Chapter 3. Articulatory system
– Spectrographic analysis of
vowels and consonants (stops, fricatives, affricates) – Lecture 8 –
02/10/14
· Spectrographic analysis
o
Spectrography is a method of identifying frequency, amplitude,
and duration of sounds.
o
Frequency is displayed on the vertical axis, time is represented
on the horizontal axis, and intensity of acoustic energy is represented by the
darkness of the trace on the screen.
· Spectrographic analysis of
vowels
o
Vowels are characterized by first three formants.
o
They appear as wide, dark horizontal stripes, reflecting the
concentrations of intense acoustic energy at those harmonic frequencies that
have been amplified by the vocal tract formants.
o
The first format (F1) in vowels is inversely related to vowel
height
§ The higher the vowel, the
lower the first format (and vice versa)
· High
/i/ and /u/ à F1 ~ 280-310
· Mid-high
/I/ and /ʊ/ à F1 ~ 400-450
· Mid-low
/ɛ/ and /ɔ/ à F1 ~ 550-590
· Low /æ/ and /a/ à F1 ~ 690-710
o
The
second format (F2) in vowels is somewhat related to degree of backness
§ The
more front the vowel, the higher the second format (but affected by lip
rounding).
|
Front
|
Back
|
Frequency Range
|
|
/i/
|
/u/
|
F2 ~ 2250 ~870
|
|
/I/
|
/ʊ/
|
F2 ~ 1920 ~ 1030
|
|
/ɛ/
|
/ɔ/
|
F2 ~ 1770~880
|
|
/æ/
|
/a/
|
F2 ~ 1660~1100
|
·
· Spectrographic analysis of
Diphthongs
o
A diphthong is a vowel that changes its resonance
characteristics during its production.
o
Characterized by first three formant frequencies
o
Produced by uttering two vowels as one unit.
§ This results in formant
transitions.
§ Steady-state formants at
the beginning of the sound, followed by formant transition, and then another
steady-state portion.
· Spectrographic analysis of
Stops
o
Four important acoustic features
§ Silent gap
· Time
during which the articulators are forming the blockage and oral pressure is
building up. Seen as voice bar for voiced stops.
§ Release burst
· Follows
the silent gap. Seen as vertical line extending into high frequencies. Lasts
for 10 – 30 ms.
· Seen
for stops in initial and medial position, but not for final position.
· Bilabial
stops – diffuse spectrum, with energy spread out over a wide range of
frequencies and more energy in the lower frequencies than in the higher.
· Alveolars
– diffuse with increasing energy in the higher frequencies or spread out
evenly.
· Velars
– Compact with energy concentrated in a relatively narrow region.
· Voiceless
stops have bursts that are longer in duration due to aspiration.
§ Format Transitions
· Articulators
move from constricted position to open position following the production of the
sound.
· Can
occur either from a voiced sound occurring before the stop, or both, lasts
around 50 ms.
· The
slope of transitions – depends on the place of articulation of the stop and
vocal tract position following the sound
· A
very low F1 – complete constriction of the vocal tract.
· The
transition for F1 starts from nearly zero and increases to the
appropriate frequency for the following vowel
§ Voice onset time (VOT)
· VOT
– time between the beginning of the stop burst and the onset of the vocal fold
vibration for the following vowel.
· This
is measured in milliseconds and indicates the coordination between laryngeal
and articulatory systems.
· Four
categories of VOT – depending on the timing between the release of the burst
and the onset of the vocal fold vibration:
o
VOT lead – VOT is negative indicating
that the vocal folds are vibrating before the articulatory release (occurs for
voiced stops).
o
Simultaneous voicing – voice onset and
articulatory release occur at the same time.
o
Short lag – onset of vocal vibration
follows shortly after the release burst.
o
Long lag time – vocal fold vibration
time is delayed relatively long time after the articulatory release
· VOTs
for Voiced stops – - 20 ms to + 20 ms
· VOTs
for voiceless stops – +25 ms to 100
ms
· VOTs
depend on the place of the stop articulation and increases as place of
articulations moves backward.
· Bilabials
- shortest VOTs.
· Alveolars
– intermediate VOTs
· Velars
– longest VOTs
· Fricatives
o
On a spectrogram – wide band of energy distributed over a broad
range of frequencies.
o
Energy in fricatives is much longer in duration than stops.
o
Specific range of frequencies and the intensity of aperiodic
sounds depend on the place of articulation of the fricative.
o
Fricatives are characterized by white noise, which is aperiodic
sound that has its energy distributed fairly evenly throughout the spectrum.
o
The spectrum for fricatives depends on place of articulation
because fricative noise is resonated most strongly anterior to the articulatory
constriction.
o
Fricatives that are produced anteriorly (e.g., /f/, /v/, /ϴ/) do
not have much of a front resonating cavity, so they have very low-intensity
spectrum spread out over a broad range of frequencies.
o
The alveolar and palatal fricatives have more intense acoustic
energy at high frequencies because of the way that they are resonated.
§ Stridents (/s/, /z/, /ʃ/, /ᴣ/)
have much more intense energy than the nonstridents (/f/, /v/, /ϴ/, /ᶞ/)
· Affricates
o
They are produced quickly combining a stop with a fricative, so
the acoustic characteristics of affricates have elements of both stops and
fricatives.
o
Affricates have a silent gap associated with the stop part of
the sound.
o
Frication noise related to the fricative portion of the
affricate sound follows the silent gap.
o
Fricatives and affricates look very similar on a spectrogram,
except that affricates are shorter in duration.
· Relevant questions for
exam(s)
o
Identifying the spectrograms of vowels (your clue would be
mainly based on the first two formant frequencies)
o
Identifying the spectrograms of vowels (your clue would be the
transition of formant frequencies)
o
Identification of spectrograms of stop consonants (clue would be
voice bar (if any), release burst, VOT, and formant transitions)
o
Identification of fricatives (based on the spread of intense
acoustic energy)
o
Learn the VOT for voiced and voiceless stops
