Exercise 7.2

You will probably have noticed that the spectrogram portions corresponding to the nasals seem very pale, particularly in the top two thirds of the spectrogram. Nasal sounds are produced with the velum (the flap at the back of the palate) lowered so that air can flow through the nose. Nasal consonants also have a closure in the mouth (e.g. the lips for /m/). The air flow through the nose causes antiresonances, i.e. dampens the amplification of sound in the oral cavity that would otherwise happen.

The vowels in to and the are both unstressed, reduced vowels. While these words in isolation would have different vowels (and the spelling partly shows this), when they are spoken in sentences and with little emphasis they both have the weak schwa-quality.

The /i/ vowels differ in their duration and also in some of the movements of the formants (the horizontal or near-horizontal darker patches in the spectrograms). These differences can be linked to where the vowels are in the speech. Vowels in words that are emphasised or that are just before the end of a clause or sentence tend to be longer - notice that team has a long /i/. Also, in English, vowels before voiced consonants like /d/ tend to be longer than those before voiceless consonants like /t/ - see how the vowel in lead is a little longer than the vowel in Pete.