Figure 1: Cross-section of the laryngeal tube. Note the ventricular folds labeled as the vestibular folds (also known as the false vocal folds) and the vocal folds (Larynx, Online Image).

Figure 2: A tube model illustrating the vocal tract. The vocal folds are illustrated at the left end of the model and the larynx tube and ventricular fold are depicted as constrictions within the vocal tract. This particular model is illustrating constrictions in the vocal tract during Tuvan Throat Singing (Imagawa et al., 2003).

Modal Voicing Data

Figure 3: Spectrogram, waveform, and textgrid marking vowel boundaries and preceding [w] boundary. This is the comparative modal spectrogram for the following data. Note the amount of white noiseless space and banding of the first formants. The formants are labeled by their number.

Figure 4: Waveform of modal voicing displaying three glottal cycles.

Figure 5: Formant tracking of modal voicing during vowel articulation of first three formants.

Creaky Voice Data

Figure 6: Spectrogram, waveform, and textgrid representations. Note the vertical banding in the spectrogram corresponding to the high amplitude pulses in the waveform.

Figure 7: Three creaky voice glottal cycles in waveform.

Figure 8: Tracking the first three formants of the creaky voice during vowel articulation.

Breathy Voice Data

Figure 9: Spectrogram, waveform, and textgrid illustrating breathy voicing style. Note the high level of low amplitude noise compared to modal voicing.

Figure 10: Three glottal cycles in breathy voice displayed as waveform.

Figure 11: Tracking the first three formants of breathy voice.

Raspy Voice Data

Figure 12: Spectrogram, waveform, and textgrid illustrating raspy voice. Again, there is a high level of noise in the spectrogram. The first two formants are less obscured than breathy voice and there is a zero (white space) around 4234Hz.

Figure 13: Waveform displaying three glottal cycles.

Figure 14: The first three formants of the rasp tracked.

Scream Data

Figure 15: Scream spectrogram, waveform, and textgrid represented. We see the highest amplitude noise with this voicing type obscuring the third formant. The overall frequency of the formants in scream voicing are higher than modal and the other voicing techniques.

Figure 16: Three glottal cycles with scream voicing.

Figure 17: Formant tracking of the first three formants in scream.

Death Grunt Data

Figure 18: Death grunt spectrogram, waveform, and textgrid with first three formants labeled. Notice the zero that extends from [w] into the vowel and the high amplitude noise above the third formant.

Figure 19: Three death grunt glottal cycles shown as soundwave.

Figure 20: Death grunt formant track of the first three formants.

Additional Data

Figure 21: Spectrogram of raspy voicing technique drawn with narrow banding to give more details. Notice the bands below the first formant. These represent the vibrations of the ventricular during articulation in raspy voice.

Works Referenced

The Larynx. Online Image. 7 April. 2009 <http://www.hopkinsmedicine.org/voice/anatomy.html>.

Imagawa, Hiroshi, Sakakibara, Ken-Ichi, Tayama, Niro, and Seji Niimi. “The effect of the hypopharyngeal and supra-glottic shapes on the singing voice” Stockholm Music Acoustics Conference August 2003: SMAC 1-4.