Measurement of Air Volumes

Respiratory research in human communication has focused primarily on the measurement of the air volumes that are typically expended during selected speech and singing tasks, and on specifying the ranges of lung inflation levels across which such tasks are normally performed (cf. Hixon, Goldman, and Mead, 1973; Watson and Hixon, 1985; Hoit and Hixon, 1987; Hoit et al., 1990). Air volumes are measured in standard metric units (liters, cubic centimeters, milliliters) and lung inflation levels are usually specified in terms of a percentage of the vital capacity or total lung volume.

Both direct and indirect methods have been used to measure air volumes expended during phonation. Direct measurement of orally displaced air volumes during phonatory tasks can be accomplished, to a limited extent, by means of a mouthpiece or face mask connected to a measurement device such as a spirometer (Beckett, 1971) or pneumotachograph (Isshiki, 1964). The use of a mouthpiece essentially limits speech production to sustained vowels, which are sufficient for assessing selected volumetric-based phonatory parameters. There are also concerns that face masks interfere with normal jaw movements and that the oral acoustic signal is degraded, so that auditory feedback is reduced or distorted and simultaneous acoustic analysis is limited. These limitations, which are inherent to the use of devices placed in or around the mouth to directly collect oral airflow, plus additional measurement-related restrictions (Hillman and Kobler, 2000) have helped motivate the development and application of indirect measurement approaches.

Most speech breathing research has been carried out using indirect approaches for estimating lung volumes by means of monitoring changes in body dimensions. The basic assumption underlying the indirect approaches is that changes in lung volume are reflected in proportional changes in body torso size. One relatively cumbersome but time-honored approach has been to place subjects in a sealed chamber called a body plethysmograph to allow estimation of the air volume displaced by the body during respiration (Draper, Ladefoged, and Whitteridge, 1959). More often used for speech breathing research are transducers (magnetometers: Hixon, Goldman, and Mead, 1973; inductance plethysmographs: Sperry, Hillman, and Perkell, 1994) that unobtrusively monitor changes in the dimensions of the rib cage and abdomen (referred to collectively as the chest wall) that account for the majority of respiratory-related changes in torso dimension (Mead et al., 1967). These approaches have been primarily employed to study respiratory function during continuous speech and singing tasks that include both voiced and voiceless sound production, as opposed to assessing air volume usage during phonatory tasks that involve only laryngeal production of voice (e.g., sustained vowels). There are also ongoing efforts to develop more accurate methods for noninvasively monitoring chest wall activity to capture finer details of how the three-dimensional geometry of the body is altered during respiration (see Cala et al., 1996).

Measurement of Airflow

Airflow associated with phonation is usually specified in terms of volume velocity (i.e., volume of air displaced per unit of time). Volume velocity airflow rates for voice production are typically reported in metric units of volume displaced (liters or cubic centimeters) per second.

Estimates of average airflow rates can be obtained by simply dividing air volume estimates by the duration of the phonatory task. Average glottal airflow rates have usually been estimated during vowel phonation by using a mouthpiece or face mask to channel the oral air stream through a pneumotachograph (Isshiki, 1964). There has also been somewhat limited use of hot wire anemometer devices (mounted in a mouthpiece) to estimate average glottal airflow during sustained vowel phonation (Woo, Colton, and Shangold, 1987). Estimates of average glottal airflow rates can be obtained from the oral airflow during vowel production because the vocal tract is relatively nonconstricted, with no major sources of turbulent airflow between the glottis and the lips.

There have also been efforts to obtain estimates of the actual airflow waveform that is generated as the glottis rapidly opens and closes during flow-induced vibration of the vocal folds (the glottal volume velocity waveform). The glottal volume velocity waveform cannot be directly observed by measuring the oral airflow signal because the waveform is highly convoluted by the resonance activity (formants) of the vocal tract. Thus, recovery of the glottal volume velocity waveform requires methods that eliminate or correct for the influences of the vocal tract. This has typically been accomplished aerodynamically by processing the output of a fast-responding pneumotachograph (high-frequency response) using a technique called inverse filtering, in which the major resonances of the vocal tract are estimated and the oral airflow signal is processed (inverse filtered) to eliminate them (Rothenberg, 1977; Holmberg, Hillman, and Perkell, 1988).

Measurement of Air Pressure

Measurements of air pressures below (subglottal) and above (supraglottal) the vocal folds are of primary interest for characterizing the pressure differential that must be achieved to initiate and maintain vocal fold vibration during normal exhalatory phonation. In practice, air pressure measurements related specifically to voice production are typically acquired during vowel phonation when there are no vocal tract constrictions of sufficient magnitude to build up positive supraglottal pressures. Under these conditions, it is usually assumed that supraglottal pressure is essentially equal to atmospheric pressure and only subglottal pressure measurements are obtained. Air pressures associated with voice and speech production are usually specified in centimeters of water (cm H₂O).

Both direct and indirect methods have been used to measure subglottal air pressures during phonation. Direct measures of subglottal air pressure can be obtained by inserting a hypodermic needle into the subglottal airway through a puncture in the anterior neck at the cricothyroid space (Isshiki, 1964). The needle is connected to a pressure transducer by tubing. This method is very accurate but also very invasive. It is also possible to insert a very thin catheter through the posterior cartilaginous glottis (between the arytenoids) to sense subglottal air pressure during phonation, or to use an array of miniature transducers positioned directly above and below the glottis (Cranen and Boves, 1985). These methods cannot be tolerated by all subjects, and the heavy topical anesthetization of the larynx that is required can affect normal function.

Indirect estimates of tracheal (subglottal) air pressure can be obtained via the placement of an elongated balloon-like device into the esophagus (Liberman, 1968). The deflated esophageal balloon is attached to a catheter that is typically inserted transnasally and then swallowed into the esophagus to be positioned at the midthoracic level. The catheter is connected to a pressure transducer and the balloon is slightly inflated. Accurate use of this invasive method also requires simultaneous monitoring of lung volume.

Noninvasive, indirect estimates of subglottal air pressure can be obtained by measuring intraoral air pressure during specially constrained utterances (Smitheran and Hixon, 1981). This is usually done by sensing air pressure just behind the lips with a translabially placed catheter connected to a pressure transducer. These intraoral pressure measures are obtained as subjects produce strings of bilabial /p/ + vowel syllables (e.g., /pi-pi-pi-pi-pi/) at constant pitch and loudness. This method works because the vocal folds are abducted during /p/ production, thus allowing pressure to equilibrate throughout the airway, making intraoral pressure equal to subglottal pressure (Fig. 1).

Figure 1..  

Instrumentation and resulting signals for simultaneous collection of oral airflow, intraoral air pressure, the acoustic signal, and chest wall (rib cage and abdomen) dimensions during production of the syllable string /pi-pi-pi/. Signals shown in the bottom panel are processed and measured to provide estimates of average glottal airflow rate, average subglottal air pressure, lung volume, and glottal waveform parameters.

Additional Derived Measures

There have been numerous attempts to extend the utility of aerodynamic measures by using them in the derivation of additional parameters aimed at better elucidating underlying mechanisms of vocal function. Such derived measures usually take the form of ratios that relate aerodynamic parameters to each other, or that relate aerodynamic parameters to simultaneously obtained acoustic measures. Common examples include (1) airway (glottal) resistance (see Smitheran and Hixon, 1981), (2) vocal efficiency (Schutte, 1980; Holmberg, Hillman, and Perkell, 1988), and (3) measures that interrelate glottal volume velocity waveform parameters (Holmberg, Hillman, and Perkell, 1988).

Normative Data

As is the case for most measures of vocal function, there is not currently a set of normative data for aerodynamic measures that is universally accepted and applied in research and clinical work. Methods for collecting such data have not been standardized, and study samples have generally not been of sufficient size or appropriately stratified in terms of age and sex to ensure unbiased estimates of underlying aerodynamic phonatory parameters in the normal population. However, there are several sources in the literature that provide estimates of normative values for selected aerodynamic measures (Kent, 1994; Baken, 1996; Colton and Casper, 1996).

See also voice production: physics and physiology.