This diagram shows the relationship between subglottal and intraoral air pressure.
Here are three graphs obtained simultaneously. The dotted line on the lowest graph is subglottal air pressure, obtained with a tracheal puncture. A hypodermic needle is hooked to a transducer, and it measures pressure below the vocal folds.
The whole pressure trace is a breath group (the phrase “phonetic aspects.”). The dotted line is lung pressure—we have to maintain a fairly constant driving pressure (8cm of water) for speech. At the end of the trace, pressure falls off to 0 (REL, or respiratory expiration level). The Po trace is intraoral air pressure, between the vocal folds and teeth. It was made by putting a pressure transducer on a wire, through a mask, behind the tongue to account for tongue movements of /k/ and /g/. A pressure transducer has to be behind the most posterior constriction, and in English, that constriction is at the velum for the sounds /k/ and /g/. The mask used to record this trace covers the mouth and nose, and there’s a partition between the oral and nasal areas. With the sounds /k, s, p/--the voiceless consonants, the glottis is open. With the /t/, there is some voicing. If you say the phrase “phonetic aspects,” the /t/ sounds more like /d/.
If we look at the oral airflow graph, we see that the peak flow rate is 2 times as great for voiceless consonants as it is for voiced sounds. We can see higher peaks on the sounds /f, k/ and /s/, with pressure around 500; the other sounds are around 0. Look at the /n/, and compare the nasal air flow and oral air flow. The nasal airflow is about 250 cc, which is about ½ that of voiceless consonants. Voiced consonants have greater resistance to air flow because of the vocal folds. The vibrating vocal folds are the first resistor to air flow. We see in the middle trace the oral airflow for voiced consonants is at “half mast”—it is not as large because of some impedance from vocal folds.
If we go to the bottom graph, for vowels, intraoral air pressure is 0 – because there is no constriction in the vocal tract. In the production of the vowels in this phrase, the vocal folds are vibrating but the oral cavity is open. The air pressure is developed below the vocal folds—air is impeded somewhat. Impedance is high because the vocal folds are touching. The pressure below vocal folds is very large. The example of /ae/ probably carries the major linguistic stress of the phrase. The subglottal pressure is 8 cm H20 and Po (intraoral air pressure) is 0. Subglottal pressure tends to be the same as the atmosphere. The reason for this is that during production of vowels, the vocal tract is relatively open; the mouth is open.
There are 2 major classes of consonants: obstruents and sonorants. Obstruents are also called pressure consonants, because pressure builds up in the oral cavity. Obstruents are stops, fricatives, and affricates. Here, they are the /k/ and /t/, /s/ and /f/. With voiceless obstruents, oral pressure = subglottal pressure, because the vocal folds are open. See the peaks in the air pressure graph at the bottom. We do see the cessation in airflow with the stops (the /k/ and /p/), and all of the release of the consonant occurs through the mouth.
Looking at the top trace, with /n/, you can see co-articulation. Some nasality spreads over to the other adjacent speech sounds. The /t/ is more like a /d/—the vocal folds are partially together. There is a stoppage of air at the larynx (vocal folds) and in the alveolar ridge.
Let’s look at the fricatives. /s/ has quite a bit of airflow—about 500 cc. With this fricative, oral pressure = subglottal pressure, because it is voiceless.