Skip to Content

Department of Linguistics

SPEECH ACOUSTICS

Frequency Transformations of Speech
Introduction

Robert Mannell

The auditory system is a non-linear tranducer of sound. The inner ear converts sound waves into patterns of action potentials in the auditory nerve which are then transmitted to the auditory cortex via the auditory processing centres in the brainstem. This complex neural pattern includes spatial and spectral representations of the sounds entering the ears. The spatial representations help us to identify the position of the sound source in space (ie. directional hearing) and these spatial representations are also important in permitting us to attend to a single voice in the presence of background noise or other competing voices. The spectral representations are a transformed version of the spectra of the sounds entering the ear.

Psychoacoustics (a division of psychophysics) is the study of the ways in which the auditory system proceses sound, but from a perceptual perspective. Psychoacoustics is interested in such issues as perceivability, discrimination between sounds and the auditory resolution of spectral components in complex sounds. For detailed information on psychoacoustics see the psychoacoustics pages.

In this lecture we will examine some aspects of auditory spectral representations of speech. We will look mainly at frequency representations of sound in the auditory system and how they affect the salience (perceivability) of acoustic features that are important in the preception of speech. We will also look briefly at intensity transformations of speech sounds. The auditory system also distorts the temporal and phase characteristics of sound but these characteristics are less important for our understanding of how the auditory transduction of sound affects speech perception.

It should also be noted that auditory frequency scales are sometimes used in algorithms used by speech recognition systems.