimination skill are not well understood. In the study presented here, pairs of isolated listeners, who were only allowed to talk to each other, were given a collaborative soundordering task in which only words and phrases could be used to refer to and identify a set of impulsive sonar echoes. The outcome supports the premise that verbal descriptions of unfamiliar sounds are often difficult for listeners to immediately grasp. The method of “collaborative referring” used in the study is proposed as new technique for obtaining a verified perceptual vocabulary for a given set of sounds and for studying human aural identification and discrimination skills.