Acoustic Phonetics

by

Allard Jongman

• LAST REVIEWED: 23 June 2023
• LAST MODIFIED: 23 June 2023
• DOI: 10.1093/obo/9780199772810-0047

Introduction

Acoustic phonetics is the study of the acoustic characteristics of speech, including an analysis and description of speech in terms of its physical properties, such as frequency, intensity, and duration. Descriptions of speech sounds in these terms date back as far as 1830 (Willis), but the invention of the sound spectrograph (1945) was the major technological breakthrough that made the analysis and visualization of the speech signal possible. Subsequent developments in digital signal processing, most notably the discrete Fourier transform, have made it possible to conduct all acoustic analyses with a basic microcomputer. Rousselot (b. 1846–d. 1924) is widely regarded as the “father of experimental phonetics.” Rousselot applied the kymograph to the study of speech. The kymograph, invented in the 1840s by Ludwig, was originally used for measuring blood pressure and other physiological processes. For speech, the kymograph consisted of a rotating drum covered with paper coated with soot; speakers spoke into a rubber tube and the sound vibrations were captured by a stylus that registered the variations in air pressure, from which duration, intensity, and pitch could be measured. Edison’s invention of the phonograph in 1877 was of crucial importance since it was the first device that allowed the recording and reproduction of sound. This invention meant that speech was no longer a fleeting event but could be repeatedly heard and analyzed. A number of researchers developed additional devices to visualize and analyze the sound waveforms, including Hermann, Scripture, and Verner. Although the speech waveform (oscillogram) is the basis of all acoustic speech research, it is rarely used as a source. One important reason is that it is “too rich” since it contains information about frequency, intensity, and phase of the signal components, while human perception disregards the latter. Hence, the spectrogram provides a better representation.

Foundational Works

Acoustic phonetics is an instrumental science that depends on ways to store, replicate, visualize, and analyze the speech signal. Acoustic phonetics is also a cumulative science in which older research continues to be influential. While the early work of researchers, such as Rousselot 1897–1908 and Scripture 1906 (see also the section Sound Spectrograph), provided a first glimpse at the complexity of the speech signal, this research is not widely cited nowadays. In contrast, a number of seminal contributions on vocal tract acoustics and research based on the sound spectrograph remain influential. Among these, Chiba and Kajiyama 1958 introduces the acoustic theory of vowel production, which was further developed in Fant 1960, a seminal dissertation on the acoustic theory of speech production. Potter, et al. 1947 provides the first systematic overview of the acoustic characteristics of English vowels and consonants as illustrated by spectrograms. Joos 1948 still is a very accessible general introduction to acoustic phonetics. The authors of Jakobson, et al. 1952 use new insights from speech processing to define distinctive features in acoustic rather than articulatory terms. Finally, Stevens 1998 is currently the most comprehensive source for the acoustics of vowels and consonants. The work is based on nearly fifty years of research by one of the most important figures in speech science.

• Chiba, T., and M. Kajiyama. 1958. The vowel: Its nature and structure. Tokyo: Kaiseikan.

  English translation of the original Japanese work published in 1941. The authors introduced the electric-circuit analog to simulate a resonance of the vocal tract and were aware of the existence of multiple formants. Their calculations of the first two formant frequencies of vowels based on vocal tract shapes were remarkably accurate. This research laid the foundation for perturbation theory and the acoustic theory of speech production.

• Fant, G. 1960. Acoustic theory of speech production. The Hague: Mouton.

  A seminal contribution by one of the pioneers in speech acoustics, this book provides a detailed account of the relation between vocal tract shape and formant pattern for vowels and consonants. The acoustic theory of speech production remains the dominant framework for research in acoustic phonetics.

• Jakobson, R., G. Fant, and M. Halle. 1952. Preliminaries to speech analysis: The distinctive features and their correlates. Technical Report 13. Cambridge, MA: Acoustics Laboratory, Massachusetts Institute of Technology.

  Before the introduction of the spectrograph, most phonetic research focused on articulation. In this monograph, the authors used the latest advances in speech acoustics to forge a close link between acoustic phonetics and phonology, specifically by defining binary distinctive features in acoustic terms. Later editions were published by MIT Press.

• Joos, M. 1948. Acoustic phonetics. Language Monograph 23. Baltimore: Linguistic Society of America.

  The author was the first linguist with extensive access to the sound spectrograph during World War II, well before the first public mention of this new equipment in 1945. This monograph covers basic aspects of wave analysis and source-filter theory and introduces the sound spectrogram as a new way to visualize and study vowels and consonants.

• Potter, R. K., G. A. Kopp, and H. C. Green. 1947. Visible speech. New York: D. Van Nostrand.

  The first comprehensive overview of the acoustic properties of (English) vowels and consonants, richly illustrated with spectrograms.

• Rousselot, P. J. 1897–1908. Principes de phonétique expérimentale. Vol. 1. Paris: H. Welter.

  Beautifully illustrated, this volume provides a detailed description of the equipment and techniques Rousselot developed for the display and analysis of speech. Also includes good coverage of the work of his contemporaries.

• Scripture, E. W. 1906. Researches in experimental phonetics: The study of speech curves. Washington, DC: Carnegie Institution of Washington.

  DOI: 10.1037/13670-000

  “Speech curves” were essentially automatically traced from phonograph records and subjected to various quantitative analysis methods. These data were used to evaluate competing theories of vowel production as proposed by Helmholtz, Hermann, and others.

• Stevens, K. N. 1998. Acoustic phonetics. Cambridge, MA: MIT Press.

  The definitive work on speech acoustics based on fifty years of research by one of its pioneers. Based on the acoustic theory of speech production, this book provides an acoustic analysis of vowels and consonants in unparalleled detail.