All images in this file are courtesy of MIT Press. Used with permission.
6.54IJ Handout
tit4 Hdoota/
S Lkfcie-
2/26/04
-~~~~~_
4
L
5
L
-yi
11.
3
L
Figure 2.2 Two-mass model of the vocal folds, showing compliances coupling the two masses
to the lateral walls and to each other.
2.0 - .,
u1
0-o
.
.
I I
.
.
.l
.
.l
.
I I I I I
.
.
.
I I I I I
.E
o3
e
0
-
.
.-
D
I
. . . .
a.
·
.
.
·
·
I I
·-
.
.
....
\.
.0
O
0.5
FRACTION OF CYCLE
Figure 2.1 The upper part of the figure shows estimated lateral sections through the vibrating
vocal folds at eight equally spaced times through the glottal vibration cycle. Data were obtained
by Baer (1975) from an excised dog larynx. The scale of 1 mm is indicated by horizontal and
vertical lines in the upper right of each panel. The curves in the lower panel show the lateral
width of the glottal opening measured at the upper edge ot the vocal folds and at a point 13
mm below the upper edge. These curves illustrate the time lag between the displacements of the
lower and upper portions of the vocal folds. (From Baer, 1975.)
C,,
T
O3
TIME (ms)
Figure
-
A
.4
r Ao
-nf
jmrrr rnh'n rrF IIIra
S$;r
cnemanc reprULnrauNn
o
I 1.u
l*e
I Illu
vI uS
PL_
A
urr
-ingmasses
u--.-.
...
u
a cde
- -J --
of vibration of the two-mass model in figure 2.2. The displacement of the lower mass is given by
ABCFH, whereas the trajectory of the upper mass is EFI. The upper mass begins its lateral
motion at E, when the displacement of the upper mass is x1o. The dashed line to point D is the
projected trajectory of the lower mass if the glottis did not open and if steady pressure on
the lower mass were maintained. The glottis is open and airflow occurs in the region marked by
the line EFH. The beginning of the second cycle is shown at the right. Assumed physical parameters of vocal folds are appropriate for an adult male, as described in the text.
Figure 2.3
Tracings of vocal fold shapes at four instants of time during the dosed phase, from
figure 2.1. (From Baer, 1975.)
--
-·-----
-----
�--
(-2-
M�he\
(a) a
E
I­
0
Figure 2.28 Waveform of derivative of glottal airflow Ug based on the LF model (Fant et al,
1985). The period is To and the open quotient OQ = 50 percent. The solid line is for an abrupt
discontinuity in the derivative of Us (i.e, parameter T2 = 0) and the dashed line corresponds to
T2 = 0.025T7. See text.
Lw.
(b)
=
T2
60
E
ta,
OQ =
o ms
0.5
I
a)
0.
Ca
E
AA
4)
4
3
12
0
OIz)
FREQ
I
(C)
60
.40
i
40
0.1 ms
T2
|
|
|
OQ =
I
a
I
0
A
X
. a,
\,
VV V
0
I
3
'
2
FREQUENCY (kHz)
0.1 ms
re
'4
I.
, \
=
OQ = 0.7
0.3
FREO(-Iz)
A
T2
60
]
I,-
,,
-­
1
3
2
4
5
FREQ Of)
al
Vy
Iy·I y
Figure 2.29 Spectrum of derivative of glottal airflow for LF model for several combinations of
values of return time T2 and open quotient OQ as indicated. Fundamental frequency= 125 Hz.
These spectra were calculated from a version of the LF model developed by Dennis Klatt (Klatt
and Klatt, 1990).
~­
U
-.
_-
I
Figure 2.12 (a) The solid line is the volume velocity-waveform corresponding to the model
with increased stiffness, for which displacements of masses are as shown in figure 2.11, and
supra- and subglottal configurations are as in figure 2.8. The dashed line corresponds to the
waveform in the absence of the effect of supra- and subglottal acoustic masses. (b) Derivative of
waveform given by solid line in (a). (c) Spectrum of dUs/dt in (b). The dashed line has a -6 dB
per octave slope.
,~~.~i-----··~~~-------- ---------------
~~~~~~1
Vv \
t~·C5
1
I
IhAe%
/w/'
ats-3
Figure 9.19 Midsagittal sections when the vocal tract reaches its most constricted config­
uration for the glides /w/ (left) and /j/ (right). In the case of /w/, the frontal lip contour is also
shown, to illustrate the lip rounding. (Adapted from Bothorel et al, 1986.)
/iv
'/I
I
0
o
I
EE
:
I
0
0
a
5E
Figure 9.24 Spectrograms of the utterances /'jt/
and /a'wlt/ produced by a female speaker
(top) and a male speaker (bottom). Measured F trajectories are superimposed on the spectro­
grams.
�
��_1_�__1_______3__1__I_
�11_
wY
4eWiT-t/
l
-
I_
-
�
-.
--
­
--
60
­
60
40
40
i
-H
3
a
q.
2O
I'
=
0
O1
I
A
AL I
2
2
6
33
t--
L
0
4
2
1
FREQ (kH L)
k
-A
-
j-
3
FREQ (dz)
~ IV
V
4
5
CiN
. -
70ms
som
200ms
210ms
V V
-V
i
E
C
I
II
26
_
_I
40
m
A
I
1
2
34
2
3
FRE0 (kHz)
C
3
0
1
2
3
FREQ (kHz).
4
5
0
1
4
5
320ms
Figure 9.27 Spectra sampled at several points through the utterance /'wt/, produced by a
female speaker (left) and a male speaker (right). In each case, the spectrum in the middle is sampled
near the point of maximum constriction, and the spectra on the top and bottom are sampled in
the preceding unstressed vowel and the following stressed vowel, respectively. Waveforms and
time windows (30 ms) are shown below the spectra.
��7----��--���-�----�--~~C""-P--�
I
.1s
a-t
/d
le"4
v -A
I4/
40
3
20
i1
2
3
4
5
Z4
i
7Oms
BWnis
8Oms
I
:3
I
C
1
0
2
4
3
FREO (kz)
5
170ms
O
1
2
3
FRE Oz)
4
140m5
0
0
I
2
IAtU
A,
V
V
-
A V
,
VV
vv
U
I
~V
3
jdi
4
A
An.
v-vv
250nhs
5
,
An
A
vv yvvi
fv
210ms
Figure 9.33 Same as figure 927, except that the utterance is /'jat/. The female speaker is rep­
resented in the left panels and the male speaker in the right panels.
�--·-3�
5
NI I4CL bA
·L·�_I_,
con s
0oKntS
Figure 9.1 Midsagittal sections through the vocal tract during the production of nasal con­
sonants produced by dosing the lips (left) and raising the tongue tip to the alveolar ridge (right).
For the labial consonant the context is "Mets tes beaux habits," and for the alveolar consonant
the context is "Une rponse ... " (Adapted from Bothorel et al, 1986.)
SUPRAGLOTTAL
CONSTRICTION
10
z
0 -
C­
U)
-I-
cnv
O0
VELOPHARYNGEAL
OPENING
o
-200
TIME
-100
0
100
FROM RELEASE
(ms)
Figure 9.2 Schematization of the time course of the cross-sectional area of the oral constric­
tion (upper panel) and of the velopharyngeal opening (lower panel) during production of a nasal
consonant in intervocalic position. The detailed shapes of these trajectories depend on the place
of articulation for the consonant and for the adjacent sounds.
-----
�
T.�-�*-----.TI---------------
I
�
2­
I
Cr
TIME FROM
RELEASE (ms)
Figure 9.5 Schematization of trajectories of poles and zeros of the transfer function of the
vocal tract for a labial nasal consonant in intervocalic position. The transfer function is the ratio
of the total volume velocity from the nose and mouth to the volume velocity at the glottis. The
labial closure occurs between 0 and -I00 ms on the abscissa The lowest zero is labeled as Z1,
and is shown as adashed line. Zeros above ZI are not shown. Poles are indicated by solid lines,
including the two low-frequency poles RI and R2 that replace the first formant during the inter­
val when the velopharyngeal port is open. For times remote from the closure, poles and zeros
due to acoustic coupling to the nasal cavity are canceled, and only the natural frequencies of the
oral cavity remain labeled Fl through F4. The vowel configuration is taken to have a uniform
cross-sectional area The points labeled 1,2, and 3 indicate times at which the spectrum enve­
lopes in figure 9.6 are calculated.
5
___ F5
I
I
I
I
- 4
F,
0
W
0
.r 2
LA­
I
I
-100
O
I
=I
0
-200
I00
TIME FROM RELEASE (ms)
Figure 9.9 Estimated trajectories of poles and zeros of the transfer function of the vocal tract
for an alveolar nasal consonant in intervocalic position. Poles and zeros are labeled as in figure
9.5. The zero Zf accounts for acoustic coupling to the front cavity during the nasal murmur, as
discussed in the text. The points labeled 1, 2, and 3 indicate times at which the spectrum enve­
lopes in figure 9.10 are calculated
"II----------------�
--· · P)-
� -
�..�I
k) 3
I
2
a
. I'
-
_
. .
I
1
0
Cc:
J 81D IIoffollffliI
-
0
100
300
200
400
S0
TIME (ms)
C
il 1 I l l I
lI
,q
i
2i
e
_o
2
1
3
0 -
5
4
I
3
2
II
4
4
6
d~~~~~2S16~~~~~~
x
.
i
11
CIN
5;5)s
3
0
1
2
3~5
106m
2
I
3
5S
4
$
\/\
r
r
I I i ~~~~\l/
-v ~ -a.,,-^FRE (Mi.
- -V'--
-v^
@QFR.e
(
1
0
2
Iv
FRE0
---)
-
)
4
5
.
_
fi
JL..
a
9 --
---��e
I��-I------�-�
�---�-ir_---�-�2-_IlL=L1
I"
I,-A,
.
EI
RwI
--_
I
vI
I
v
v-
v,,
-
Ah-
-L.
v,-
~--·--1~I-i
I
I
I
I
I
---
I':
I-I-
I- I
I
-1
I --
i
ml
--
I
/ip
f?
,/
,+f7
s
.-
'PL.,
v
1.0
I i! ii!
704
-
4
t_
--7--T--T--l
:F
-f-+
3 -
I I I
.rA
I
I
-,
.. !
I
.
: 1=
!
F| -. !' I
-
v
---
4-
- v-
j
v,'
0,
nV"V"v
--U]T: I
II
. I . -. . .
I
v
0E
E
I
t
4
I.
II
�)71-L. --
I.
.
*b~
V -
-4
-\
i
t ~=V._
-I
I I
.
I
-1
iz
tI
rei
;
.
s�lth�?i�s�·4��
F
I -
v
-
.
I .7re-
{ev
v
v
-V
v
.0
_..k_*
_I _-_. .J.A
._. _rL-C\~J4A/e~(*
4yV:
i'"
'
.~a -
d
A.
~o p
tUV'
itaU$
I ls
l/,
"1~
1
1
j I1
F
M I
r-ly-F-W -11I1 1ILIII II II IV
I F I
ln l /-r 1
gv!
-
4!
v- v--
1S I'H
'-... 1 ll
'p
v-- v-
^'
----
-V
I
.
I
.r
I
l'
yv-,
V"'V.V"" 5
!;
BH
!-W
474-H
...
I I
LI ,i
I...
I-
"�'
VVV
v
i
. I. . .
"
V
ff
f~
V
Vvv
t
Ct-l
V
A_"
,,
-·--�
V
,.
,-
l
I j--
V
I
4 7­
.*.
ntiL.
W_L_V4_ULUl~YFI~
"
v -
i
I
I
-
III
v-' VI'"
a~
~~
--