Practica oto-rhino-laryngologica. Suppl.

Show abstractHide abstract

The vibration of the body wall during phonation was measured using accelerationtype pick-ups attached to five different points on the body wall. The voice was recorded simultaneously. The test subjects were 18 amateur singers. Each subject was asked to phonate five Japanese vowels /a/, /i/, /u/, /e/ and /o/ in both trained singing voice (trained voice) and untrained singing voice (untrained voice). This distinction was possible for all subjects after some practice.
Spectral analysis of the body wall vibration was performed. The intensities of the vibration were also analysed with respect to the vibration proportion (VP), which is the ratio of the body wall vibration (G) to the sound pressure level. 'Singing formant'was found not only in the spectral structure of the voice sound but also in that of body wall vibration. The VP of the cheek was greatest with the trained head voice, but it was not so great during the production of trained chest voice. With the untrained voice, there was no difference in the VP among the five locations of the body wall. The VP at the cheek in music students was greater than that of members of a glee club. The difference of VP among five vowels was smaller with the trained voice than with the untrained voice. These results suggest the important role of the vibration of the cheek and great resonance of oral cavity in the trained voice.
Listening tests were also performed. Five vowels of voice sound that had been recorded simultaneously with the vibration of body wall were presented to a group of 18 listeners for identification. Articulation score for trained voices was lower than that for untrained voices. The difference of articulation score among five Japanese vowels was smaller with the trained voice.
These results suggest that the difficulty in perceiving sung vowels arises from ‘covering’.

View full abstract

Download PDF (6806K)

Register with J-STAGE for free!