Page 44 - Winter 2020

ONE SINGER, TWO VOICES
Figure 2. A and B: typical waveform and spectrum, respectively, of the glottal airflow during phonation. C and D: vocal tract transfer function for the vowel /ae/ and the corresponding radiated spectrum, respectively.
a strong effect on these frequencies and protruding the lips makes the VT longer, thus lowering the formant fre- quencies. The shape of the VT can be varied within wide limits. Moreover, by bulging the tongue more or less and in various directions, the VT can be narrowed or widened in almost any place along its length axis from the deep
pharynx to the hard palate. Also, the jaw and lip open- ings contribute to determining the VT shape. As a result, the formant frequencies can be varied within quite wide ranges: the first formant between about 150 and 1,000 Hz, the second from about 500 and 3,000 Hz, and the third from about 1,500 Hz and 4,000 Hz.
Overtone Singing
What is overtone singing, then? The term covers several different styles. Overtone singing was described as early as the nineteenth century by a famous singing teacher Manuel Garcia (Wendler, 1991) and has attracted the interest of several researchers. Smith and colleagues (1967) described “throat singing,” a type of chant per- formed by Tibetan Lamas, (see, e.g., s.si.edu/37QCSOZ). It is produced by males with special types of vocal fold vibrations, referred to as vocal fry register. Its pitch is stable and very low and is produced by a vocal fold vibration pattern in which every second or third airflow pulse is attenuated. Consequently, the pitch period of this drone is doubled or tripled. A similar type of phonation often occurs in phrase endings in conversational speech but is then typically aperiodic. In throat singing, two of the overtones are quite strong, and audible, thereby together giving the impression of a “chord.” Throat sing- ing is regarded as sacred in some Asian cultures.
The overtone singing demonstrated by AMH in the above link can be produced by both females and males. How- ever, the fundamental frequency of the drone is not as low as in throat singing. In AMH’s case, it is in the range typical of female speaking voices. The melody is played in a much higher pitch range by very strong overtones. Figure 3 shows some examples where overtones number 9, 7, and 4 are the strongest in the spectrum.
Figure 3. Examples of spectra produced in overtone singing by AMH. FE, frequency of the enhanced overtone.
44 Acoustics Today • Spring 2021