A Hypothesis about Whistle Voice
Ingo R. Titze
Abstract
Ingo R. Titze
Abstract
YOUNG ADULT FEMALES SOMETIMES STUMBLE upon a very high pitched piccolo-like sound they can produce in their larynx. It is known as whistle voice, or whistle register. When first experienced by a singer, it often startles her as much as a listener. It seems to be produced with ease at pitches starting from G^sub 6^ (about 1570 Hz) to G^sub 7^ (3140 Hz). Some authors have claimed that whistle voice can start below 1000 Hz, especially in untrained females,1 but trained lyric and coloratura sopranos usually sing notes well above 1000 Hz in nonwhistle phonations. The hypothesis here is that whistle voice makes use of a source-vocal tract interaction based on acoustic inertance below the third formant. Acoustic inertance of the vocal tract helps to set the vocal folds into vibration.2 For a female vocal tract (see eleven vowel shapes in Figure 1a), the third formant occurs above 3000 Hz and vocal tract inertance is found for most vowel shapes if frequencies are in the 1500-3000 Hz range. This range is on the upskirt of the third formant F^sub 3^ (see Figure 1b, where formants F^sub 1^, F^sub 2^, and F^sub 3^ are labeled for the vowel /o/). Formants (the resonances of the vocal tract) are where the inertance curves make an upward turn followed by a sharp drop. When the inertance is high (above the zero line), source frequencies are reinforced. In particular, the fundamental frequency F^sub 0^ gets a boost. Note that for the /o/ vowel shape, the inertance curve is above zero for pitches slightly below and above C^sub 7^. But the higher harmonics (2F^sub 0^, 3F^sub 0^, 4F^sub 0^ ... ) are not systematically reinforced. They face variable inertances from the vocal tract. Note that the harmonics 2F^sub 0^, 3F^sub 0^, and 4F^sub 0^ lie in the midst of a cluster of formants above F^sub 3^. Similar situations occur for most of the other vowel shapes. Thus, the frequency range below and above C^sub 7^ (1500-3000 Hz) is particularly advantageous and consistent for the fundamental F^sub 0^, but quite irregular for higher harmonics. Two examples of spectrograms of whistle voice of recorded artists are shown in Figures 2 and 3. Figure 2 is from Georgia Brown, a Brazilian recording artist.3 She produces a long sustained whistle note at F^sub 0^ = 2400 Hz (about D^sub 7^) without vibrato, followed by a lower note with vibrato at 1860 Hz (about B-flat^sub 6^). For both notes, the fundamental F^sub 0^ is very strong, apparently reinforced by vocal tract inertance. The second harmonic 2F^sub 0^ is also strong, but no higher harmonics are evident in the first note. In the second note, the third harmonic has some energy. The exact vowel shape used by the performer is not known, but it would appear from Figure lb that vowel shapes like /a/ and /i/ would have both F^sub 0^ and 2F^sub 0^ reinforcement, whereas a shape like /c/ could also have 3F^sub 0^ reinforcement. The second example (Figure 3) is a spectrogram of Mariah Carey, an American female pop singer.4 In the recording, she produces a series of vocal glides to a maximum fundamental frequency of 2400 Hz (about D^sub 7^, where F^sub 0^ is labeled). Some energy is seen in both 2F^sub 0^ and 3F^sub 0^, but the energy is mostly in F^sub 0^ at all pitches. …
OpenAlex reports 3 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
YOUNG ADULT FEMALES SOMETIMES STUMBLE upon a very high pitched piccolo-like sound they can produce in their larynx. It is known as whistle voice, or whistle register. When first experienced by a singer, it often startles her as much as a listener. It seems to be produced with ease at pitches starting from G^sub 6^ (about 1570 Hz) to G^sub 7^ (3140 Hz). Some authors have claimed that whistle voice can start below 1000 Hz, especially in untrained females,1 but trained lyric and coloratura sopranos usually sing notes well above 1000 Hz in nonwhistle phonations. The hypothesis here is that whistle voice makes use of a source-vocal tract interaction based on acoustic inertance below the third formant. Acoustic inertance of the vocal tract helps to set the vocal folds into vibration.2 For a female vocal tract (see eleven vowel shapes in Figure 1a), the third formant occurs above 3000 Hz and vocal tract inertance is found for most vowel shapes if frequencies are in the 1500-3000 Hz range. This range is on the upskirt of the third formant F^sub 3^ (see Figure 1b, where formants F^sub 1^, F^sub 2^, and F^sub 3^ are labeled for the vowel /o/). Formants (the resonances of the vocal tract) are where the inertance curves make an upward turn followed by a sharp drop. When the inertance is high (above the zero line), source frequencies are reinforced. In particular, the fundamental frequency F^sub 0^ gets a boost. Note that for the /o/ vowel shape, the inertance curve is above zero for pitches slightly below and above C^sub 7^. But the higher harmonics (2F^sub 0^, 3F^sub 0^, 4F^sub 0^ ... ) are not systematically reinforced. They face variable inertances from the vocal tract. Note that the harmonics 2F^sub 0^, 3F^sub 0^, and 4F^sub 0^ lie in the midst of a cluster of formants above F^sub 3^. Similar situations occur for most of the other vowel shapes. Thus, the frequency range below and above C^sub 7^ (1500-3000 Hz) is particularly advantageous and consistent for the fundamental F^sub 0^, but quite irregular for higher harmonics. Two examples of spectrograms of whistle voice of recorded artists are shown in Figures 2 and 3. Figure 2 is from Georgia Brown, a Brazilian recording artist.3 She produces a long sustained whistle note at F^sub 0^ = 2400 Hz (about D^sub 7^) without vibrato, followed by a lower note with vibrato at 1860 Hz (about B-flat^sub 6^). For both notes, the fundamental F^sub 0^ is very strong, apparently reinforced by vocal tract inertance. The second harmonic 2F^sub 0^ is also strong, but no higher harmonics are evident in the first note. In the second note, the third harmonic has some energy. The exact vowel shape used by the performer is not known, but it would appear from Figure lb that vowel shapes like /a/ and /i/ would have both F^sub 0^ and 2F^sub 0^ reinforcement, whereas a shape like /c/ could also have 3F^sub 0^ reinforcement. The second example (Figure 3) is a spectrogram of Mariah Carey, an American female pop singer.4 In the recording, she produces a series of vocal glides to a maximum fundamental frequency of 2400 Hz (about D^sub 7^, where F^sub 0^ is labeled). Some energy is seen in both 2F^sub 0^ and 3F^sub 0^, but the energy is mostly in F^sub 0^ at all pitches. …
Key concepts: Inertance, Formant, Vocal tract, Vibrato, Acoustics, Vowel, Vocal folds, Speech recognition