Effects of voicing onset timing on the production and perception of fricative–vowel sequences

Fricative–vowel sequences such as /su/ involve a transition from the voiceless fricative /s/, produced by turbulent noise, to the vowel /u/, which involves vocal fold vibration. Previous studies have reported cases where vocal fold vibration was not observed during the fricative portion of /zu/, leaving the mechanism by which speakers distinguish between /su/ and /zu/ unclear. This study aims to clarify how the timing of voicing onset affects the acoustic and perceptual characteristics in a fricative–vowel sequence using a simplified mechanical voice production model. The mechanical model consists of a vocal tract with a movable tongue model and a reed-type artificial larynx. Voicing onset was controlled by an electromagnet with a metal reed, and the tongue was moved vertically to simulate the transition from fricative to vowel. Results showed that the model reproduced the natural fricative–vowel co-articulation in human speech, while allowing control of voicing onset. Perception tests revealed that /su/ and /zu/ were distinguished in sounds where the frication and voicing intervals overlapped by approximately 20–60%. These findings suggest that the distinction between /su/ and /zu/ does not rely solely on voicing presence, but rather on the degree of temporal overlap between frication and voicing.