Speech recognition in noise using the Danish Sentence Test (DAST): Audio-only vs. audio-visual stimulation

Understanding speech in noise can be challenging for individuals with and without hearing loss. This study investigated the effect of visual speech information on speech recognition in noise, using the newly developed Danish Sentence Test (DAST). Fifteen normal-hearing adults (18-55 years) participated. Speech recognition thresholds (SRTs) were measured in noise under audio-only and audio-visual stimulation. Each condition was tested twice, and the first condition was repeated to examine potential learning effects. Results showed a significant benefit from visual speech information, with a mean improvement of 1.54 dB compared to audio-only stimulation. Furthermore, good test-retest reliability was observed, with no significant differences between repeated measurements. A significant learning effect was observed in the audio-only and audio-visual conditions, with an average improvement of approximately 2.3 dB across repetitions. These findings confirm that visual speech cues enhance speech understanding in noise. They also indicate that the DAST is a reliable tool for assessing speech recognition under more ecologically<br/>relevant conditions.