Physiological Processes of Speech Production--Reading Notes (4)

标签：speech production   speech science   

Vocal Fold and its Oscillation

The larynx includes several structures such as the subglottic dome, vocal folds, ventricles, vestibular folds, epiglottis, and aryepiglottic folds, as shown in Fig. 4a. The vocal folds run anteroposteriorly from the vocal processes of the arytenoid cartilages to the internal surface of the thyroid cartilage. The vocal fold tissue consists of the thyroarytenoid muscle, vocal ligament, lamina propria, and mucous membrane. They form a special layer structure that yields to aerodynamic forces to oscillate, which is often described as the body-cover structure.

During voiced speech sounds, the vocal folds are set into vibration by pressurized air passing through the membranous portion of the narrowed glottis. The glottal airflow thus generated induces wave-like motion of the vocal fold membrane, which appears to propagate from the bottom to the top of the vocal fold edges. When this oscillatory motion builds up, the vocal fold membranes on either side come into contact with each other, resulting in repetitive closing and opening of the glottis. Figure 4b shows that vocal fold vibration repeats four phases within a cycle: the closed phase, opening phase, open phase, and closing phase. The conditions that determine vocal fold vibration are
the stiffness and mass of the vocal folds, the width of the glottis, and the pressure difference across the glottis. 

The aerodynamic parameters that regulate vocal fold vibration are the transglottal pressure difference and glottal airflow. The former coincides with the measure of subglottal pressure during mid and low vowels, which is about 5–10 cm H2O in comfortable loudness and pitch (1 cm H2O = 0.98 hPa). The latter also coincides with the average measure of oral airflow during vowel production, which is roughly 0.1–0.2 l/s. These values show a large individual variation: the pressure range is 4.2–9.6 cm H2O in males and 4.4–7.6 cm H2O in females, while the airflow rate ranges between 0.1–0.3 l/s in males and 0.09–0.21 l/s in females.

Figure 4a,b: Vocal folds and their vibration pattern. (a) Coronal section of the larynx, 

showing the tissues of the vocal and vestibular (false) folds. The cavity of the larynx 

includes supraglottic and subglottic regions. (b) Vocal-fold vibration pattern and 

glottal shapes in open phases. As the vocal-fold edge deforms in a glottal cycle, 

the glottis follows four phases: closed, opening, open and closing

Figure 5 shows schematically the relationship between the glottal cycle and volumic airflow change in normal and breathy phonation. The airflow varies within each glottal cycle, reflecting the cyclic variation of the glottal area and subglottal pressure. The glottal area curve roughly shows a triangular pattern, while the airflow curve shows a skew of the peak to the right due to the inertia of the air mass within the glottis. The closure of the glottis causes a discontinuous decrease of the glottal airflow to zero, which contributes the main source of vocal tract excitation,

as shown in Fig. 5a. When the glottal closure is more abrupt, the output sounds are more intense with richer harmonic components. When the glottal closure is incomplete in soft and breathy voices or the cartilaginous portion of the glottis is open to show the glottal chink, the airflow includes a direct-current (DC) component and exhibits a gradual decrease of airflow, which results in a more sinusoidal waveform and a lower intensity of the output sounds, as shown in Fig. 5b.

Laryngeal control of the oscillatory patterns of the vocal folds is one of the major factors in voice quality control. In sharp voice, the open phase of the glottal cycle becomes shorter, while in soft voice, the open phase becomes longer. The ratio of the open phase within a glottal cycle is called the open quotient (OQ), and the ratio of the closing slope to the opening slope in the glottal cycle is called the speed quotient (SQ). These two parameters determine the slope of the spectral envelope. When the open phase is longer (high OQ) with a longer closing phase (low SQ), the glottal airflow becomes more sinusoidal, with weak harmonic components. Contrarily, when the open phase is shorter (low OQ), glottal airflow builds up to pulsating waves with rich harmonics. In modal voice, all the vocal fold layers are involved in vibration, and the membranous glottis is completely closed during the closed phase of each cycle. In falsetto, only the edges of the vocal folds vibrate, glottal closure becomes incomplete, and harmonic components reduce remarkably. 

The oscillation of the vocal folds during natural speech is quasiperiodic, and cycle-to-cycle variation are observed in speech waveforms as two types of measures: jitter (frequency perturbation) and shimmer (amplitude perturbation). These irregularities appear to arise from combinations of biomechanical (vocal fold asymmetry), neurogenic (involuntary activities of laryngeal muscles), and aerodynamic (fluctuations of airflow and subglottal pressure) factors. In sustained phonation of normal voice, the jitter is about 1% in frequency, and the shimmer is about 6% in amplitude.

1.5 glottal cycles from glottal opening, with glottal shapes at peak opening (in the 

circles). (a) In modal phonation with complete glottal closure in the closed phase, 

glottal closure causes abrupt shut-off of glottal airflow and strong excitation of the 

air in the vocal tract during the closed phase. (b) In breathy phonation, the glottal 

closure is incomplete, and the airflow wave includes a DC component, which results 

in weak excitation of the tract

Physiological Processes of Speech Production--Reading Notes (4)

标签：speech production   speech science   

原文地址：http://blog.csdn.net/jojozhangju/article/details/46847243