Temporal Dynamics of Repetitions during the Early Stage of Stuttering: An Acoustic Study

The purpose of this study is to compare the duration characteristic of sound repetitions in the speech of adults who stutter (S=10) recorded near the onset of their stuttering to those of controlled nonstuttering adults (C=10). Dysfluent episodes are identified in digital recordings of the clients read speech. The digitized signals are analyzed by means of Cool Edit Pro software. Using visual displays of sound spectrograms, the durations of the spoken repetition units, the silent intervals between the units and the total dysfluency are measured. The stutterers exhibit shorter silent intervals between spoken repetitions units, which is used as one of the parameter for objective assessment of early stuttering. The total duration of the stutterer’s dysfluency is significantly shorter because of their shorter silent intervals when compared to dysfluency of equal repetition units produced by control subjects. Analysis reveal that silent interval duration is capable of differentiating stuttering from control client with 77.4 - 95 % accuracy. In this work 80% of data are used for training and remaining 20% for testing.