Speech Production

Speech is produced by joint activation of three subsystems—respiratory, phonatory, and articulatory (see Figure 1). The respiratory system produces an outflow of air under relatively constant pressure. The modulation of this airflow to produce the sounds we know of as speech is characterized by the source-filter theory, according to which the phonatory system provides the sound source for speech and the articulatory system filters this source.

The phonatory system acts on the outward airflow of the respiratory system to produce pho-nation, or voicing. The sound source is the vibration of the vocal cords (or folds), which are two opposed horizontal muscular slivers running from front to back within the cartilaginous larynx. They are drawn apart during quiet breathing. Prior to the first cycle of vocal fold vibration for speech, the vocal folds are brought together. This blocks the airway causing air pressure to build up beneath the folds. The increasing pressure evokes an elliptical opening between the folds through which a puff of air flows. But the opening evokes two counter-effects that restore the closed state. They are the elastic recoil of the stretched folds and the Bernoulli effect—the development of low pressure in regions of high flow—which sucks the folds together. The folds close immediately, pressure builds to eventually blow them apart again, and the next cycle continues. In an adult male phonating a vowel, the rate of vibration is about 125 hertz (Hz) (cycles per second). The phonatory system also produces a basic distinction between voiced consonants (during which the vocal folds are vibrating), and voiceless consonants (during which they do not vibrate). Vibration is eliminated by moving the folds laterally (to the sides) into the open breathing position. In this position, they are not activated by the air flow.

The successive puffs of air generated by vocal fold vibration enter the supralaryngeal vocal tract—the air space between the larynx and the lips. They not only produce vibration that propagates through the air to the listener at their repetition rate, called the fundamental frequency (f0), but they also produce vibration at the first 40 or so whole numbered multiples (harmonics) of f0.

The main role of the articulatory system is to modulate the acoustic package (collectively known as the source spectrum), which enters the supra-laryngeal vocal tract. The supralaryngeal vocal tract (oral and pharyngeal cavities) acts as a filter. It provides a set of variable resonators selectively amplifying different parts of the source spectrum for vowels and voiced consonants depending on the size and shape of cavities formed in the tract.

Figure 1 Sketch of the Three Subcomponents of the Speech Production Apparatus

This is analogous to how the tuning knob on a hi-fi stereo selectively amplifies a certain small subset of radio frequencies in each individual position that one turns it to. In the simplest speech-related case of the so-called neutral vowel (e.g., the first vowel in the phrase “the book”), the tract assumes the shape of a single tube of uniform cross-sectional area. In a typical male adult, this configuration maximally amplifies frequencies of 500, 1500, and 2500 Hz. In general terms, these frequencies are known as resonant frequencies, but in the specific context of speech they are called formants. For other vowels, and voiced consonants, involving a nonuniform tract, the oral and pharyngeal cavities tend to make separate contributions, with the pharyngeal cavity responsible for the first (lowest) formant, and the oral cavity responsible for the second. Topics discussed in this entry include the syllable, individual consonants and vowels, suprasegmental phenomena, and Tinbergen's four questions.

...