Tutorials - Voice Production

About The National Center for Voice and Speech Summer Vocology Institute NCVS Groups Library Tutorials Information Links


About NCVS
NCVS Locations
NCVS People

Graphing Phonation: Waves & Spectra

Glottal Flow Waveform

Study the two cycles of a typical glottal flow waveform above. The variables shown are as follows:
Tp the part of the glottal cycle in which airflow rate is increasing (thus, the slope of the curve is going upward during this time). Think of 'p' for positive.
Tn the part of the glottal cycle in which airflow rate is decreasing (downward slope in the curve). Here, 'n' represents negative.
To the length of time during each cycle in which air is flowing (i.e., the folds are open)
T the total duration of each vibrational cycle
uo the average rate of airflow
uac maximum rate of flow

In addition to the variables shown in the graph, two other useful variables can be derived, which define how the waveform is shaped:

  • Qo, the open quotient, which is equal to To / T. This can be expressed simply as the percentage of time in each cycle during which the folds are open.

  • Qs, the skewing quotient, which is equal to Tp / Tn. This is the portion of time in each cycle during which the folds are moving outward, divided by the time in which they are moving inward. More simply, the skewing quotient is a number that explains how far from symmetric (how skewed) the waveform "bump" is.
The shape of the glottal waveform helps to determine the loudness of the sound generated, and also its timbre. A 'jagged' waveform (with a large skewing quotient or small open quotient) which represents sudden changes in airflow will produce more high frequencies, and result in a 'brassy' timbre. A smoother curve, representative of gradual changes in airflow (open quotient and skewing quotient both high) will tend to produce a more 'fluty' sound.

A vocalist can control the open quotient (To) by moving the vocal processes (tips of arytenoid cartilages at one end of each vocal fold) closer together or further apart during phonation.

Secondary Sound Sources in the Larynx
In addition to the desired vocal sound produced by the larynx, other sounds are also often produced. Turbulence, a 'non-smooth' flow in the glottal airstream, can create a 'hissing' sound, which is called aspiration if it is combined with sound from vocal fold vibration, and whisper if it is not. Aspiration is a contributing factor to 'breathy' voice.

Other sources of extra sounds in the larynx include glottal 'clicks' or 'stops', which are produced by an especially effortful and sudden opening or closing of the vocal folds. Also, any extra mucus or other fluids which are on or near the vocal folds can produce a rough or 'gurgly' noise as they move around during phonation.

Audible Frequencies in Sound Sources
Generally, human detection of sound is limited to waves having frequencies from 20 Hz to 20,000 Hz. The precise limits of hearing vary on an individual basis, and tend to degrade with age and other causes of hearing loss. The sensitivity of human hearing is not constant across this frequency range; frequencies in the thousands of hertz (1,000 to 10,000 Hz) are usually more easily heard than those at the extremes of the audible range. Audiophiles enhance these less discernable frequencies with various graphic equalizers, woofers, and tweeters.

Frequency Spectra
All of the sounds we hear in everyday life are made up of many component frequencies, intermixed in varying proportion to one another. A sound spectrum is an array of these components of an sound, separated and arranged in order of frequency. Each component frequency is a pure sine-wave tone; the graph of the shape of such a pure tone is called a sinusoid. In general, an infinite collection of sinusoids is needed to construct any complex sound.

Spectral Slope
The spectral slope of a given spectrum describes how rapidly the amplitudes of successive partials (component frequencies) decrease as they get higher in frequency. Three examples of spectral slopes are shown below:

Spectral slope influences the timbre of the sound, just as waveform shape does, as described above. A spectral slope of around 6 dB/octave, the least severe slope in the graph, results in stronger high frequencies, which yield a more 'brassy' or strident sound. The middle slope depicted, 12 dB/octave, is that of a normal vocal quality. The most extreme slope shown, 18 dB/octave, would result in a more 'fluty' sound; it has stronger low frequencies, as compared to the higher ones, which rapidly drop off in strength.

Tutorials Homepage                                                                  

The National Center for Voice and Speech is a Division of The Denver Center for the Performing Arts and a Center at The University of Iowa.
Site Map - a text-based navigation of the website                                                               Contact the Webmaster
© National Center for Voice and Speech ~ National Center for Voice and Speech, www.ncvs.org
National Institute on Deafness and Other Communication Disorders