Acoustic Phonetics: Vowels and Consonants

Acoustic Phonetics

• Vowels are always voiced, meaning the vocal folds are vibrating, and the vocal tract is open and unrestricted.
• Consonants may be voiced or unvoiced, and the vocal tract is obstructed.
• Acoustic differences between phonemes are related to their production.

Phonetic Features

• Properties shared by a group of phonemes can be acoustic, articulatory, or auditory.
• Examples of phonetic features include long/short, nasality, and sibilance.
• Classification of phonetic features can be clear, but not always.

Speech Waveforms and Spectra

• A waveform is a graph of pressure or amplitude against time.
• Speech waveforms have a complex time and frequency structure.
• They are made up of many frequency components, ranging from 50Hz to 10kHz, with 100Hz to 3kHz being the most important.
• Periodic and aperiodic components are present in speech waveforms.
• Amplitude envelope gives important cues for suprasegmental speech perception.

Vowel Quadrilateral

• Vowels are articulated with characteristic tongue and lip shapes.
• The vocal tract is relatively open and acts as a resonator or filter of vocal fold vibration.
• The source of vowels is a periodic vocal fold waveform.
• Lips, tongue, jaw, and velum change the shape of the vocal tract.
• Position of the tongue is the most important factor in determining vowel sounds.

Source-Filter Model of Speech Production

• The production of sound has three stages: initiation, resonance, and radiation.
• Two mechanisms of initiation in speech are vocal fold vibration and turbulence of air through a constriction.
• The source-filter model of speech production shows how air power is converted into a periodic acoustic signal or an aperiodic signal at a constriction and then modified by vocal tract filtering or resonance.

Acoustics of Vocal Fold Vibration

• The frequency of vocal fold vibration is affected by the length of the vocal folds, tension in the vocal folds, and pressure in the trachea.
• The waveform of vocal fold vibration is triangular, not sinusoidal.
• Different people have different waveforms for vocal fold vibration.

Vocal Tract Resonance

• Vocal tract resonance affects the spectra of voiced speech sounds.
• The vocal tract acts as a resonator or filter of vocal fold vibration.
• The resonance of the vocal tract is affected by the shape of the tract, which is changed by the position of the tongue, lips, jaw, and velum.

Acoustic Properties of Vowels

• Vowels have characteristic acoustic properties, such as formants.
• Formants are broad resonances in the spectral envelope.
• The first formant (F1) is inversely related to vowel height, and the second formant (F2) is related to degree of backness.
• Vowel formants are affected by the shape of the vocal tract, which is changed by the position of the tongue, lips, jaw, and velum.

Diphthongs

• Diphthongs involve a smooth movement of the tongue during a vowel.
• Corresponding formant movements occur during a diphthong.
• Diphthongs are long vowels, typically lasting 250-300ms.

Spectrograms

• Spectrograms show the time and frequency variations of a sound.
• The y-axis shows frequency, and the x-axis shows time variation.
• Darkness (or lightness) shows amplitude.
• Spectrograms are produced using a set of band-pass filters.
• Trade-offs occur between time and frequency resolution.

Examples of Spectrograms

• Examples of spectrograms are provided for various words, such as "audiology", "heed", "who'd", "horde", "hid", "head", "had", "hood", "hod", "hud", "boy", "buy", "bay", "bow", "bough", and "tour", "tare", "tear".