Abstract
In this study, two contrastive typical tongue shape methods were studied. One was Ship-bottom Tongue (ST) and the other was Rolled Tongue (RT) against upper jaw. An Italian operatic trained singer (dramatic tenor, age 37) sang five Japanese vowels, /a, i, u, e, o/, at the same pitch frequency of 415 Hz and loudness in an unechoic room using the two methods, and the vocalizations were recorded on a digital audio tape recorder. In an acoustic study, two different types of frequency spectra were measured : a long-term average spectrum to investigate differences in the harmonic structures in the low frequency range and a short-term spectrum to observe spectral differences in the high frequency range. The first two formant frequencies were also carefully measured using an analysis-by-synthesis method. In a physiologic study, Magnetic Resonance Imaging (MRI) was adopted to take a sagittal section picture of the articulatory organ during singing in which differences in the two tongue shapes could be observed. All the voice samples were subjected to perceptual evaluation experiments by 48 professional opera singers and student singers, in which ST voices were compared with RT ones in terms of “pithiness impression.” Findings were summarized as follows : (1) ST voices mostly possessed rich harmonics in the frequency range up to 20 kHz, while RT voices revealed less energy in the frequency range above 10 kHz, (2) all the RT voices showed larger distances between the first and second formant frequencies than the ST counterparts, (3) the oral cavity in ST was much larger than that in RT, while the pharyngeal cavity in ST was much narrower than in RT, (4) RT voices were judged more pithy than ST voices. Therefore, it was concluded that pithy impression of RT voices was closely related to the distance between the first and second formant frequencies which was clearly caused by the different control of the tongue shape in singing and that the tongue shape could effect vibration of the vocal folds.
