Clinical Application of the Synchronized Sentence Set (S3)

Document Type

Conference Proceeding

Publication Date

12-1-2009

Journal / Book Title

16th International Congress on Sound and Vibration, Kraków, Poland

Abstract

Nearly 15 years ago, the US Army (ARL-HRED) developed the Synchronized Sentence Set (S3) to assess the benefits of spatialized speech presentation for military personnel working in multi-talker environments. The S3 consists of 2034 sentences (10 syllables each) constructed from 104 token phrases and recorded by four male talkers. When presented together, sentences are temporally synchronized in a manner allowing key words in individual sentences to begin and end at the same time. The S3 companion software enables the user to simultaneously present one target (T) message and up to three competing (C) messages during a test condition. The user can independently route the T- and C-messages through one or more transducers (e.g., earphones or loudspeakers) to create a variety of divided and selective attention tasks. The purpose of this study was to determine whether the S 3 and companion software could be used to assess speech perception in multi-talker noise in a clinical setting. Sentences from the S3 were presented via (1) standard supra-aural earphones (diotically or dichotically), (2) four loudspeakers, and (3) a 3-D audio earphone system that simulated the positions of the four loudspeakers. Participants (N=26) listened to T-messages in the presence of 0, 1, 2 and 3 C-messages. The listeners' task was to record key words from the T-messages on a custom response sheet (640 trials/participant). Results indicated significant decreases in performance as the number of T-messages increased, for all listening modes. In conditions involving two or three C-messages, however, participants performed significantly better when T-messages were routed through loudspeakers or the 3-D audio system. Findings suggest that the S3 and companion software may offer clinicians and researchers with a reliable and flexible tool for assessing listener performance in multi-talker noise.

This document is currently not available here.

Share

COinS