Elements of Acoustic Phonetics

Citation

Ladefoged, P. (1995). Elements of Acoustic Phonetics (2nd ed.). University of Chicago Press.

Chapter Summary

This is a short book, about 220 pages with 122 line drawings, that takes a reader with no background in physics or mathematics from the nature of sound to the digital techniques used in speech laboratories. The second edition has eleven chapters where the first had seven. Four chapters are new, covering experimental techniques made possible by computers, and the topics the publisher lists as new (resonances of the vocal tract, digital speech processing, Fourier analysis, Linear Predictive Coding) correspond to Chapters 8–11. Chapters 1–7 move from the basic properties of sound waves through to hearing and speech production.

Chapters 1–2: Sound Waves; Loudness and Pitch

The opening chapters describe what a sound wave is and how variations in its physical properties relate to what a listener hears as loudness and as pitch. They are the base for everything that follows.

Chapters 3–4: Quality; Wave Analysis

Chapter 3 addresses the quality of speech sounds, alongside loudness and pitch, and Chapter 4 the analysis of waves. Karen Steffen Chung, reviewing the book in Language, describes it as taking the reader from the most basic descriptions of what sound is to complex waves, Fourier analysis and the decibel scale.

Chapters 5–7: Resonance; Hearing; The Production of Speech

Chapter 5 introduces resonance; Chung describes the complexity of the vocal tract's many resonators being built up through simplified graphic analogies. The chapters on hearing and on speech production then connect the physics to the listener's ear and the speaker's vocal organs.

Chapter 8: Resonances of the Vocal Tract

The chapter on the vocal tract applies the earlier ideas about resonance to speech. The publisher describes this material as explaining how formants are related to different cavities of the tract.

Chapters 9–11: Digital Speech Processing; Fourier Analysis; Digital Filters and LPC Analysis

The last three chapters cover how sound waves are stored and processed by computer, the mathematics of Fourier analysis, and digital filtering with Linear Predictive Coding, which the publisher identifies as the equations used most often in analysing speech sounds. Chung singles out the treatment of digital speech processing for clarifying ideas with wider application, such as the range of 8-bit against 16-bit quantization, the sampling frequency needed for different signals, and the use of filters.

Key Concepts

Chapters 1–4: Sound and its properties

• Sound waves: the physical signal whose properties are described in the first chapter.
• Loudness and pitch: the auditory impressions related to a wave's amplitude and frequency.
• Quality and wave analysis: the quality of speech sounds and the analysis of waves, the topics of Chapters 3 and 4.

Chapters 5–8: Resonance, hearing and the vocal tract

• Resonance: the topic of Chapter 5, extended to the vocal tract in Chapter 8.
• Formants and cavities: the relation between the resonances of the vocal tract and its different cavities.

Chapters 9–11: Computer analysis

• Digital speech processing: storing sound waves in a computer, including quantization and sampling.
• Fourier analysis: breaking a wave into its frequency components.
• Linear Predictive Coding: a technique used with digital filters to analyse speech sounds.

Critical Analysis

Strengths

• Written for readers without a science background. The publisher says the book assumes no background in physics or mathematics, and Chung observes that Ladefoged guides the reader by the hand through anything mathematical, with everything carefully explained for the numerically challenged.
• Builds understanding step by step. The review describes a picture of the many resonators of the vocal cavity constructed slowly through simplified graphic analogies, and notes that points such as why 70.7% is a special number, what the radiation factor is, and how damped and undamped resonators behave become clear.
• Extended to current laboratory practice. The second edition adds four chapters on computer-based methods; Chung calls the coverage of digital speech processing excellent and expects readers who go on to further reading on computer speech processing to find the information solid.
• Compact. The book is about 220 pages, with 13 tables and 122 line drawings.

Limitations

• Not for browsing. Chung calls it above all a content book, suited to careful reading and rereading rather than skimming, unless the reader already knows the concepts.
• Heavier calculations at the end. In the last chapter, she says, the calculations become somewhat lengthy and overwhelming, although they are in principle not hard to follow.
• Dated in its technology. The digital chapters reflect computer speech processing as of 1995, so the principles remain useful but current software and practice are not covered.
• Acoustics only. Apart from one chapter on the production of speech, the contents are acoustic, so a general phonetics course or text is still needed for articulation and transcription.

Where It Fits in the Course

Ladefoged's book is the compact route into the physics of speech in the phonetics and phonology course. Ladefoged and Johnson's A Course in Phonetics is the hands-on introduction to the production and acoustic properties of speech sounds, with transcription practice, and is the natural first book; Elements follows it for readers who want a more thorough grounding in acoustics. Johnson's Acoustic and Auditory Phonetics covers the physical signal of speech and the perceptual systems that interpret it, and overlaps with Ladefoged's treatment of the speech signal and of hearing, so the two can be compared or read in sequence.

Clark and colleagues' An Introduction to Phonetics and Phonology combines articulatory and acoustic phonetics with phonological theory and gives wider context, while Odden's Introducing Phonology and Carley, Mees and Collins's Practical English Phonetics and Phonology concentrate on sound patterns and on English, and do not replace Ladefoged's treatment of waves, spectra and digital analysis.

How to Use This Book

The book suits students who have taken an introductory phonetics course and now want to understand what the spectrograms and analysis tools in the laboratory are doing, and speech science students who need a primer in acoustics. No physics or calculus is required.

A practical plan: read Chapters 1–4 in order and slowly, working through each explanation before continuing. Take Chapters 5–8 as the core for linguists, since they connect resonance to formants and to the cavities of the vocal tract. Chapters 9–11 are the technical part: read Chapter 9 for the practical consequences of digitization (sampling rate, quantization) even if the later mathematics is skipped, and return to Chapters 10–11 when Fourier analysis or LPC becomes necessary for a project. Pair the reading with recordings and software for sound analysis, and reread the early chapters when the later ones refer back to them.