Spectra of Speech Sounds
Nigel Hewlett, Janet Mackenzie Beck · 2013
An explanation of the acoustic basis for these finer differences in quality cannot be achieved merely by a close examination of waveform shapes. It is necessary to delve more deeply into the properties of the sound. Speech sounds are made up of components at different frequencies. That is why we were careful to call the frequency of repetition of the whole cycle of a periodic wave the fundamental frequency. There are other frequencies above it. The quality of a sound is determined by the relative amplitudes among these higher frequency components. The analysis of the frequency components that make up a sound is called frequency analysis or frequency domain analysis in acoustic phonetics, as distinct from the time domain analysis that previous chapters have been concerned with. The results of a frequency analysis are typically displayed on a graph called a spectrum, thus the term spectral analy- sis. Spectral analysis has been most extensively and successfully applied in the phonetic description of vowels (of monophthongal vowels, in fact), to which the bulk of this chapter is devoted. Spectra of consonants are dealt with more briefly at the end.