Decoding the Human Voice: An Introduction to Phonetics

All speech starts with moving air. The vast majority of sounds in the world's languages are pulmonic egressive. This simply means air is pushed out (egressive) from the lungs (pulmonic). Take a deep breath and say "hello." You are using a pulmonic egressive airstream.

However, other mechanisms exist. Glottalic Airstreams: The air is moved by the action of the glottis (the space between your vocal folds). This produces ejectives (egressive, like the "p'" sound in some Caucasian languages, which sounds like a sharp, spit-out 'p') and implosives (ingressive, where air is sucked in, common in some African and Southeast Asian languages). Velaric Airstreams: The air is moved by the back of the tongue against the velum (soft palate). This is how click consonants are made, famous in the Khoisan languages of Southern Africa. The "tsk-tsk" sound of disapproval is a dental click.

For the rest of our discussion, we will focus on pulmonic egressive sounds, as they form the basis of English and most other languages.

As air travels up from the lungs, it passes through the larynx, or voice box. Inside the larynx are the vocal folds (or vocal cords). The state of these folds is a critical factor in sound production. This process is called phonation.

The primary distinction here is voicing.

Voiceless: The vocal folds are held apart, allowing air to pass through freely without vibrating. Produce a long, sustained /s/ sound (like a snake hiss: "ssssss"). Now, place your fingers gently on your throat. You will feel no vibration. Sounds like /p/, /t/, /k/, /f/, and /s/ are voiceless.

Voiced: The vocal folds are brought close together, and the air passing through causes them to vibrate rapidly. Now, produce a long, sustained /z/ sound ("zzzzzz"). With your fingers on your throat, you will feel a distinct buzzing vibration. That's voicing. Sounds like /b/, /d/, /g/, /v/, and /z/, as well as all vowels, are voiced.

This single distinction—voiced versus voiceless—is one of the most fundamental contrasts in phonetics, responsible for differentiating pairs of words like pat and bat, or fan and van.

After passing the larynx, the airstream enters the vocal tract (the pharynx, oral cavity, and nasal cavity), where it is shaped into specific sounds by the articulators. Some articulators are passive (the upper teeth, the alveolar ridge, the hard palate), while others are active (the tongue, the lower lip).

We describe consonants using a three-part label: Voicing - Place of Articulation - Manner of Articulation. We've already covered voicing. Let's explore the other two.

Place of Articulation: Where the Sound is Made

This refers to the point in the vocal tract where the airstream is constricted. We describe it by moving from the front of the mouth to the back.

Bilabial: Both lips are used. Examples: [p] (pat), [b] (bat), [m] (mat). Labiodental: The lower lip contacts the upper teeth. Examples: [f] (fan), [v] (van). Dental: The tip of the tongue contacts the back of the upper teeth. Examples: [θ] (thin), [ð] (this). Alveolar: The tip or blade of the tongue contacts the alveolar ridge, the bony ridge just behind the upper teeth. This is a very common place of articulation in English. Examples: [t] (top), [d] (dog), [n] (no), [s] (sip), [z] (zoo), [l] (lip). Postalveolar: The tongue contacts the area just behind the alveolar ridge. Examples: [ʃ] (ship), [ʒ] (measure), [tʃ] (chip), [dʒ] (judge). Retroflex: The tip of the tongue is curled back to contact the area behind the alveolar ridge. While not prominent in standard British or American English, it's the characteristic "r" sound in many varieties of Indian English. The IPA symbol is [ɻ]. Palatal: The body of the tongue contacts the hard palate (the roof of your mouth). Example: [j] (yes). Velar: The back of the tongue contacts the velum, or soft palate. Examples: [k] (cat), [g] (go), [ŋ] (sing). Uvular: The back of the tongue contacts the uvula, the small fleshy appendage hanging at the back of the throat. Not found in English, but the "r" sound in French and German is often a uvular fricative or trill. Example: [ʁ] (the French 'r' in Paris). Pharyngeal: The root of the tongue is retracted into the pharynx (the throat). These are characteristic of Semitic languages like Arabic. Glottal: The constriction occurs at the glottis itself. Examples: [h] (hat), and the glottal stop [ʔ] (the sound in the middle of "uh-oh").

Manner of Articulation: How the Sound is Made

This describes the type and degree of constriction of the airstream.

Plosive (or Stop): The airstream is completely blocked, pressure builds up, and is then released in a burst. Examples: [p], [b], [t], [d], [k], [g], [ʔ]. Nasal: The velum is lowered, allowing air to escape through the nasal cavity while the oral cavity is blocked at some point. Examples: [m], [n], [ŋ]. All nasals are typically voiced. Fricative: The articulators are brought close together, but not so close as to block the airstream completely. The air is forced through a narrow channel, creating turbulence and a hissing sound. Examples: [f], [v], [θ], [ð], [s], [z], [ʃ], [ʒ], [h]. Affricate: This is a combination of a plosive followed immediately by a fricative at the same place of articulation. The sound begins as a stop and is released as a fricative. Examples: [tʃ] (as in church), [dʒ] (as in judge). Approximant: The articulators are brought near each other, but not close enough to create turbulent airflow. They are the most vowel-like consonants. Liquids: [l] (a lateral approximant, where air flows around the sides of the tongue) and [ɹ] (the typical 'r' in American English). Glides (or Semivowels): [w] (a voiced labial-velar glide, as in we) and [j] (a palatal glide, as in yes). Trill: An articulator (like the tongue tip or the uvula) is held in a position where the airstream causes it to vibrate rapidly. The "rolled r" in Spanish (perro) is an alveolar trill, [r]. Tap (or Flap): A very rapid, ballistic articulation where the tongue strikes an articulator and immediately moves away. The sound in the middle of the American English pronunciation of butter or writer is an alveolar flap, [ɾ].

With this three-part system, we can give a precise, unambiguous description for any consonant. For example, [s] is a voiceless alveolar fricative. [g] is a voiced velar plosive. [m] is a voiced bilabial nasal.

Vowels are fundamentally different from consonants. While consonants involve a significant constriction in the vocal tract, vowels are produced with a relatively open vocal tract. The specific sound of a vowel is determined by the shape and volume of the oral cavity, which is controlled by the position of the tongue and the lips.

We describe vowels using three main criteria:

1. Vowel Height: How high or low the tongue is in the mouth. High (or Close): The tongue is raised high in the mouth. Examples: the vowel in seet [i], the vowel in suit [u]. Mid: The tongue is in a neutral, intermediate position. Examples: the vowel in set [ɛ], the vowel in thought [ɔ]. Low (or Open): The tongue is low in the mouth, and the jaw is dropped. Example: the vowel in father [ɑ].
2. Vowel Backness: How far forward or backward the highest point of the tongue is. Front: The tongue is pushed forward. Examples: seet [i], set [ɛ], sat [æ]. Central: The tongue is in a neutral position. The most common central vowel is schwa [ə], the unstressed vowel sound in the first syllable of a*bout or the last syllable of soda. Back: The tongue is pulled back. Examples: suit [u], thought [ɔ], father [ɑ].
3. Lip Rounding: Whether the lips are rounded or unrounded (spread). Rounded: The lips are pursed. In English, back vowels are typically rounded. Examples: suit [u], thought [ɔ]. Unrounded: The lips are in a neutral or spread position. In English, front vowels are unrounded. Examples: seet [i], set [ɛ].

Phoneticians map these vowels onto a chart called the Vowel Quadrilateral, which represents the space inside the oral cavity.

A special category of vowels is diphthongs, which involve a glide from one vowel position to another within a single syllable. Examples in English include the vowel in buy [aɪ] (a glide from a low vowel to a high front one) and the vowel in boy* [ɔɪ].

Frequency: This is the rate at which the sound wave vibrates, measured in Hertz (Hz). We perceive frequency as pitch. A higher frequency means a higher pitch. The fundamental frequency (F0) of a speech sound is determined by the vibration rate of the vocal folds. Amplitude: This is the intensity or power of the sound wave. We perceive amplitude as loudness. Greater amplitude means a louder sound.

• Duration: This is simply the length of time a sound lasts, measured in milliseconds (ms).

A spectrogram plots these properties visually: The x-axis represents Time. The y-axis represents Frequency.

• The darkness or color intensity of a point on the graph represents the Amplitude at that specific time and frequency.

With a spectrogram, we can see the invisible structure of speech. A plosive like [p] will appear as a blank space (the closure) followed by a burst of noise. A fricative like [s] will show up as a chaotic, high-frequency static.

The most important acoustic feature for distinguishing vowels is the presence of formants. When you produce a vowel, your vocal tract acts as a resonant chamber, like the body of a guitar or a violin. It amplifies certain frequencies and dampens others. These amplified frequency bands are the formants.

While there are many formants, the first two (F1 and F2), the lowest in frequency, are the most crucial for vowel identification. Their values are directly linked to the articulatory positions we discussed earlier.

First Formant (F1) corresponds to Vowel Height. There is an inverse relationship. High vowels (like [i] and [u]) have a low F1. This is because the large pharyngeal cavity created by the high tongue position resonates at a low frequency. Low vowels (like [æ] and [ɑ]) have a high F1. The smaller pharyngeal cavity resonates at a higher frequency.

Second Formant (F2) corresponds to Vowel Backness. Front vowels (like [i] and [ɛ]) have a high F2. The short oral cavity in front of the tongue constriction resonates at a high frequency. Back vowels (like [u] and [ɔ]) have a low F2. The longer cavity in front of the tongue constriction resonates at a lower frequency.

Lip rounding also has an effect, generally lowering all formant frequencies, especially F2.

This acoustic information is remarkably consistent. You can plot the F1 and F2 values of a speaker's vowels on a graph, and it will produce a chart that looks almost identical to the articulatory vowel quadrilateral. This provides a powerful physical confirmation of our articulatory descriptions. The shape of your mouth creates a specific acoustic signature.

The sound wave is funneled by the outer ear to the eardrum, which vibrates. These vibrations are transferred through the three tiny bones of the middle ear to the cochlea in the inner ear. The cochlea is a fluid-filled, spiral-shaped organ lined with thousands of tiny hair cells. Different parts of the cochlea's basilar membrane are sensitive to different frequencies. High frequencies are detected at the base, and low frequencies at the apex.

When a specific frequency enters the cochlea, the corresponding hair cells are stimulated and send a neural signal along the auditory nerve to the brain. The brain receives this incredibly complex stream of frequency information and begins its remarkable job of interpretation.

The brain doesn't just passively record sound like a microphone. It actively filters, categorizes, and even manufactures perception based on context, experience, and other sensory inputs.

Categorical Perception

One of the most profound discoveries in auditory phonetics is categorical perception. We do not perceive speech sounds along a smooth, continuous spectrum. Instead, our brains force them into discrete categories.

Consider the difference between [b] and [p]. One key acoustic difference is Voice Onset Time (VOT)—the time between the release of the stop closure and the beginning of vocal fold vibration. For [b], the vibration starts almost immediately (short VOT). For [p], there is a significant delay (long VOT).

If you were to create a series of sounds in a lab, starting with a VOT of 0 ms and gradually increasing it to 60 ms, you wouldn't hear a sound that slowly becomes more "p-like." Instead, you would hear [b], [b], [b], [b]... and then, suddenly, at a certain threshold (around 25-30 ms for English speakers), you would hear [p]. The perception "flips" from one category to the other. You are physically incapable of hearing the sounds in between as a blend; your brain forces a choice. This is categorical perception, and it's how we manage the infinite acoustic variation in the real world and sort it into a finite set of meaningful phonemes.

The McGurk Effect: Seeing is Hearing

Auditory perception is not isolated. The brain integrates information from other senses, especially vision. The McGurk Effect is a stunning demonstration of this.

In the classic experiment, a participant is shown a video of a person saying the syllable [ga]. However, the audio track is dubbed with the sound [ba]. The overwhelming majority of people do not hear [ba] or [ga]. They hear [da]. Their brain takes the visual information of a velar articulation ([g]) and the auditory information of a bilabial articulation ([b]) and computes a third, intermediate perception: the alveolar [d]. This proves that what we "hear" is a construction, a best-guess synthesis of all available sensory data.

Phonemic Restoration Effect

The brain also uses context to fill in missing information. If you take a recording of the sentence "The state governors met with their respective legislatures convening in the capital city" and replace the first /s/ in "legislatures" with a cough, listeners will not notice the missing sound. They will report hearing the full word and a cough. Their brain, using the surrounding linguistic context, "restores" the missing phoneme. It predicts what should be there and inserts it into our perception.