The analysis discussed here was motivated by a comment in
Message-ID: <bu6tlq$ech9b$1@ID-162222.news.uni-berlin.de>
to the effect
that my vowel in "win" seemed to have a glide.
I've recorded the words "hib hid hif hig hij hik hil him hin hip hir his hit hiv hiz"
in order to show how glide varies with the sound following the vowel. I think this is
what Professor Ladefoged calls "anticipatory coarticulation". You can
click WAV (420 kb) or
MP3 (29 kB)
to hear the sound file
I've used the Praat* phonetic-analysis software, version 4.1.22, to make a composite
plot. The entire plot is obtained with Praat simply by selecting the "Edit" option
when a sound recording has been selected.
Praat provides for copying some plots to the Windows clipboard, but it doesn't have
that provision for the "Edit" plot, so I've used screen capture in another program to
retrieve the composite plot and put it into a JPEG file, which you can see by clicking
here.
The plot shows the following traces:
The formants are referred to by the terms "F1" thru "F5", ordered from bottom to top
in the composite plot. Of primary interest are F1 and F2. Vowel glide is
mostly evident in the slope of F2. The phonetics terms "forward" and "back" correspond
to separation of the first and second formants: The greater the separation, the farther
forward the vowel is. The phonetics terms "low" and "high" correspond to frequency
of the first formant: The higher the frequency of the first formant, the lower the
vowel is.
Most of the vowels show in varying degrees a down slope of the second formant, while
the first formant has in general much less slope. This means that the vowel glides
back, while the lowness doesn't change much. Greatest glide is in "him", least in
"hig", while "hin" has an intermediate amount of glide.
Of some interest is the closeness of the second and third formants and the downward slope
of both of them in "hir". This seems to be typical of an "r" sound following a vowel.
Also note the scattered pattern of dots at the end of "hif", "hij", "hik", "his", "hit", and "hiz".
Sibilants and plosives produce something like white noise, which properly shows up in a formant
plot as it has done in these cases.
* The Praat phonetic-analysis software is copyright 1992-2004 by Paul Boersma and David Weenik.
You can learn more about it here .
