What Is Acoustic Phonetics?

Acoustic phonetics is a branch of linguistics concerned with the physical properties of speech sounds: the frequency, amplitude, duration, and spectral structure of the sound waves produced by the human vocal tract during spoken language. It sits at the intersection of linguistics, physics, and engineering, providing the empirical measurements that connect articulatory gestures made by a speaker to the perceptual experience of a listener. Where articulatory phonetics describes how sounds are produced by the vocal organs, and auditory phonetics concerns the perception of speech, acoustic phonetics occupies the middle domain of the signal itself.

The field draws on classical acoustics, signal processing, and digital measurement technology. Its analytical tools, including the sound spectrograph and digital FFT analysis, allow researchers to visualize and quantify speech sounds that are otherwise too rapid and complex to analyze by ear alone. Acoustic phonetics is foundational to automatic speech recognition, text-to-speech synthesis, forensic voice analysis, and the study of language variation and change.

Speech Sound Production

The human vocal tract functions as an acoustic resonator whose shape is continuously modified by movements of the tongue, lips, jaw, and velum. Voiced sounds, including vowels and voiced consonants, are produced when the vocal folds vibrate to create a periodic source signal whose fundamental frequency (F0) corresponds to perceived pitch, typically ranging from around 80 Hz for adult male speakers to over 300 Hz for children. This source signal is filtered by the resonances of the vocal tract above the glottis, which selectively amplify certain frequency bands. Voiceless sounds such as fricatives (/s/, /f/) use turbulent noise generated at a constriction in the vocal tract as their source rather than periodic vocal fold vibration, producing aperiodic spectra. The University of Sheffield's acoustic phonetics research group notes that waveform and spectral analysis together reveal the periodic and aperiodic patterns that distinguish voiced from voiceless sounds and characterize each phoneme class.

Formants and Spectral Structure

The resonant frequencies of the vocal tract, called formants, are the primary acoustic correlates of vowel identity. Formants appear as broad peaks in the speech spectrum and are labeled F1, F2, F3, and so forth in order of increasing frequency. The first formant (F1) is inversely related to vowel height: high vowels like /i/ have low F1 values (around 270 Hz for adult males) while low vowels like /a/ have high F1 values (around 730 Hz). The second formant (F2) tracks vowel backness, with front vowels producing higher F2 values than back vowels. Consonants are distinguished by transitions in formant frequencies immediately before and after the consonant, by the presence or absence of voice bar energy at the lower frequencies, and by burst spectra in the case of plosives. The Macquarie University speech acoustics group provides detailed documentation of the acoustic representations that define each class of English speech sound.

Spectrographic Analysis

The sound spectrogram is the central visualization tool of acoustic phonetics. It plots time on the horizontal axis, frequency on the vertical axis, and acoustic energy as darkness or color, making it possible to read off formant trajectories, duration, voicing onset time, and spectral transitions in a single display. Wideband spectrograms (analysis window around 5 ms) resolve individual pitch periods as vertical striations, enabling measurement of F0 and voice quality. Narrowband spectrograms (analysis window around 45 ms) resolve the individual harmonic components of the voice source, useful for studying phonation and prosody. Modern digital analysis tools allow the extraction of formant frequencies, fundamental frequency, intensity, and other parameters frame by frame at millisecond resolution. The Journal of the Acoustical Society of America has published foundational and ongoing research on acoustic measurement of speech since 1929.

Applications

Acoustic phonetics has applications in a range of disciplines, including:

• Automatic speech recognition and natural language processing systems
• Text-to-speech synthesis and voice conversion
• Forensic speaker identification and voice comparison
• Clinical assessment of speech and voice disorders
• Second-language teaching and pronunciation training