抄録
It is well accepted that nasality must be regarded as a matter of resonance. However the acoustical characteristics of the nasalized vowels are yet remained in spite of many investigations.
Reviewing previous investigations, the authers attempted to synthesize the nasalized vowels by modifying the normal non-nasalized vowels acoustically. Five Japanese vowels uttered by a normal subject with three different pitches were acoustically modified using a bandpass filter, and the way of modification, by which the vowels were most markedly nasalized was discussed.
It was clarified that each vowel was most markedly nasalized when the harmonic in a specific frequncy range was strengthened. This frequency range was found to vary according to vowels but not to pitches of the foundamental tone, i. e. to be ca. 800cps for vowel “u”, ca. 800cps for vowel “o”, ca. 1000cps for vowel “a”, ca. 800 to 1000cps for vowel “e”, and ca. 600cps for vowel “i”. What these facts should mean was discussed by sonagrams and it was concluded that each vowel was most markedly nasalized when the bandwidth of the first formant was increased in the direction of higher frequency range.
