Effect of F0 on the intelligibility of emotional speech in noise for younger and older listeners

Abstract

This study investigated if F0 variability could explain why intelligibility in noise is better for speech spoken to portray fear compared to emotionally neutral speech. Word recognition accuracy was measured for stimuli produced with neutral vocal emotion, for intact stimuli portraying fear, and for six versions of the fear stimuli with varying reductions in F0 variability. Younger and older adults were tested in speech-spectrum noise and two-talker babble. Younger adults outperformed older adults. As F0 variability in the fear stimuli was reduced, performance in speech-spectrum noise for both age groups decreased and approached performance for the neutral stimuli. In the two-talker babble, even when F0 variability was most reduced, performance remained higher for the fear stimuli than the neutral stimuli, especially for older adults. The mean F0 for the fear stimuli was higher than for the neutral stimuli, while mean F0 of the neutral stimuli and the two-talker masker were similar. Thus, although the greater F0 variability in the fear stimuli confers an advantage to both age groups in speech-spectrum noise, F0 mean and variability both contribute to the effect of emotion on intelligibility in two-talker babble, especially for older listeners.