Perceptual and acoustic evidence for reduced fluency in the vicinity of stuttering episodes

A perceptual experiment and acoustic analyses were conducted to address the question whether stuttering occurs only at specific "moments" or whether it also affects the surrounding speech. Sections of stutterers' speech were extracted from clauses which were spoken completely fluently (control) or contained one stutter (experimental). In the experimental sections, only speech up to or following the stuttered word was employed. All sections were rated by independent groups of subjects for fluency, the nature of the excised stutter (repetition or prolongation), and the temporal position of the stutter relative to the fluent section that they heard (before or after). Two additional groups of listeners were asked to select from experimental-control pairs the section that had been drawn from near a stutter, and to indicate type and position of the stutter. Listeners could reliably judge which sections had been near a stutter and the type of that stutter, but not its position. Acoustic analyses showed that there were no differences in duration, rate, number of pauses, and average intensity between the experimental and control sections. However, there were significant differences in terms of the drop in intensity between the syllables in the respective sections. The perceptual identification of experimental versus control sections showed a significant relationship with this acoustic measure and with speech rate. The judgments about the type of stutter only correlated with drop in intensity. We conclude that stuttering episodes affect the intensity-time profile of the speech in their vicinity, and that listeners can use this acoustic information to infer the presence and type of the stutter.