I
AE
S In
t
er
na
t
io
na
l J
o
urna
l o
f
Art
if
icia
l In
t
ellig
ence
(
I
J
-
AI
)
Vo
l.
15
,
No
.
5
,
Octo
b
er
20
26
,
p
p
.
4
8
7
8
~
489
2
I
SS
N:
2
2
5
2
-
8
9
3
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijai.v
15
.i
5
.
p
p
4
8
7
8
-
489
2
4878
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
a
i
.
ia
esco
r
e.
co
m
Exploring
Arabic
non
-
sibila
nt
frica
tive con
so
na
nts:
a
co
mprehens
iv
e cl
a
ss
ificatio
n
Yo
us
s
ef
E
lf
a
hm
1
,
Nesrine
A
ba
j
a
dd
i
1
,
B
a
dia
M
o
un
ir
2
,
L
a
ila
E
lm
a
a
zo
uzi
2
,
I
lha
m
M
o
un
ir
2
,
Abdelm
a
j
id F
a
rc
hi
1
1
I
M
I
I
La
b
o
r
a
t
o
r
y
,
D
e
p
a
r
t
me
n
t
o
f
El
e
c
t
r
i
c
a
l
a
n
d
C
o
mp
u
t
e
r
E
n
g
i
n
e
e
r
i
n
g
,
F
a
c
u
l
t
y
o
f
S
c
i
e
n
c
e
s
a
n
d
Te
c
h
n
i
q
u
e
s
,
H
a
s
san
F
i
r
st
U
n
i
v
e
r
si
t
y
,
S
e
t
t
a
t
,
M
o
r
o
c
c
o
2
LA
P
S
S
I
I
La
b
o
r
a
t
o
r
y
,
H
i
g
h
S
c
h
o
o
l
o
f
Te
c
h
n
o
l
o
g
y
,
C
a
d
i
A
y
y
a
d
U
n
i
v
e
r
si
t
y
,
S
a
f
i
,
M
o
r
o
c
c
o
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Oct
16
,
2
0
2
5
R
ev
is
ed
Au
g
1
,
2
0
2
6
Acc
ep
ted
Au
g
12
,
2
0
2
6
Vo
ice
in
tera
c
ti
o
n
se
rv
e
s
a
s
a
n
e
fficie
n
t
to
o
l
fo
r
imp
ro
v
i
n
g
c
o
m
m
u
n
ica
ti
o
n
b
e
twe
e
n
h
u
m
a
n
s
a
n
d
m
a
c
h
i
n
e
s,
e
n
a
b
li
n
g
q
u
ick
e
r
e
x
c
h
a
n
g
e
s
w
it
h
o
u
t
th
e
n
e
e
d
fo
r
p
h
y
sic
a
l
c
o
n
tac
t
with
d
e
v
ice
s.
Th
is
st
u
d
y
i
n
tr
o
d
u
c
e
s
a
m
e
th
o
d
o
l
o
g
y
a
ime
d
a
t
id
e
n
ti
fy
in
g
Ara
b
ic
n
o
n
-
sib
il
a
n
t
frica
ti
v
e
c
o
n
so
n
a
n
ts
a
n
d
re
c
o
g
n
izin
g
t
h
e
ir
si
g
n
ifi
c
a
n
c
e
w
it
h
in
th
is
c
o
n
tex
t.
O
u
r
a
p
p
ro
a
c
h
in
v
o
l
v
e
s
a
n
a
ly
z
in
g
th
e
n
o
rm
a
li
z
e
d
e
n
e
rg
y
d
istri
b
u
ti
o
n
wit
h
in
c
o
n
so
n
a
n
t
-
v
o
we
l
p
h
o
n
e
ti
c
sy
ll
a
b
les
,
w
h
ich
se
rv
e
s
a
s
th
e
a
c
o
u
stic
i
n
p
u
t
si
g
n
a
l
fo
r
o
u
r
sp
e
e
c
h
re
c
o
g
n
it
i
o
n
sy
ste
m
.
No
tab
ly
,
t
h
e
re
lativ
e
ly
li
m
it
e
d
a
tt
e
n
ti
o
n
g
i
v
e
n
t
o
n
o
n
-
sib
il
a
n
t
c
o
n
s
o
n
a
n
ts
i
n
e
x
isti
n
g
l
it
e
ra
tu
re
u
n
d
e
rsc
o
re
s
th
e
imp
o
r
tan
c
e
o
f
t
h
is
re
se
a
rc
h
e
n
d
e
a
v
o
r.
T
h
e
e
v
a
lu
a
t
i
o
n
o
f
o
u
r
c
las
sifica
ti
o
n
m
e
th
o
d
a
p
p
li
e
d
t
o
o
u
r
d
a
tas
e
t
h
i
g
h
l
ig
h
ts
t
h
e
p
i
v
o
tal
r
o
le
o
f
v
o
c
a
l
si
g
n
a
l
e
n
e
rg
y
in
c
h
a
ra
c
teriz
in
g
Ara
b
ic
n
o
n
-
si
b
il
a
n
t
frica
ti
v
e
c
o
n
so
n
a
n
ts.
Ac
h
iev
in
g
a
c
las
sifica
ti
o
n
a
c
c
u
ra
c
y
o
f
9
0
%
f
o
r
d
isti
n
g
u
ish
i
n
g
v
o
ice
d
a
n
d
v
o
ice
les
s
non
-
sib
il
a
n
t
c
o
n
so
n
a
n
ts,
wit
h
a
ra
te
e
x
c
e
e
d
in
g
8
5
%
f
o
r
o
t
h
e
r
c
a
teg
o
ries
,
u
n
d
e
rsc
o
re
s
t
h
e
e
ffe
c
ti
v
e
n
e
ss
o
f
o
u
r
a
p
p
r
o
a
c
h
.
N
o
tew
o
rt
h
y
is
t
h
e
o
u
t
p
e
rfo
rm
a
n
c
e
o
f
o
u
r
c
las
sif
ica
ti
o
n
tec
h
n
iq
u
e
c
o
m
p
a
re
d
t
o
c
u
rre
n
t
a
lg
o
rit
h
m
s
b
a
se
d
o
n
su
p
p
o
rt
v
e
c
to
r
m
a
c
h
in
e
s
(S
V
M
)
a
n
d
a
rti
f
i
c
ial
n
e
u
ra
l
n
e
two
rk
s
(AN
Ns
)
,
in
d
ica
ti
n
g
it
s
p
o
ten
ti
a
l
t
o
e
n
h
a
n
c
e
Ara
b
ic
a
u
to
m
a
ti
c
sp
e
e
c
h
re
c
o
g
n
it
io
n
(ASR)
s
y
ste
m
s.
Ov
e
ra
ll
,
th
is
re
se
a
rc
h
c
o
n
t
rib
u
tes
t
o
a
d
v
a
n
c
in
g
sp
e
e
c
h
re
c
o
g
n
it
io
n
tec
h
n
o
lo
g
y
,
p
a
rti
c
u
larly
i
n
t
h
e
re
a
lm
o
f
Ara
b
ic
li
n
g
u
isti
c
s,
o
f
fe
rin
g
p
r
o
m
isin
g
i
m
p
li
c
a
ti
o
n
s
f
o
r
fu
t
u
re
d
e
v
e
l
o
p
m
e
n
ts
in
t
h
is
field
.
K
ey
w
o
r
d
s
:
Ar
ab
ic
lan
g
u
a
g
e
Ar
tific
ial
n
eu
r
al
n
etwo
r
k
s
Au
to
m
atic
s
p
ee
ch
r
ec
o
g
n
itio
n
E
n
er
g
y
in
f
r
e
q
u
en
c
y
b
a
n
d
s
Fric
ativ
e
co
n
s
o
n
an
ts
No
n
-
s
ib
ilan
t
Su
p
p
o
r
t
v
ec
to
r
m
ac
h
in
es
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Yo
u
s
s
ef
E
lf
ah
m
I
MI
I
L
ab
o
r
ato
r
y
,
Dep
ar
tm
e
n
t
o
f
E
lectr
ical
an
d
C
o
m
p
u
ter
E
n
g
in
ee
r
in
g
Facu
lty
o
f
Scien
ce
s
an
d
T
ec
h
n
iq
u
es
Hass
an
First Un
iv
er
s
ity
R
o
ad
o
f
C
asab
lan
ca
B
.
P: 5
7
7
,
Settat,
Mo
r
o
cc
o
E
m
ail: y
.
elf
ah
m
@
u
h
p
.
ac
.
m
a
1.
I
NT
RO
D
UCT
I
O
N
I
n
th
e
er
a
o
f
tech
n
o
lo
g
ical
p
r
o
g
r
ess
,
r
o
b
o
ts
h
av
e
b
ec
o
m
e
in
c
r
ea
s
in
g
ly
in
teg
r
ated
in
to
b
o
th
in
d
u
s
tr
ial
an
d
d
o
m
esti
c
asp
ec
ts
o
f
o
u
r
d
aily
liv
es.
Acc
o
r
d
in
g
to
th
e
I
n
ter
n
atio
n
al
Fed
er
atio
n
o
f
R
o
b
o
tics
(
I
FR
)
,
in
2
0
2
2
alo
n
e,
th
e
2
7
m
em
b
er
s
tates
o
f
th
e
E
u
r
o
p
ea
n
Un
io
n
in
s
talled
n
ea
r
l
y
7
2
,
0
0
0
r
o
b
o
tic
u
n
its
[
1
]
,
s
h
o
wca
s
in
g
th
e
p
er
v
asiv
e
p
r
esen
ce
o
f
m
ac
h
in
es.
T
h
is
wid
esp
r
ea
d
ad
o
p
tio
n
h
as
s
p
u
r
r
ed
th
e
ad
v
a
n
ce
m
en
t
o
f
v
o
ice
in
ter
ac
tio
n
as
an
ef
f
icien
t
m
o
d
e
o
f
co
m
m
u
n
icatio
n
b
etwe
en
h
u
m
an
s
an
d
m
ac
h
in
es.
T
o
c
o
m
p
r
e
h
en
d
h
u
m
an
s
p
ee
ch
,
m
ac
h
in
es
m
u
s
t
b
e
eq
u
ip
p
ed
w
ith
an
au
to
m
atic
s
p
ee
ch
r
ec
o
g
n
itio
n
(
ASR
)
s
y
s
tem
th
at
tr
an
s
lates
s
p
o
k
en
wo
r
d
s
in
to
tex
t,
f
ac
ilit
atin
g
ef
f
ec
tiv
e
co
m
m
u
n
icatio
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
E
xp
lo
r
in
g
A
r
a
b
ic
n
o
n
-
s
ib
ila
n
t fri
ca
tive
co
n
s
o
n
a
n
ts
:
a
co
mp
r
eh
en
s
ive
cla
s
s
ifica
tio
n
(
Yo
u
s
s
ef
E
lfa
h
m
)
4879
T
h
is
p
ap
er
in
tr
o
d
u
ce
s
an
Ar
ab
ic
s
p
ee
ch
r
ec
o
g
n
itio
n
s
y
s
tem
em
p
lo
y
in
g
an
an
aly
tical
ap
p
r
o
ac
h
b
ased
o
n
th
e
id
e
n
tific
atio
n
o
f
p
h
o
n
e
m
es,
th
e
f
u
n
d
am
en
tal
u
n
its
o
f
s
o
u
n
d
.
Ph
o
n
em
es
ca
n
m
an
if
e
s
t
as
co
n
s
o
n
a
n
ts
o
r
v
o
wels
[
2
]
,
with
co
n
s
o
n
an
t
p
r
o
d
u
ctio
n
in
v
o
lv
i
n
g
th
e
o
b
s
tr
u
ctio
n
o
f
air
f
lo
w
th
r
o
u
g
h
th
e
v
o
ca
l
tr
ac
t
b
y
p
r
ec
is
e
p
o
s
itio
n
in
g
a
n
d
m
a
n
ip
u
latio
n
o
f
ar
ticu
lato
r
s
s
u
ch
as
th
e
t
o
n
g
u
e,
lip
s
,
a
n
d
teeth
.
T
h
is
o
b
s
tr
u
ctio
n
r
esu
lts
in
d
is
tin
ct
co
n
s
o
n
an
t
s
o
u
n
d
s
,
s
u
ch
as
th
e
clo
s
u
r
e
o
f
lip
s
p
r
o
d
u
cin
g
t
h
e
p
l
o
s
iv
e
/p
/
o
r
th
e
n
ar
r
o
win
g
o
f
s
p
ac
e
b
etwe
en
teeth
an
d
lo
wer
lip
p
r
o
d
u
cin
g
th
e
f
r
icativ
e
/f/
.
C
o
n
tr
asti
n
g
ly
,
v
o
wels
ar
e
f
o
r
m
ed
b
y
allo
win
g
u
n
r
estricte
d
air
f
lo
w
th
r
o
u
g
h
th
e
v
o
ca
l
tr
ac
t
[
3
]
.
Ou
r
o
b
jectiv
e
is
to
ch
ar
ac
ter
ize
an
d
id
en
tify
f
r
icativ
e
co
n
s
o
n
an
ts
,
a
s
ig
n
if
ican
t
co
m
p
o
n
en
t
o
f
th
e
Ar
ab
ic
s
o
u
n
d
s
y
s
tem
,
co
n
s
titu
tin
g
ab
o
u
t
h
alf
o
f
th
e
co
n
s
o
n
an
ts
in
th
e
alp
h
a
b
et.
T
h
ese
f
r
icativ
es
ar
e
d
is
tin
g
u
is
h
ed
b
y
th
e
h
is
s
in
g
s
o
u
n
d
g
en
er
ate
d
th
r
o
u
g
h
t
h
e
n
ar
r
o
win
g
o
f
th
e
v
o
ca
l
tr
ac
t,
f
o
r
ci
n
g
air
th
r
o
u
g
h
a
r
estricte
d
ch
an
n
el
[
4
]
.
T
h
e
class
if
icatio
n
o
f
f
r
icativ
e
co
n
s
o
n
an
ts
,
p
ar
ticu
lar
ly
s
ib
ilan
t
an
d
non
-
s
ib
ilan
t
[
5
]
,
is
ex
p
lo
r
ed
in
th
e
s
cien
tific
liter
atu
r
e.
Sib
ilan
t
f
r
icativ
es
in
v
o
lv
e
d
ir
ec
tin
g
air
f
lo
w
with
th
e
to
n
g
u
e
to
w
ar
d
s
clo
s
ed
teeth
,
p
r
o
d
u
ci
n
g
a
wh
is
tlin
g
n
o
is
e.
I
n
Ar
ab
i
c,
s
ib
ilan
t
f
r
icativ
es
in
clu
d
e
/s
,
s
ҁ,
z
,
Ӡ
,
an
d
∫/
,
wh
ile
non
-
s
ib
ilan
t
f
r
icativ
es
in
cl
u
d
e
/f,
θ,
ð
,
ð
ҁ,
χ
,
ɣ
,
ʕ,
ћ
,
a
n
d
h/
.
Vo
icin
g
f
u
r
t
h
er
ca
teg
o
r
izes
f
r
icativ
es
in
to
v
o
iced
an
d
u
n
v
o
iced
,
wh
er
e
v
o
iced
f
r
icativ
es
in
v
o
lv
e
v
o
ca
l
f
o
ld
v
ib
r
atio
n
,
an
d
u
n
v
o
ice
d
o
n
es d
o
n
o
t.
I
n
th
i
s
w
o
r
k
p
r
o
p
o
s
e
s
an
an
a
l
y
t
i
ca
l
a
p
p
r
o
a
ch
f
o
r
ac
c
u
r
a
t
e
l
y
id
en
t
i
f
y
i
n
g
n
o
n
-
s
i
b
i
l
an
t
A
r
ab
i
c
f
r
i
c
a
t
iv
e
s
/
f
,
χ
,
ɣ
,
ʕ
,
ћ
,
a
n
d
h/
.
T
h
e
m
e
t
h
o
d
e
x
tr
a
c
t
s
a
s
u
b
s
e
t
o
f
a
c
o
u
s
t
i
c
in
d
i
ce
s
co
n
t
a
in
i
n
g
r
e
l
e
v
a
n
t
i
n
f
o
r
m
a
t
io
n
ab
o
u
t
t
h
e
s
e
f
r
i
ca
t
i
v
e
s
.
O
u
r
c
en
t
r
a
l
r
e
s
e
a
r
ch
q
u
e
s
t
i
o
n
i
s
:
c
a
n
th
e
en
e
r
g
y
d
i
s
t
r
i
b
u
t
i
o
n
w
i
t
h
in
s
p
e
c
i
f
i
c
f
r
eq
u
e
n
cy
b
a
n
d
s
o
f
t
h
e
v
o
ca
l
s
i
g
n
a
l
p
r
o
v
i
d
e
a
b
a
s
i
s
f
o
r
th
e
ac
c
u
r
a
t
e
c
la
s
s
i
f
i
c
a
t
i
o
n
o
f
Ar
a
b
i
c
non
-
s
i
b
i
l
an
t
f
r
i
c
a
t
iv
e
s
?
T
h
i
s
w
o
r
k
h
y
p
o
t
h
e
s
i
z
e
s
th
a
t
an
a
l
y
z
i
n
g
t
h
e
e
n
e
r
g
y
l
ev
e
l
s
i
n
c
a
r
e
f
u
l
l
y
s
e
l
e
c
t
ed
f
r
e
q
u
en
c
y
b
an
d
s
w
i
l
l
y
ie
l
d
d
i
s
t
i
n
c
t
p
a
t
t
e
r
n
s
t
h
a
t
a
l
lo
w
f
o
r
t
h
e
r
e
li
a
b
l
e
d
if
f
er
e
n
t
ia
t
i
o
n
o
f
th
e
s
e
f
r
i
c
a
ti
v
e
s
,
p
o
t
en
t
i
a
l
l
y
o
f
f
e
r
in
g
a
co
m
p
u
ta
t
i
o
n
a
l
ly
e
f
f
i
c
i
en
t
a
l
t
e
r
n
a
t
i
v
e
t
o
m
o
r
e
c
o
m
p
l
ex
m
a
ch
i
n
e
l
e
ar
n
i
n
g
a
p
p
r
o
a
ch
e
s
.
T
h
e
ar
t
i
c
l
e
c
o
m
p
ar
e
s
th
e
p
r
ed
i
c
t
i
v
e
p
o
w
er
o
f
th
e
s
e
p
ar
a
m
e
te
r
s
w
i
t
h
t
h
e
c
o
m
m
o
n
a
p
p
r
o
a
ch
b
a
s
e
d
o
n
n
eu
r
a
l
n
e
t
w
o
r
k
s
an
d
s
u
p
p
o
r
t
v
e
c
t
o
r
m
a
c
h
in
e
s
(
S
V
M
)
.
T
h
e
s
tr
u
ctu
r
e
o
f
th
e
ar
ticle
in
clu
d
es
a
liter
atu
r
e
r
e
v
iew
o
n
ac
o
u
s
tic
cu
es
f
o
r
f
r
icativ
es,
f
o
llo
wed
b
y
th
e
p
r
esen
tatio
n
o
f
m
eth
o
d
s
an
d
to
o
ls
u
s
ed
in
th
e
s
tu
d
y
.
T
h
e
r
esu
lts
o
f
th
e
an
aly
s
i
s
an
d
class
if
icatio
n
alg
o
r
ith
m
s
ar
e
th
en
d
is
cu
s
s
e
d
,
co
m
p
a
r
in
g
th
em
with
two
d
is
tin
ct
class
if
ier
s
:
n
eu
r
al
n
etwo
r
k
s
an
d
SVM
m
ac
h
in
es.
T
h
e
co
n
clu
s
io
n
ad
d
r
ess
es
th
e
m
eth
o
d
s
'
ad
v
an
tag
es
an
d
lim
itatio
n
s
,
d
is
cu
s
s
in
g
t
h
e
im
p
licatio
n
s
f
o
r
u
n
d
er
s
tan
d
i
n
g
n
o
n
-
s
ib
ilan
t
f
r
icativ
es in
th
e
Ar
ab
ic
lan
g
u
ag
e
.
T
h
e
r
esear
ch
liter
atu
r
e
o
n
f
r
i
ca
tiv
e
co
n
s
o
n
an
ts
r
e
v
ea
ls
a
wea
lth
o
f
s
tu
d
ies
ex
p
lo
r
in
g
th
e
s
e
s
o
u
n
d
s
,
wh
ich
h
av
e
g
a
r
n
er
ed
e
x
ten
s
iv
e
atten
tio
n
.
Desp
ite
th
is
,
o
u
r
u
n
d
er
s
tan
d
in
g
o
f
th
e
r
ela
tio
n
s
h
ip
s
b
etwe
en
ar
ticu
lato
r
y
o
r
g
an
s
,
ac
o
u
s
tic
ch
an
g
es,
an
d
p
er
ce
p
tu
al
m
ec
h
an
is
m
s
ass
o
ciate
d
with
f
r
icativ
es
r
em
ain
s
in
co
m
p
lete.
T
h
e
p
r
o
d
u
ctio
n
o
f
f
r
icativ
e
c
o
n
s
o
n
a
n
ts
in
v
o
lv
e
s
two
p
ar
ts
o
f
th
e
v
o
ca
l
tr
ac
t
ap
p
r
o
ac
h
in
g
ea
ch
o
th
er
,
cr
ea
tin
g
a
n
a
r
r
o
w
p
ass
ag
e
wh
er
e
tu
r
b
u
len
t a
ir
f
lo
w
p
r
o
d
u
ce
s
th
e
f
r
icativ
e
s
o
u
n
d
[
5
]
.
W
ith
in
th
e
th
eo
r
y
o
f
i
n
v
ar
ia
n
ts
[
6
]
,
s
tu
d
ies
em
p
h
asize
th
e
cr
u
cial
r
o
le
o
f
tem
p
o
r
al,
s
p
ec
tr
al,
an
d
am
p
litu
d
e
c
h
ar
ac
ter
is
tics
in
a
co
u
s
tic
s
ig
n
als
as
ess
en
tial
p
er
ce
p
tu
al
cu
es.
Stab
le
ac
o
u
s
tic
f
ea
tu
r
es,
s
u
c
h
as
v
o
ice
o
n
s
et
tim
e
(
VOT
)
f
o
r
p
l
o
s
iv
e
co
n
s
o
n
a
n
ts
an
d
f
o
r
m
an
t
s
f
o
r
v
o
wels,
h
av
e
b
ee
n
id
en
ti
f
ied
.
Ho
wev
e
r
,
th
e
ex
is
ten
ce
o
f
s
im
ilar
in
v
ar
ian
t
ac
o
u
s
tic
p
r
o
p
e
r
ties
f
o
r
f
r
icativ
es
h
as
s
p
ar
k
ed
co
n
tr
o
v
e
r
s
y
,
with
s
o
m
e
s
tu
d
ies
ch
allen
g
in
g
th
e
n
o
tio
n
o
f
u
n
iv
er
s
al
f
ea
tu
r
es
an
d
ad
v
o
c
atin
g
f
o
r
d
if
f
er
e
n
tiatio
n
b
et
wee
n
s
ib
ilan
t
an
d
non
-
s
ib
ilan
t
v
ar
ieties.
Mo
s
t
s
tu
d
ies
aim
ed
at
c
h
ar
ac
ter
izin
g
f
r
icativ
e
co
n
s
o
n
a
n
ts
h
av
e
p
r
im
ar
ily
ce
n
ter
ed
ar
o
u
n
d
th
e
v
o
icele
s
s
f
r
icativ
es
o
f
E
n
g
lis
h
[
7
]
–
[
9
]
.
Ne
v
er
th
eless
,
th
er
e
is
a
n
o
tab
le
d
ef
i
cien
cy
in
r
esear
c
h
co
n
ce
r
n
in
g
f
r
icativ
es
in
o
th
e
r
lan
g
u
ag
es,
with
a
p
ar
ticu
lar
d
ea
r
th
o
f
in
v
esti
g
atio
n
s
in
to
th
e
o
v
er
all
c
o
n
s
o
n
an
t
s
y
s
tem
o
f
Ar
ab
ic,
esp
ec
ially
f
o
cu
s
in
g
o
n
f
r
icativ
es.
C
o
n
ce
r
n
in
g
t
h
e
ef
f
ec
ts
o
f
v
a
r
io
u
s
v
o
ca
l
tr
ac
t
co
n
f
i
g
u
r
atio
n
s
d
u
r
in
g
f
r
icativ
e
p
r
o
d
u
ctio
n
o
n
v
a
r
io
u
s
ac
o
u
s
tic
m
ea
s
u
r
es,
it is
g
en
er
ally
ac
ce
p
ted
th
at
th
e
s
ize
an
d
s
h
ap
e
o
f
th
e
v
o
ca
l tr
ac
t d
eter
m
i
n
e
th
e
s
p
ec
tr
u
m
o
f
a
f
r
icativ
e
[
1
0
]
.
T
h
is
is
co
n
s
id
er
ed
well
-
d
escr
ib
e
d
b
y
ac
o
u
s
tic
f
ea
tu
r
es
s
u
ch
as
th
e
lo
ca
t
io
n
o
f
th
e
s
p
ec
tr
al
p
ea
k
an
d
th
e
f
o
u
r
s
p
ec
tr
al
m
o
m
en
ts
(
s
p
ec
tr
al
m
ea
n
,
s
p
r
ea
d
,
s
k
ewn
ess
,
an
d
k
u
r
to
s
is
)
[
1
1
]
–
[
1
4
]
.
Ad
d
itio
n
ally
,
f
r
icativ
es
ar
e
n
o
t
i
m
m
u
n
e
to
c
o
ar
ticu
latio
n
,
an
d
ar
t
icu
lato
r
m
o
v
em
en
ts
h
av
e
s
ig
n
if
ican
t
ac
o
u
s
tic
co
n
s
eq
u
en
ce
s
f
o
r
s
p
ec
tr
al
e
n
er
g
y
d
is
tr
ib
u
tio
n
.
Sev
e
r
al
s
tu
d
ies
em
p
h
asize
th
e
im
p
o
r
tan
ce
o
f
c
o
n
s
id
er
in
g
lan
g
u
ag
e
-
s
p
ec
if
ic
an
d
s
p
ea
k
er
-
s
p
ec
if
ic
ch
a
r
ac
ter
is
tics
wh
en
s
tu
d
y
i
n
g
f
r
icativ
es.
C
o
n
s
eq
u
en
tly
,
th
e
r
ep
r
esen
tatio
n
o
f
s
p
ec
tr
o
-
tem
p
o
r
al
d
is
tr
ib
u
tio
n
h
as
b
ee
n
s
u
cc
ess
f
u
lly
u
tili
ze
d
to
ex
am
in
e
s
u
b
tle
d
if
f
er
en
ce
s
am
o
n
g
v
o
icele
s
s
f
r
icativ
es
[
1
5
]
.
Fu
r
th
er
m
o
r
e,
it
h
as
b
ee
n
o
b
s
er
v
ed
th
at
th
e
v
o
wel
f
o
llo
w
in
g
th
e
f
r
icativ
e
h
as
a
m
o
r
e
s
ig
n
if
ica
n
t
im
p
ac
t
o
n
th
e
ac
o
u
s
tic
ch
ar
ac
ter
is
tics
o
f
/f/
th
an
/s
/
,
an
d
ev
en
less
o
n
/∫/
[
1
0
]
.
T
h
ese
f
in
d
in
g
s
s
u
g
g
est
th
e
n
ec
ess
ity
o
f
co
n
s
id
er
in
g
v
ar
io
u
s
v
ar
ia
b
les
th
at
ca
n
in
f
lu
e
n
ce
th
e
ac
o
u
s
tic
p
r
o
p
er
ties
o
f
f
r
icativ
es to
b
etter
u
n
d
er
s
tan
d
th
e
u
n
d
e
r
ly
in
g
m
ec
h
an
is
m
s
o
f
th
eir
p
r
o
d
u
ctio
n
.
T
h
e
s
p
ec
tr
al
p
ea
k
,
r
ep
r
esen
tin
g
th
e
m
o
s
t
in
ten
s
e
f
r
eq
u
e
n
c
y
in
th
e
v
o
ca
l
s
ig
n
al,
is
an
ac
o
u
s
tical
p
ar
am
eter
e
x
ten
s
iv
ely
s
tu
d
ie
d
in
th
e
co
n
te
x
t
o
f
f
r
icativ
e
s
.
Nu
m
er
o
u
s
r
esear
ch
ef
f
o
r
ts
h
av
e
f
o
cu
s
ed
o
n
u
n
d
er
s
tan
d
i
n
g
th
e
r
elatio
n
s
h
i
p
b
etwe
en
to
n
g
u
e
p
o
s
itio
n
an
d
th
e
f
r
eq
u
en
cy
o
f
th
e
s
p
ec
tr
al
p
ea
k
d
u
r
in
g
f
r
icativ
e
p
r
o
d
u
ctio
n
at
d
i
f
f
er
e
n
t
lo
ca
tio
n
s
in
t
h
e
m
o
u
th
.
I
t
is
ass
u
m
ed
th
at
th
is
f
r
eq
u
e
n
cy
d
ec
r
ea
s
es
f
r
o
m
h
i
g
h
to
lo
w
f
r
eq
u
en
cies
as
th
e
to
n
g
u
e
m
o
v
es
f
r
o
m
th
e
f
r
o
n
t
to
th
e
b
ac
k
o
f
th
e
m
o
u
th
[
1
6
]
,
[
1
7
]
.
Ho
wev
e
r
,
th
is
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
15
,
No
.
5
,
Octo
b
er
20
26
:
4
8
7
8
-
489
2
4880
r
elatio
n
s
h
ip
h
as
n
o
t
b
ee
n
c
o
n
f
ir
m
e
d
f
o
r
Gr
ee
k
f
r
icativ
es,
as
d
em
o
n
s
tr
ated
b
y
Nir
g
ian
ak
i
et
a
l.
[
1
8
]
.
Mo
r
eo
v
er
,
th
e
s
p
ec
tr
al
p
ea
k
h
as
p
r
o
v
en
v
alu
a
b
le
in
d
is
tin
g
u
is
h
in
g
b
etwe
en
s
ib
ilan
t
an
d
non
-
s
ib
ilan
t
f
r
icativ
es,
s
p
ec
if
ically
d
if
f
er
e
n
tiatin
g
b
etwe
en
alv
eo
lar
an
d
p
o
s
t
-
alv
eo
lar
s
ib
ilan
ts
[
8
]
,
[
1
9
]
–
[
2
2
]
.
T
h
e
f
r
eq
u
en
cies
an
d
i
n
ten
s
ity
o
f
th
e
f
r
ictio
n
n
o
is
e
in
v
o
ice
d
f
r
icativ
es
ar
e
ac
o
u
s
tical
f
ea
tu
r
es
th
at
d
if
f
er
en
tiate
b
etwe
en
co
n
s
o
n
an
ts
b
ased
o
n
th
eir
ar
ticu
latio
n
p
o
in
t.
L
ab
io
d
en
tal
f
r
icativ
es
p
r
o
d
u
ce
lo
w
-
in
ten
s
ity
n
o
is
e
with
d
if
f
u
s
e
p
ea
k
s
ar
o
u
n
d
3
.
5
k
Hz
an
d
8
k
Hz.
Alv
eo
lar
f
r
icativ
es,
o
n
th
e
o
th
er
h
an
d
,
h
a
v
e
m
o
r
e
in
te
n
s
e
n
o
is
e
r
a
n
g
in
g
b
etwe
en
4
k
Hz
an
d
8
k
Hz,
wi
th
p
ea
k
s
a
r
o
u
n
d
5
k
Hz
an
d
8
k
Hz.
Fin
ally
,
f
o
r
p
alata
l
f
r
icativ
es,
tu
r
b
u
len
ce
n
o
is
e
is
b
etwe
en
2
k
Hz
an
d
7
k
Hz,
with
a
d
if
f
u
s
e
p
ea
k
av
er
ag
in
g
ar
o
u
n
d
4
k
Hz
[
2
3
]
–
[
2
5
]
.
An
o
th
er
ac
o
u
s
tical
p
ar
a
m
eter
f
r
e
q
u
en
tly
em
p
l
o
y
ed
in
th
e
s
tu
d
y
o
f
f
r
icativ
es
i
s
th
e
f
ir
s
t
s
p
ec
tr
al
m
o
m
en
t,
wh
ich
ca
n
b
e
d
ef
in
e
d
eith
er
as
th
e
av
e
r
ag
e
o
f
th
e
s
p
ec
tr
al
en
er
g
y
d
i
s
tr
ib
u
tio
n
o
r
as
th
e
ce
n
ter
o
f
g
r
a
v
ity
(
C
o
G)
o
f
th
e
f
r
icativ
e
s
ig
n
al
[
2
6
]
.
R
esear
ch
in
d
icate
s
th
at
th
e
C
o
G
ca
n
d
is
tin
g
u
is
h
b
etwe
en
s
ib
ilan
t
an
d
n
o
n
-
s
ib
ilan
t
co
n
s
o
n
an
ts
,
as
well
as
d
if
f
er
e
n
t
ty
p
es
o
f
s
ib
ilan
ts
.
Stu
d
ies
h
av
e
s
h
o
wn
th
at
s
ib
ilan
ts
h
av
e
h
ig
h
er
C
o
G
v
alu
es
th
a
n
n
o
n
-
s
ib
ilan
t
s
[
2
7
]
,
[
2
8
]
,
a
n
d
th
e
co
n
s
o
n
an
t
/s
/
h
as
a
h
i
g
h
er
C
o
G
th
an
th
e
co
n
s
o
n
an
t
/∫/
[
1
7
]
,
[
2
9
]
–
[
3
2
]
.
As
f
o
r
th
e
s
ec
o
n
d
s
p
ec
tr
al
m
o
m
en
t,
wh
ic
h
m
ea
s
u
r
es
th
e
a
m
o
u
n
t
o
f
s
p
r
ea
d
o
r
v
ar
ian
ce
o
f
s
p
ec
tr
al
en
er
g
y
a
r
o
u
n
d
th
e
m
ea
n
,
p
r
e
v
io
u
s
r
esear
ch
h
as
d
em
o
n
s
tr
a
ted
th
at
s
ib
ilan
t
f
r
icativ
es
ex
h
ib
it
lo
wer
s
p
ec
tr
al
v
ar
ian
ce
th
an
n
o
n
-
s
ib
ilan
t
s
[
2
7
]
,
with
th
e
p
o
s
t
-
alv
eo
lar
f
r
ic
ativ
e
/∫/
h
av
in
g
th
e
lo
west
s
p
ec
tr
al
v
ar
ian
ce
[
1
1
]
.
Oth
er
s
tu
d
ies
h
av
e
r
ep
o
r
ted
f
in
d
in
g
s
r
eg
ar
d
i
n
g
t
h
e
th
ir
d
an
d
f
o
u
r
th
m
o
m
en
ts
o
f
th
e
s
p
ec
tr
u
m
,
n
a
m
ely
s
k
ewn
ess
an
d
k
u
r
to
s
is
.
Sk
ew
n
ess
ch
ar
ac
ter
izes
th
e
s
lo
p
e
o
f
th
e
s
p
ec
tr
u
m
an
d
p
r
o
v
id
e
s
a
m
ea
s
u
r
e
o
f
th
e
o
v
er
all
asy
m
m
etr
y
o
f
th
e
e
n
e
r
g
y
d
is
tr
ib
u
tio
n
.
A
s
k
ewn
ess
v
alu
e
o
f
ze
r
o
in
d
icate
s
a
s
y
m
m
etr
ic
d
is
tr
ib
u
tio
n
ce
n
ter
ed
ar
o
u
n
d
th
e
m
ea
n
.
Po
s
itiv
e
s
k
ewn
ess
im
p
lies
a
n
eg
ativ
e
s
lo
p
e
with
en
er
g
y
co
n
ce
n
tr
atio
n
in
l
o
w
f
r
eq
u
e
n
cies,
wh
ile
n
eg
ativ
e
s
k
ewn
ess
in
d
icate
s
a
p
o
s
itiv
e
s
lo
p
e
an
d
en
er
g
y
p
r
e
v
alen
ce
in
h
ig
h
er
f
r
eq
u
e
n
cies
[
3
3
]
,
[
3
4
]
.
K
u
r
to
s
is
,
o
n
th
e
o
th
er
h
an
d
,
r
ef
er
s
t
o
th
e
“
p
ea
k
ed
n
ess
”
o
r
f
latn
ess
o
f
th
e
d
is
tr
ib
u
tio
n
.
A
k
u
r
t
o
s
is
o
f
3
i
n
d
icate
s
a
n
o
r
m
al
d
is
tr
ib
u
tio
n
,
w
h
ile
a
v
al
u
e
lo
wer
th
an
3
s
u
g
g
ests
a
f
la
t
d
is
tr
ib
u
tio
n
,
an
d
a
h
ig
h
er
v
alu
e
m
ea
n
s
a
s
h
ar
p
er
d
is
tr
ib
u
tio
n
[
3
3
]
,
[
3
4
]
.
R
esear
ch
er
s
h
av
e
s
u
g
g
ested
th
at
s
k
e
wn
ess
an
d
k
u
r
to
s
is
ca
n
p
o
ten
tially
d
if
f
er
e
n
tiate
b
etwe
en
th
e
co
n
s
o
n
an
ts
/s
/
an
d
/∫/
,
f
in
d
in
g
th
at
/s
/
ex
h
ib
its
n
eg
ativ
e
s
k
ewn
ess
wh
ile
th
e
f
r
icativ
e
/∫/
h
as p
o
s
itiv
e
s
k
ewn
ess
[
2
7
]
,
[
3
1
]
,
[
3
5
]
.
Ho
wev
er
,
o
th
er
s
tu
d
ies h
av
e
a
ls
o
r
ep
o
r
ted
th
at
/s
/
h
ad
a
h
ig
h
e
r
p
o
s
itiv
e
s
k
ewn
e
s
s
th
an
/∫/
[
2
7
]
.
I
n
ter
m
s
o
f
k
u
r
to
s
is
,
r
esear
ch
h
as
r
ev
ea
led
th
at
/s
/
h
as
a
lar
g
e
p
o
s
itiv
e
v
alu
e
wh
ile
/∫/
h
as a
s
m
all
p
o
s
itiv
e
o
r
n
e
g
ativ
e
v
alu
e
[
1
7
]
,
[
2
7
]
,
[
3
1
]
,
[
3
5
]
.
R
eg
ar
d
in
g
t
h
e
c
h
ar
ac
ter
izatio
n
o
f
Ar
a
b
ic
f
r
icativ
es,
th
e
wo
r
k
o
f
Al
-
K
h
air
y
[
3
6
]
d
is
co
v
er
ed
th
at
non
-
v
o
iced
Ar
a
b
ic
f
r
icativ
es
h
av
e
a
lo
n
g
er
av
er
a
g
e
d
u
r
atio
n
o
f
f
r
ictio
n
n
o
is
e
th
an
v
o
ice
d
f
r
icativ
es.
Sib
ilan
t
f
r
icativ
es
h
av
e
a
lo
n
g
er
ab
s
o
l
u
te
av
er
ag
e
d
u
r
atio
n
o
f
f
r
icti
o
n
n
o
is
e
co
m
p
ar
ed
to
non
-
s
ib
ilan
t
f
r
icativ
es.
He
f
o
u
n
d
th
at
am
p
litu
d
e
m
ea
s
u
r
e
m
en
ts
ca
n
d
is
tin
g
u
is
h
b
etwe
e
n
s
ib
ilan
t
an
d
n
o
n
-
s
ib
ilan
t
f
r
i
ca
tiv
es
b
u
t
ca
n
n
o
t
id
en
tify
in
d
i
v
id
u
al
c
o
n
s
o
n
an
t
s
with
in
ea
ch
g
r
o
u
p
.
B
y
u
s
in
g
th
e
lo
ca
tio
n
o
f
t
h
e
s
p
ec
tr
al
p
ea
k
o
f
f
r
icativ
es,
Al
-
K
h
air
y
[
3
6
]
f
o
u
n
d
th
at
it
is
p
o
s
s
ib
le
to
d
is
tin
g
u
is
h
b
et
wee
n
s
ib
ilan
t
an
d
non
-
s
ib
ilan
t
f
r
icativ
es
b
u
t
n
o
t
b
etwe
en
th
e
p
o
s
t
-
alv
eo
lar
non
-
s
ib
ilan
t
s
/∫/
an
d
th
e
al
v
eo
lar
o
n
es
/s
/
an
d
/z
/
.
I
n
t
h
e
s
am
e
co
n
tex
t,
th
e
r
esu
lts
o
f
th
e
s
tu
d
y
co
n
d
u
cted
b
y
B
en
a
m
r
an
e
[
2
3
]
class
if
ied
Ar
a
b
ic
f
r
icativ
es
in
to
t
h
r
ee
g
r
o
u
p
s
b
ased
o
n
th
ei
r
p
lace
o
f
ar
ticu
latio
n
:
an
ter
io
r
f
r
icativ
es
/f
,
Ө
,
ð
,
ð
ʕ/
,
s
ib
ilan
t
f
r
icativ
es
/s
,
sҁ
,
z
,
Ӡ
,
∫/
,
an
d
p
o
s
ter
io
r
f
r
icativ
es
/χ
,
ɣ
,
ʕ
,
ћ
,
an
d
h/
.
An
ter
io
r
f
r
icativ
es
ex
h
ib
it
a
f
lat
s
p
ec
tr
u
m
with
f
ew
en
er
g
y
p
ea
k
s
an
d
r
elativ
ely
l
o
w
in
ten
s
ity
.
T
h
eir
f
r
eq
u
e
n
cy
C
o
G
is
m
o
d
er
ately
h
ig
h
b
u
t
is
th
e
h
ig
h
est
am
o
n
g
all
f
r
icativ
es
f
o
r
/ð
an
d
ð
ʕ/
.
Sib
ilan
t
f
r
icativ
es
h
av
e
a
co
m
p
ac
t sp
ec
tr
u
m
,
wit
h
a
m
ain
r
eg
io
n
o
f
h
ig
h
er
en
e
r
g
y
th
an
th
e
o
th
er
g
r
o
u
p
s
,
as w
ell
as h
ig
h
in
ten
s
ity
an
d
f
r
eq
u
e
n
cy
C
o
G
.
Fric
ativ
es
/s
/
an
d
/s
ҁ/
h
a
v
e
a
h
ig
h
e
r
C
o
G
th
an
/z
,
Ӡ,
an
d
∫/
.
Po
s
ter
io
r
f
r
icativ
es
h
av
e
in
ter
m
ed
iate
o
r
h
ig
h
i
n
ten
s
ity
an
d
a
m
o
d
er
ate
o
r
lo
w
f
r
eq
u
e
n
cy
C
o
G
.
B
u
ild
in
g
u
p
o
n
o
u
r
p
r
ev
io
u
s
wo
r
k
ch
ar
ac
ter
izin
g
Ar
ab
ic
f
r
icativ
es
u
s
in
g
en
er
g
y
d
is
tr
ib
u
tio
n
in
f
r
eq
u
e
n
cy
b
a
n
d
s
[
3
7
]
,
th
is
s
tu
d
y
ad
d
r
ess
es
th
e
m
o
r
e
c
h
allen
g
in
g
task
o
f
d
is
tin
g
u
is
h
in
g
b
etwe
en
in
d
iv
i
d
u
al
non
-
s
ib
ilan
t
f
r
icativ
es.
W
h
ile
p
r
io
r
r
esear
c
h
,
in
cl
u
d
in
g
o
u
r
o
wn
,
h
as
s
u
cc
ess
f
u
lly
d
if
f
er
en
tiated
b
etwe
en
s
ib
ilan
t
an
d
non
-
s
ib
ilan
t
f
r
icat
iv
es,
a
f
in
er
-
g
r
ai
n
ed
class
if
icatio
n
o
f
n
o
n
-
s
ib
ilan
t
s
r
em
ain
s
lar
g
ely
u
n
e
x
p
lo
r
e
d
,
p
ar
ticu
lar
ly
with
in
th
e
Ar
a
b
ic
lan
g
u
ag
e.
T
h
e
r
ar
ity
o
f
r
esear
ch
f
o
c
u
s
in
g
s
p
ec
if
ic
ally
o
n
in
d
i
v
id
u
al
non
-
s
ib
ilan
t
f
r
icativ
es
is
lar
g
ely
d
u
e
to
th
e
in
h
er
en
t
ac
o
u
s
tic
co
m
p
lex
ities
an
d
th
e
s
ig
n
if
ican
t
d
if
f
icu
lty
i
n
ac
cu
r
ately
class
if
y
in
g
an
d
id
e
n
tify
in
g
th
ese
p
h
o
n
em
es
co
m
p
ar
ed
to
th
eir
s
ib
ilan
t
co
u
n
ter
p
ar
ts
.
T
h
is
r
esear
ch
is
im
p
o
r
tan
t
b
ec
au
s
e
ac
cu
r
ate
id
en
tific
atio
n
o
f
th
ese
p
h
o
n
em
es
is
cr
u
cial
f
o
r
ad
v
an
cin
g
Ar
ab
ic
ASR
s
y
s
tem
s
,
im
p
r
o
v
in
g
th
eir
r
o
b
u
s
tn
ess
to
v
ar
iatio
n
s
in
s
p
ee
ch
a
n
d
ac
ce
n
ts
.
Fu
r
th
er
m
o
r
e
,
o
u
r
n
o
v
el
a
p
p
r
o
ac
h
lev
er
a
g
es
a
co
m
p
u
tatio
n
ally
ef
f
icien
t
m
e
th
o
d
,
an
al
y
zin
g
e
n
er
g
y
d
is
tr
ib
u
tio
n
in
ca
r
e
f
u
lly
s
elec
ted
f
r
eq
u
e
n
cy
b
a
n
d
s
,
o
f
f
er
in
g
a
p
o
ten
tial
ad
v
an
ta
g
e
o
v
er
c
o
m
p
lex
m
ac
h
in
e
lear
n
in
g
m
o
d
els
in
r
eso
u
r
ce
-
c
o
n
s
tr
a
in
ed
en
v
ir
o
n
m
e
n
ts
.
T
h
ese
f
in
d
in
g
s
n
o
t
o
n
ly
c
o
n
tr
ib
u
te
to
a
m
o
r
e
d
etailed
u
n
d
er
s
tan
d
i
n
g
o
f
th
e
ac
o
u
s
tic
p
r
o
p
er
ties
o
f
Ar
ab
ic
n
o
n
-
s
ib
ilan
t
f
r
icativ
e
s
b
u
t
also
p
av
e
th
e
way
f
o
r
m
o
r
e
ef
f
icien
t
an
d
ac
cu
r
ate
Ar
ab
ic
s
p
ee
c
h
r
ec
o
g
n
itio
n
tech
n
o
l
o
g
ies.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
E
xp
lo
r
in
g
A
r
a
b
ic
n
o
n
-
s
ib
ila
n
t fri
ca
tive
co
n
s
o
n
a
n
ts
:
a
co
mp
r
eh
en
s
ive
cla
s
s
ifica
tio
n
(
Yo
u
s
s
ef
E
lfa
h
m
)
4881
2.
M
E
T
H
O
D
I
n
th
is
s
ec
tio
n
,
th
e
o
v
er
all
p
r
i
n
cip
le
o
f
th
is
p
r
o
p
o
s
ed
m
eth
o
d
is
f
ir
s
tly
d
escr
ib
ed
.
Af
ter
th
at,
th
is
wo
r
k
d
escr
ib
e
s
th
e
d
etails
o
f
th
e
k
ey
s
tep
s
in
th
e
f
o
llo
win
g
s
u
b
s
ec
tio
n
s
.
T
h
e
m
eth
o
d
u
s
ed
to
ch
ar
ac
ter
ize
non
-
s
ib
ilan
t
Ar
ab
ic
f
r
icativ
e
c
o
n
s
o
n
an
ts
is
d
escr
ib
ed
as
f
o
llo
ws:
i
n
th
e
ca
s
e
o
f
a
c
o
n
s
o
n
a
n
t
–
v
o
wel
(
CV
)
ty
p
e
s
ig
n
al,
th
is
wo
r
k
h
as
d
ev
elo
p
ed
a
m
eth
o
d
to
an
aly
ze
th
e
en
er
g
y
d
is
tr
ib
u
tio
n
in
th
e
s
p
ee
ch
s
ig
n
al,
as illu
s
tr
ated
in
Fig
u
r
e
1
.
Ou
r
ap
p
r
o
ac
h
r
el
ies
o
n
th
e
id
e
n
tific
atio
n
o
f
tw
o
s
p
ec
if
ic
i
n
ter
v
als:
th
e
m
id
d
l
e
o
f
th
e
co
n
s
o
n
a
n
t
an
d
th
e
m
i
d
d
le
o
f
th
e
v
o
wel.
T
h
e
k
e
y
ac
o
u
s
tic
p
ar
am
ete
r
s
o
f
th
is
m
eth
o
d
i
n
clu
d
e
th
e
ac
o
u
s
tic
en
e
r
g
y
in
th
e
m
i
d
d
le
o
f
th
e
co
n
s
o
n
an
t,
d
en
o
te
d
b
y
th
e
ac
r
o
n
y
m
“M
C
i
”,
an
d
th
e
en
e
r
g
y
in
th
e
m
id
d
le
o
f
t
h
e
v
o
wel,
d
en
o
ted
b
y
“M
Vi”
.
Her
e
th
e
s
u
b
s
cr
ip
t
(
i)
s
ig
n
if
ies
th
e
r
elev
a
n
t
f
r
e
q
u
en
c
y
b
a
n
d
,
b
ec
au
s
e
th
is
wo
r
k
in
ten
d
s
to
e
x
am
in
e
th
e
en
er
g
y
d
is
tr
ib
u
tio
n
i
n
s
p
ec
if
ic
f
r
e
q
u
e
n
cy
b
an
d
s
.
T
h
e
p
r
ec
is
io
n
o
f
d
eter
m
in
in
g
th
ese
two
in
ter
v
als
is
clo
s
ely
lin
k
ed
to
th
e
id
en
tific
atio
n
o
f
th
e
m
o
m
en
t
d
elim
itin
g
th
e
tr
an
s
itio
n
b
etwe
en
th
e
co
n
s
o
n
an
t
an
d
th
e
v
o
wel.
On
ce
th
is
tim
e
p
o
in
t
is
id
en
tifie
d
,
th
is
w
o
r
k
s
elec
ts
th
e
s
ec
o
n
d
th
ir
d
o
f
th
e
s
ig
n
al
to
ch
ar
ac
ter
ize
th
e
co
n
s
o
n
an
t
in
ter
v
al
an
d
s
im
ilar
ly
d
eter
m
i
n
e
th
e
co
r
r
esp
o
n
d
i
n
g
in
ter
v
al
f
o
r
th
e
v
o
wel.
Fig
u
r
e
1
.
E
n
er
g
y
d
is
tr
ib
u
tio
n
f
o
r
a
C
V
-
ty
p
e
v
o
ca
l sig
n
al
T
o
d
o
th
is
,
th
is
wo
r
k
u
s
ed
th
e
f
o
llo
win
g
m
eth
o
d
o
lo
g
y
(
s
e
e
p
ip
elin
e
Fig
u
r
e
2
)
:
in
itially
,
th
is
wo
r
k
b
u
ilt
a
d
atab
ase,
i.e
.
a
co
r
p
u
s
o
f
C
V
-
ty
p
e
au
d
io
f
iles
.
T
h
e
p
h
o
n
em
es
in
th
is
d
atab
ase
wer
e
th
en
s
u
b
jecte
d
to
p
r
ep
r
o
ce
s
s
in
g
to
ex
tr
ac
t
ac
o
u
s
tic
f
ea
tu
r
es,
in
clu
d
in
g
en
e
r
g
y
i
n
th
e
b
an
d
s
.
T
h
e
n
,
we
p
r
o
ce
e
d
ed
to
th
e
s
elec
tio
n
o
f
th
e
v
ec
t
o
r
o
f
in
d
ices,
w
h
ich
co
n
s
titu
tes
th
e
in
p
u
t
o
f
th
e
class
if
icatio
n
s
y
s
tem
r
esp
o
n
s
ib
le
f
o
r
r
ec
o
g
n
izin
g
th
e
non
-
s
ib
ilan
t
f
r
icativ
e
co
n
s
o
n
an
t: v
o
iced
,
u
n
v
o
ice
d
,
an
d
p
h
ar
y
n
g
ea
l
.
Fig
u
r
e
2
.
Sp
ee
c
h
s
ig
n
al
class
if
icatio
n
p
r
o
ce
s
s
in
g
p
i
p
elin
e
2
.
1
.
Co
rpus
T
o
b
u
ild
o
u
r
o
wn
co
r
p
u
s
,
th
is
wo
r
k
in
v
ited
2
0
Mo
r
o
cc
a
n
s
p
ea
k
er
s
to
r
ep
ea
t
a
C
V
C
VC
V
-
ty
p
e
s
eq
u
en
ce
f
o
u
r
tim
es.
T
h
is
s
e
q
u
en
ce
is
illu
s
tr
ated
g
r
ap
h
ica
lly
in
Fig
u
r
e
3
.
T
h
ese
s
eq
u
e
n
ce
s
wer
e
r
ec
o
r
d
ed
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
15
,
No
.
5
,
Octo
b
er
20
26
:
4
8
7
8
-
489
2
4882
u
s
in
g
Pra
at
s
o
f
twar
e,
with
a
s
am
p
lin
g
f
r
e
q
u
en
c
y
o
f
2
2
.
0
5
k
Hz.
T
h
en
,
p
r
o
ce
ed
ed
t
o
s
eg
m
en
t
th
ese
s
eq
u
en
ce
s
in
o
r
d
e
r
to
is
o
late
th
e
C
V
u
n
it,
u
ltima
tely
r
esu
ltin
g
in
a
c
o
r
p
u
s
co
m
p
r
is
in
g
4
,
3
2
0
C
V
u
n
its
.
T
o
g
u
ar
an
tee
ab
s
o
lu
te
p
r
ec
is
io
n
an
d
av
o
i
d
co
ar
ticu
latio
n
n
o
is
e,
th
e
in
it
ial
s
eg
m
en
tatio
n
o
f
th
e
t
r
an
s
itio
n
s
b
etwe
en
th
e
co
n
s
o
n
an
t a
n
d
th
e
v
o
wel
was p
er
f
o
r
m
ed
m
a
n
u
ally
o
n
th
e
f
ir
s
t h
alf
o
f
th
e
d
ataset
u
s
in
g
Pra
at.
T
h
is
m
eticu
lo
u
s
m
an
u
al
a
n
n
o
tatio
n
s
er
v
ed
as
a
r
eliab
le
g
r
o
u
n
d
tr
u
th
to
s
tu
d
y
an
d
v
alid
ate
th
e
ac
o
u
s
tic
b
eh
av
i
o
r
o
f
la
n
d
m
ar
k
s
.
On
ce
th
e
m
an
u
al
b
o
u
n
d
ar
ies
wer
e
estab
lis
h
ed
,
th
is
wo
r
k
v
alid
ated
th
at
o
u
r
au
to
m
atic
d
etec
tio
n
,
b
ased
o
n
th
e
r
ate
o
f
c
h
an
g
e
(
R
OC
)
o
f
n
o
r
m
alize
d
en
e
r
g
y
,
p
er
f
ec
tly
m
atch
ed
th
e
m
a
n
u
al
r
ef
er
en
ce
s
.
C
o
n
s
eq
u
e
n
tly
,
th
e
s
ec
o
n
d
h
alf
o
f
th
e
co
r
p
u
s
was
s
eg
m
en
ted
au
to
m
atica
lly
,
p
r
o
v
id
i
n
g
a
h
ig
h
ly
r
eliab
le
an
d
r
o
b
u
s
t set o
f
4
,
3
2
0
s
am
p
les with
o
u
t r
eq
u
ir
in
g
m
u
ltip
le
h
u
m
an
an
n
o
tato
r
s
.
Fig
u
r
e
3
.
I
ll
u
s
tr
atio
n
o
f
a
C
VC
VC
V
-
ty
p
e
s
eq
u
en
ce
f
o
r
th
e
co
n
s
o
n
an
t
/
χ
/
f
o
llo
we
d
b
y
th
e
v
o
wel
/a
/
2
.
2
.
P
re
pro
ce
s
s
ing
I
n
itial
p
r
ep
r
o
ce
s
s
in
g
o
f
a
s
p
e
ec
h
s
ig
n
al
s
am
p
le
in
v
o
lv
es
t
h
e
ap
p
licatio
n
o
f
a
p
r
e
-
em
p
h
asis
f
ilter
,
aim
ed
at
im
p
r
o
v
in
g
s
p
ee
ch
q
u
ality
b
y
im
p
r
o
v
in
g
in
tellig
ib
i
lity
an
d
atten
u
atin
g
b
ac
k
g
r
o
u
n
d
n
o
is
e.
Giv
en
th
e
s
to
ch
asti
c
n
atu
r
e
o
f
th
e
s
p
ee
c
h
s
ig
n
al,
a
win
d
o
win
g
s
tep
f
o
llo
ws
to
d
ec
r
ea
s
e
v
ar
ian
c
e
e
f
f
ec
ts
an
d
im
p
r
o
v
e
tem
p
o
r
al
r
eso
lu
tio
n
,
u
s
in
g
a
Ham
m
in
g
win
d
o
w
wh
ich
h
as
a
Gau
s
s
ian
s
h
ap
e
in
th
e
tim
e
d
o
m
ain
.
Z
er
o
-
p
a
d
d
in
g
is
u
s
ed
to
im
p
r
o
v
e
f
r
eq
u
en
cy
r
eso
lu
tio
n
,
an
d
to
ad
d
r
ess
s
ig
n
al
d
is
co
n
tin
u
ities
a
win
d
o
w
o
v
er
lap
s
o
f
9
.
6
m
s
is
im
p
lem
en
ted
.
T
h
e
n
e
x
t
s
tep
co
n
s
is
ts
o
f
ca
lcu
latin
g
th
e
f
ast
Fo
u
r
ier
tr
an
s
f
o
r
m
(
FF
T
)
t
o
m
o
v
e
f
r
o
m
th
e
tim
e
d
o
m
ain
to
th
e
f
r
eq
u
e
n
cy
d
o
m
ain
,
t
h
u
s
f
ac
ilit
atin
g
th
e
g
en
e
r
atio
n
o
f
a
s
p
ec
tr
o
g
r
am
.
Su
b
s
eq
u
en
tly
,
th
is
s
p
ec
tr
o
g
r
a
m
is
u
s
ed
to
d
eter
m
in
e
th
e
en
e
r
g
y
in
s
p
ec
if
ic
f
r
eq
u
en
cy
b
an
d
s
.
2
.
3
.
E
nerg
y
c
a
lcula
t
io
n
T
h
e
en
e
r
g
y
p
er
f
r
e
q
u
en
c
y
b
an
d
,
d
en
o
te
d
“E
b
”,
was
ca
l
cu
lated
b
y
a
d
d
in
g
th
e
s
q
u
ar
es
o
f
th
e
am
p
litu
d
e
o
f
ea
ch
s
eg
m
e
n
t o
f
th
e
s
p
ec
tr
o
g
r
am
ac
co
r
d
in
g
to
(
1
)
.
(
)
=
∑
10
10
(
|
(
,
|
2
)
(
1
)
W
h
er
e
(
,
)
r
ep
r
esen
ts
th
e
am
p
litu
d
e
o
f
th
e
s
p
ec
tr
u
m
s
m
o
o
th
ed
u
s
in
g
a
2
0
-
p
o
in
t
m
o
v
in
g
a
v
er
ag
e
at
tim
e
in
d
ex
n
.
T
h
e
in
d
ex
r
ep
r
es
en
t
s
th
e
co
n
s
id
er
ed
f
r
eq
u
e
n
cy
b
a
n
d
.
T
h
e
f
r
eq
u
en
cy
in
d
ex
v
ar
ies
b
etwe
e
n
th
e
in
d
ices o
f
th
e
Fo
u
r
ier
tr
an
s
f
o
r
m
co
r
r
esp
o
n
d
in
g
to
th
e
l
o
wer
an
d
u
p
p
er
b
o
u
n
d
s
o
f
ea
c
h
b
a
n
d
.
T
h
e
ca
lcu
latio
n
o
f
th
e
lo
g
ar
i
th
m
o
f
th
e
e
n
er
g
y
in
ea
ch
f
r
eq
u
en
c
y
b
a
n
d
is
also
j
u
s
tifie
d
b
y
th
e
f
u
n
ctio
n
in
g
o
f
th
e
h
u
m
an
a
u
d
ito
r
y
s
y
s
tem
s
in
ce
we
d
o
n
o
t
p
er
ce
iv
e
s
o
u
n
d
in
ten
s
ity
o
n
a
lin
ea
r
s
ca
le.
T
o
m
itig
ate
th
e
in
f
lu
en
ce
o
f
s
p
e
ak
er
-
r
elate
d
v
ar
iatio
n
s
,
a
n
o
r
m
aliza
tio
n
s
tep
was
im
p
lem
e
n
ted
.
T
h
is
in
v
o
l
v
ed
d
iv
id
in
g
th
e
e
n
er
g
y
in
ea
c
h
b
a
n
d
,
r
ep
r
esen
te
d
b
y
“
b
”
,
b
y
th
e
to
tal
en
e
r
g
y
,
en
s
u
r
in
g
a
n
o
r
m
alize
d
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
E
xp
lo
r
in
g
A
r
a
b
ic
n
o
n
-
s
ib
ila
n
t fri
ca
tive
co
n
s
o
n
a
n
ts
:
a
co
mp
r
eh
en
s
ive
cla
s
s
ifica
tio
n
(
Yo
u
s
s
ef
E
lfa
h
m
)
4883
m
ea
s
u
r
em
en
t
f
o
r
a
n
aly
s
is
an
d
co
m
p
ar
is
o
n
.
Usi
n
g
th
e
n
o
tatio
n
s
“
PEb
”
to
d
en
o
te
th
e
p
er
c
en
tag
e
o
f
en
er
g
y
i
n
b
an
d
(
b
)
at
f
r
am
e
(
n
)
,
a
n
d
“
ET
”
to
r
ep
r
esen
t th
e
to
tal
en
er
g
y
ac
r
o
s
s
all
b
an
d
s
as sh
o
wn
(
2
)
.
(
)
=
100
.
(
)
(
)
)
(
2
)
2
.
4
.
L
o
ca
t
io
n o
f
c
o
ns
o
na
nts
T
h
is
s
tu
d
y
ap
p
r
o
ac
h
u
s
es
two
ac
o
u
s
tic
p
ar
am
eter
s
:
th
e
en
e
r
g
y
at
th
e
m
id
d
le
o
f
t
h
e
co
n
s
o
n
a
n
t
an
d
th
e
en
er
g
y
at
th
e
m
id
d
le
o
f
th
e
v
o
wel.
T
h
e
ef
f
ec
tiv
en
ess
o
f
th
is
m
eth
o
d
d
ep
en
d
s
o
n
t
h
e
tem
p
o
r
al
p
o
in
ty
d
elim
itin
g
th
e
tr
an
s
itio
n
b
etwe
en
th
e
co
n
s
o
n
a
n
t
an
d
t
h
e
v
o
wel.
T
o
id
e
n
tify
th
is
p
o
i
n
t,
we
ex
p
lo
ited
lan
d
m
ar
k
p
o
in
ts
,
wh
ich
ar
e
m
o
m
e
n
ts
r
elativ
e
to
ab
r
u
p
t
ch
a
n
g
es
in
th
e
ac
o
u
s
tic
s
p
ec
tr
u
m
d
u
e
to
ar
tic
u
lato
r
y
m
o
v
em
en
ts
ass
o
ciate
d
with
th
e
p
r
o
d
u
ctio
n
o
f
s
o
u
n
d
s
(
c
o
n
s
o
n
a
n
t/v
o
we
l)
.
I
n
t
h
is
s
tu
d
y
n
ee
d
th
e
(
g
)
lan
d
m
ar
k
,
wh
ich
ch
ar
ac
ter
izes
th
e
v
i
b
r
atio
n
o
f
th
e
v
o
ca
l
co
r
d
s
,
allo
win
g
u
s
to
lo
ca
te
th
e
b
eg
in
n
in
g
an
d
en
d
o
f
a
v
o
wel
p
r
ec
is
ely
.
T
o
d
eter
m
i
n
e
th
e
g
lan
d
m
ar
k
,
th
is
s
tu
d
y
ca
lcu
l
ated
th
e
R
OC
o
f
th
e
n
o
r
m
al
ized
en
er
g
y
i
n
th
e
B
1
b
an
d
[
1
0
0
-
4
0
0
Hz]
u
s
in
g
(
3
)
.
(
)
=
(
)
−
(
−
)
)
(
3
)
I
n
F
ig
u
r
e
4
,
esp
ec
ially
i
n
th
e
lo
wer
p
ar
t
illu
s
tr
atin
g
th
e
R
OC
,
two
ty
p
es
o
f
p
ea
k
s
(
o
n
e
p
o
s
itiv
e
an
d
o
n
e
n
eg
ativ
e)
ca
n
b
e
o
b
s
er
v
e
d
.
E
ac
h
lan
d
m
ar
k
p
r
esen
ts
two
p
ea
k
s
,
co
r
r
esp
o
n
d
in
g
to
th
e
b
e
g
in
n
in
g
a
n
d
en
d
o
f
an
ac
o
u
s
tic
p
h
en
o
m
en
o
n
.
T
h
e
ac
o
u
s
tic
lan
d
m
ar
k
(
g
)
co
r
r
esp
o
n
d
s
to
th
e
tim
es
wh
en
th
e
v
o
ca
l
co
r
d
s
b
eg
in
to
v
ib
r
ate
(
g
+
)
an
d
w
h
en
th
is
v
ib
r
atio
n
en
d
s
(
g
-
)
.
T
h
ese
tim
es
ar
e
id
en
tifie
d
b
y
th
e
cr
o
s
s
in
g
p
o
in
ts
o
f
th
e
R
OC
o
f
th
e
n
o
r
m
alize
d
en
e
r
g
y
cu
r
v
e
f
o
r
th
e
b
a
n
d
[
0
-
4
0
0
Hz]
with
th
e
th
r
esh
o
ld
s
o
f
9
d
B
(
g
+)
an
d
-
9
d
B
(
g
-
)
r
esp
ec
tiv
ely
.
Fig
u
r
e
4
.
L
a
n
d
m
ar
k
p
o
in
ts
d
et
ec
tio
n
p
r
o
ce
s
s
in
g
.
T
h
e
b
o
tto
m
cu
r
v
e
is
th
e
R
OC
in
en
er
g
y
w
ith
in
th
e
[0
-
4
0
0
Hz]
b
an
d
2
.
5
.
Cla
s
s
if
ica
t
io
n a
lg
o
rit
hm
s
I
n
th
e
class
if
icatio
n
p
h
ase,
t
h
is
s
tu
d
y
em
p
l
o
y
ed
two
class
if
icatio
n
m
o
d
els
to
ass
ess
th
e
p
e
r
f
o
r
m
a
n
ce
o
f
o
u
r
o
w
n
class
if
icatio
n
m
eth
o
d
:
ar
tific
ial
n
eu
r
al
n
etwo
r
k
s
(
ANN)
an
d
SVM.
Neu
r
al
n
etwo
r
k
s
,
ca
p
ab
le
o
f
s
u
p
er
v
is
ed
lear
n
in
g
,
co
n
s
is
t
o
f
co
m
p
u
tatio
n
al
u
n
its
ca
lled
n
eu
r
o
n
s
.
E
ac
h
n
eu
r
o
n
u
n
d
er
g
o
es
a
lin
ea
r
s
tep
,
s
u
m
m
in
g
weig
h
ted
i
n
p
u
ts
,
f
o
l
lo
wed
b
y
a
n
o
n
-
lin
ea
r
s
tep
with
an
ac
tiv
atio
n
f
u
n
ctio
n
lik
e
th
e
s
ig
m
o
id
.
T
h
e
ar
r
an
g
em
e
n
t
o
f
n
e
u
r
o
n
s
in
th
e
lay
er
s
o
f
th
e
ANN
is
f
lex
ib
le
d
u
r
in
g
d
esig
n
.
T
h
e
p
r
o
ce
s
s
o
f
u
s
in
g
an
ANN
in
v
o
lv
es
ch
o
o
s
in
g
th
e
n
u
m
b
e
r
o
f
lay
er
s
,
n
eu
r
o
n
s
p
er
lay
er
,
an
d
ac
tiv
atio
n
f
u
n
ctio
n
s
,
f
o
llo
wed
b
y
r
a
n
d
o
m
weig
h
t in
itializatio
n
,
tr
ain
in
g
t
o
ad
ju
s
t w
eig
h
ts
,
an
d
f
in
ally
,
t
h
e
p
r
ed
ictio
n
p
h
ase
[
3
8
]
–
[
4
0
]
.
As
f
o
r
th
e
SVM
class
if
ier
,
also
k
n
o
wn
as
a
lar
g
e
-
m
a
r
g
in
s
e
p
ar
ato
r
,
it
r
elies
o
n
a
s
u
p
er
v
is
ed
m
ac
h
in
e
lear
n
in
g
co
n
ce
p
t to
f
in
d
an
o
p
t
im
al
h
y
p
er
p
lan
e
f
o
r
s
ep
ar
atin
g
d
if
f
er
en
t c
lass
es.
T
h
e
SV
M
'
s
o
p
er
atio
n
in
clu
d
es
s
elec
tin
g
a
k
er
n
el
f
u
n
ctio
n
,
a
m
at
h
em
atica
l
tr
an
s
f
o
r
m
atio
n
m
a
k
in
g
d
ata
lin
ea
r
ly
s
ep
ar
ab
le
in
a
h
ig
h
-
d
im
e
n
s
io
n
al
s
p
ac
e.
Nex
t
,
th
e
SVM
is
tr
ain
ed
b
y
i
d
en
t
if
y
in
g
th
e
h
y
p
e
r
p
lan
e
t
h
at
m
a
x
im
izes
th
e
m
ar
g
in
b
etwe
en
class
es.
On
ce
tr
ain
ed
,
th
e
SVM
ca
n
b
e
u
s
ed
t
o
p
r
e
d
i
ct
n
ew
d
ata
p
o
i
n
ts
[
4
1
]
,
[
4
2
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
15
,
No
.
5
,
Octo
b
er
20
26
:
4
8
7
8
-
489
2
4884
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
I
n
o
u
r
p
r
e
v
io
u
s
s
tu
d
ies,
o
u
r
in
itial
in
v
esti
g
atio
n
f
o
cu
s
ed
o
n
t
h
e
ca
teg
o
r
izatio
n
o
f
f
r
icativ
es
b
ased
o
n
th
eir
m
o
d
e
o
f
ar
ticu
latio
n
,
s
p
ec
if
ically
wh
eth
e
r
th
e
y
a
r
e
s
ib
ilan
t
o
r
n
o
n
-
s
ib
ilan
t
.
W
e
o
b
s
er
v
e
d
t
h
at
th
e
m
ajo
r
ity
o
f
non
-
s
ib
ilan
t
co
n
s
o
n
an
ts
ar
e
ch
ar
ac
te
r
ized
b
y
lo
w
en
er
g
y
in
h
ig
h
f
r
e
q
u
en
cie
s
(
5
,
0
0
0
-
8
,
0
0
0
Hz
)
an
d
s
ig
n
if
ican
t
en
er
g
y
i
n
lo
w
f
r
eq
u
e
n
cies
(
8
0
0
-
2
,
0
0
0
H
z)
[
3
7
]
.
A
d
d
itio
n
ally
,
we
n
o
ted
a
s
u
b
s
tan
tial
d
if
f
er
en
ce
in
e
n
er
g
y
d
is
tr
ib
u
tio
n
f
o
r
t
h
e
s
am
e
c
o
n
s
o
n
a
n
t
f
o
llo
wed
b
y
th
r
ee
d
if
f
er
en
t
v
o
we
ls
.
T
o
m
itig
ate
t
h
e
v
o
wel'
s
ef
f
ec
t
o
n
th
e
co
n
s
o
n
a
n
t,
we
o
p
ted
to
co
n
d
u
ct
a
cla
s
s
if
icatio
n
with
in
g
r
o
u
p
s
o
f
c
o
n
s
o
n
an
ts
f
o
llo
wed
b
y
th
e
s
am
e
v
o
wel.
T
h
is
d
ec
is
io
n
led
u
s
to
u
n
d
er
ta
k
e
v
o
wel
class
if
icatio
n
,
wh
er
e
we
f
o
u
n
d
th
at
th
e
v
o
wel
/a
/
d
is
tin
g
u
is
h
es
its
elf
f
r
o
m
o
t
h
e
r
v
o
wels
with
a
s
ig
n
if
ica
n
t
am
o
u
n
t
o
f
e
n
er
g
y
in
th
e
b
an
d
(
6
0
0
-
1
,
8
0
0
Hz)
.
Similar
ly
,
th
e
v
o
wel
/i/
s
tan
d
s
o
u
t
f
r
o
m
th
e
v
o
wel
/u
/
d
u
e
t
o
n
o
tab
le
en
er
g
y
in
th
e
b
an
d
(
1
,
800
-
4
,
6
0
0
Hz)
.
I
n
th
is
s
tu
d
y
alg
o
r
ith
m
s
f
o
r
class
if
y
in
g
n
o
n
-
s
ib
ilan
t
f
r
icativ
es
o
p
er
ates
as
f
o
ll
o
ws:
with
in
a
g
i
v
en
v
o
wel
c
o
n
tex
t,
th
e
alg
o
r
ith
m
ca
te
g
o
r
izes
co
n
s
o
n
an
ts
in
to
two
g
r
o
u
p
s
:
v
o
i
ce
d
an
d
u
n
v
o
iced
.
Su
b
s
eq
u
e
n
tly
,
it
p
r
o
ce
e
d
s
to
class
if
y
th
e
co
n
s
o
n
an
ts
with
in
ea
ch
g
r
o
u
p
.
3
.
1
.
E
nerg
y
dis
t
ributio
n a
na
ly
s
is
o
f
no
n
-
s
ibi
la
nt
f
rica
t
iv
es
T
h
is
s
ec
tio
n
o
u
tlin
es
th
e
f
in
d
in
g
s
f
r
o
m
th
e
an
aly
s
is
o
f
n
o
r
m
alize
d
en
er
g
y
d
is
tr
ib
u
ti
o
n
with
in
non
-
s
ib
ilan
t
f
r
icativ
e
c
o
n
s
o
n
a
n
ts
in
th
e
Ar
ab
ic
lan
g
u
a
g
e
(
/f
,
χ
,
ɣ
,
ʕ
,
ћ
,
a
n
d
h/
)
f
o
llo
wed
b
y
o
n
e
o
f
th
e
th
r
ee
s
h
o
r
t
v
o
wels
(
/a
,
i
,
an
d
u/
)
.
T
h
is
s
tu
d
y
a
n
aly
s
is
is
b
ased
o
n
th
e
ex
am
i
n
atio
n
o
f
en
er
g
y
d
is
tr
ib
u
tio
n
with
in
s
p
ec
if
ic
f
r
eq
u
en
cy
b
a
n
d
s
.
Sp
ec
if
ically
,
th
is
s
tu
d
y
s
cr
u
tin
ized
th
e
f
o
llo
win
g
b
an
d
s
:
[
1
0
0
-
4
0
0
Hz]
,
[
4
0
0
-
6
0
0
Hz]
,
[
6
0
0
-
1
0
0
0
H
z]
,
[
4
0
0
-
1
,
0
0
0
Hz]
,
[
1
,
0
0
0
-
2
,
0
0
0
Hz]
,
[
1
,
0
0
0
-
2
,
6
0
0
Hz]
,
[
2
,
0
0
0
-
3
,
2
0
0
Hz]
,
[3
,
2
0
0
-
4
,
2
0
0
Hz]
,
[
4
,
2
0
0
-
4
,
6
0
0
Hz]
,
an
d
[
4
,
600
-
8
,
0
0
0
Hz]
.
An
an
aly
s
is
o
f
Fig
u
r
es
5
t
o
7
r
ev
ea
led
t
h
at
n
o
n
-
s
ib
ilan
t
c
o
n
s
o
n
an
ts
,
wh
eth
e
r
v
o
ice
d
o
r
v
o
icele
s
s
,
ex
h
ib
it
s
ig
n
if
ica
n
t
en
e
r
g
y
in
t
h
e
f
ir
s
t
f
r
eq
u
en
cy
b
an
d
[
1
0
0
-
4
0
0
Hz]
.
T
h
is
ch
a
r
ac
ter
is
tic
ca
n
b
e
ex
p
lain
ed
b
y
th
e
f
ac
t
th
at
th
ese
co
n
s
o
n
an
ts
ar
e
p
r
o
d
u
ce
d
in
th
e
p
o
s
ter
io
r
p
ar
t
o
f
th
e
o
r
al
ca
v
ity
,
p
r
o
m
o
t
in
g
th
e
em
er
g
en
c
e
o
f
f
o
r
m
an
t
s
tr
u
ctu
r
es
f
o
r
f
r
ic
ativ
e
co
n
s
o
n
a
n
ts
,
as
s
u
g
g
este
d
b
y
Stre
v
en
s
[
1
9
]
.
Ho
we
v
er
,
th
is
b
an
d
d
o
es
n
o
t
d
if
f
er
en
tiate
b
etwe
en
v
o
iced
an
d
v
o
icele
s
s
non
-
s
ib
ilan
t
co
n
s
o
n
an
ts
.
T
o
ch
ar
ac
te
r
ize
v
o
icin
g
i
n
th
ese
co
n
s
o
n
an
ts
,
th
is
s
tu
d
y
e
x
am
i
n
ed
th
e
s
ec
o
n
d
f
r
e
q
u
en
c
y
b
a
n
d
[
4
0
0
-
6
0
0
Hz]
.
Ou
r
o
b
s
er
v
atio
n
s
s
h
o
wed
th
at
v
o
iced
c
o
n
s
o
n
a
n
ts
/ɣ
/
an
d
/ʕ/
ex
h
ib
ited
e
n
er
g
y
ex
ce
e
d
in
g
1
0
%,
r
eg
ar
d
less
o
f
th
e
f
o
llo
win
g
v
o
wel.
I
n
co
n
tr
ast,
th
is
f
r
eq
u
en
c
y
b
a
n
d
c
o
n
tain
e
d
n
o
en
er
g
y
f
o
r
v
o
icele
s
s
co
n
s
o
n
an
ts
/f
,
χ
,
ћ
,
a
n
d
h/
.
Fu
r
t
h
er
m
o
r
e
,
th
is
s
tu
d
y
o
b
s
er
v
ed
a
d
i
f
f
er
en
ce
in
e
n
er
g
y
d
is
tr
ib
u
tio
n
in
f
r
e
q
u
en
c
y
b
an
d
s
f
o
r
th
e
s
am
e
co
n
s
o
n
an
t
f
o
llo
wed
b
y
o
n
e
o
f
th
e
th
r
ee
v
o
wels.
T
h
is
o
b
s
er
v
atio
n
in
d
icate
s
th
at
v
o
wels
h
a
v
e
a
s
ig
n
if
ican
t
ef
f
ec
t
o
n
non
-
s
ib
ilan
t
co
n
s
o
n
an
ts
,
wh
ich
ca
n
b
e
attr
ib
u
ted
to
co
ar
ticu
latio
n
.
T
h
e
f
o
llo
win
g
p
r
o
v
i
d
es
a
d
etailed
ac
o
u
s
tic
an
aly
s
is
o
f
th
e
d
is
tr
ib
u
tio
n
o
f
n
o
r
m
alize
d
en
er
g
y
in
th
e
m
id
d
le
o
f
all
n
o
n
-
s
ib
ilan
t
f
r
icativ
e
co
n
s
o
n
a
n
ts
in
th
e
m
en
tio
n
e
d
f
r
eq
u
e
n
cy
b
a
n
d
s
.
3
.
1
.
1
.
T
he
/ʕ
/
co
ns
o
na
n
t
Fo
r
th
e
/ʕ/
c
o
n
s
o
n
a
n
t
f
o
llo
wed
b
y
th
e
/a
/
v
o
wel,
th
e
an
a
ly
s
is
o
f
n
o
r
m
alize
d
e
n
er
g
y
d
is
tr
ib
u
tio
n
(
Fig
u
r
e
5
(a
)
)
r
e
v
ea
ls
th
at
n
ea
r
ly
3
0
%
o
f
th
e
en
er
g
y
is
co
n
c
en
tr
ated
in
th
e
[
1
0
0
-
4
0
0
Hz]
an
d
[
6
0
0
-
1
0
0
0
Hz]
b
an
d
s
,
with
less
th
an
1
0
%
in
th
e
[
4
0
0
-
6
0
0
Hz]
b
an
d
an
d
a
p
p
r
o
x
im
ately
2
0
%
in
th
e
[
1
,
0
0
0
-
2
,
6
0
0
Hz]
b
an
d
.
Ho
wev
er
,
wh
e
n
t
h
e
/ʕ/
c
o
n
s
o
n
an
t
is
f
o
llo
wed
b
y
th
e
/i/
v
o
wel,
th
e
e
n
er
g
y
d
is
tr
ib
u
tio
n
c
h
an
g
es,
with
ar
o
u
n
d
3
0
%
in
th
e
[
1
0
0
-
4
0
0
Hz]
b
a
n
d
,
o
v
er
2
0
%
in
th
e
[
4
0
0
-
6
0
0
Hz]
b
an
d
,
ab
o
u
t
1
0
%
in
th
e
[
6
0
0
-
1
,
0
0
0
Hz]
b
an
d
,
o
v
er
1
5
%
in
th
e
[
2
,
000
-
3
,
2
0
0
Hz]
b
an
d
,
an
d
ar
o
u
n
d
1
0
%
in
th
e
[
1
,
0
0
0
-
2
,
6
0
0
Hz]
b
an
d
.
Fo
r
th
e
/u
/
v
o
wel
,
th
er
e
is
a
m
ig
r
ati
o
n
o
f
en
e
r
g
y
f
r
o
m
h
i
g
h
to
lo
w
f
r
eq
u
e
n
c
ies,
r
esu
ltin
g
in
a
c
o
n
ce
n
t
r
atio
n
o
f
4
0
%
in
th
e
[
1
0
0
-
4
0
0
Hz]
b
a
n
d
,
o
v
er
3
0
%
in
th
e
[
4
0
0
-
6
0
0
Hz]
b
an
d
,
an
d
ab
o
u
t
1
5
% in
th
e
[
6
0
0
-
1
,
0
0
0
Hz]
b
an
d
.
3
.
1
.
2
.
T
he
/ɣ
/
co
ns
o
na
nt
B
ased
o
n
th
e
an
aly
s
is
o
f
n
o
r
m
alize
d
en
er
g
y
d
is
tr
ib
u
tio
n
g
r
a
p
h
s
f
o
r
t
h
e
/ɣ
/
co
n
s
o
n
an
t
f
o
llo
wed
b
y
th
e
/a
/
o
r
/i/
v
o
wels
(
Fig
u
r
e
5
(
b
)
)
,
it
was
o
b
s
er
v
ed
th
at
th
e
[
1
0
0
-
4
0
0
Hz]
f
r
e
q
u
en
c
y
b
an
d
r
ep
r
esen
ts
ap
p
r
o
x
im
ately
6
0
%
o
f
th
e
to
t
al
en
er
g
y
,
wh
ile
th
e
[
4
0
0
-
6
0
0
Hz]
b
an
d
co
n
tain
s
o
v
er
1
0
%
o
f
th
e
en
er
g
y
.
I
n
co
n
tr
ast,
th
e
[
6
0
0
-
1
,
0
0
0
Hz]
b
an
d
h
as
alm
o
s
t
n
o
en
er
g
y
,
le
s
s
th
an
5
%.
T
h
e
[
3
,
2
0
0
-
4
,
2
0
0
Hz]
b
an
d
c
o
n
tain
s
ab
o
u
t
1
0
%
o
f
th
e
to
tal
en
er
g
y
.
W
h
en
th
e
/ɣ
/
co
n
s
o
n
an
t
is
f
o
llo
wed
b
y
th
e
/u
/
v
o
wel,
th
e
en
er
g
y
d
is
tr
ib
u
tio
n
r
em
ain
s
th
e
s
am
e
f
o
r
th
e
[
1
0
0
-
4
0
0
Hz]
a
n
d
[
4
0
0
-
6
0
0
Hz]
b
an
d
s
.
Ho
wev
er
,
a
d
if
f
e
r
en
ce
was
o
b
s
er
v
ed
f
o
r
th
e
th
ir
d
b
an
d
[
6
0
0
-
1
,
0
0
0
Hz
]
,
wh
ich
co
n
tain
s
ap
p
r
o
x
im
ately
1
0
%
o
f
th
e
to
tal
en
er
g
y
.
T
h
e
o
t
h
er
b
an
d
s
h
av
e
less
th
an
5
% o
f
t
h
e
to
tal
en
er
g
y
.
3
.
1
.
3
.
T
he
/ћ/
co
ns
o
na
nt
Fo
r
th
e
/ћ
/
co
n
s
o
n
an
t
f
o
llo
w
ed
b
y
th
e
/a
/
v
o
wel,
en
er
g
y
d
is
tr
ib
u
tio
n
is
s
u
ch
th
at
ab
o
u
t
1
0
%
o
f
en
er
g
y
is
in
ea
c
h
o
f
th
e
[
1
0
0
-
4
0
0
Hz
]
,
[
6
0
0
-
1
,
0
0
0
Hz]
,
a
n
d
[
1
,
000
-
2
,
6
0
0
Hz]
f
r
e
q
u
en
c
y
b
an
d
s
,
with
a
to
tal
ab
s
en
ce
o
f
en
er
g
y
in
th
e
[
4
0
0
-
6
0
0
Hz]
b
a
n
d
.
T
h
e
[
1
,
0
0
0
-
2
,
6
0
0
Hz]
an
d
[
3
,
2
0
0
-
4
,
2
0
0
Hz]
b
an
d
s
ea
ch
co
n
tai
n
o
v
er
2
0
% o
f
en
e
r
g
y
(
Fig
u
r
e
6
(
a)
)
.
W
h
en
th
e
/ћ
/
c
o
n
s
o
n
a
n
t is f
o
llo
wed
b
y
th
e
/i/
v
o
wel,
a
p
p
r
o
x
im
ately
1
0
% o
f
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
E
xp
lo
r
in
g
A
r
a
b
ic
n
o
n
-
s
ib
ila
n
t fri
ca
tive
co
n
s
o
n
a
n
ts
:
a
co
mp
r
eh
en
s
ive
cla
s
s
ifica
tio
n
(
Yo
u
s
s
ef
E
lfa
h
m
)
4885
th
e
en
er
g
y
is
co
n
ce
n
t
r
ated
in
t
h
e
[
1
0
0
-
4
0
0
Hz]
b
an
d
,
wh
ile
th
e
[
2
,
0
0
0
-
3
,
2
0
0
Hz]
a
n
d
[
3
,
2
0
0
-
4
,
2
0
0
Hz]
b
an
d
s
ea
ch
co
n
tain
o
v
e
r
3
0
%
o
f
e
n
er
g
y
.
T
h
e
o
th
er
b
an
d
s
h
av
e
alm
o
s
t
n
o
e
n
er
g
y
.
W
h
en
th
e
/ћ
/
co
n
s
o
n
an
t
is
f
o
llo
wed
b
y
th
e
/u
/
v
o
wel,
th
e
[
1
0
0
-
4
0
0
Hz]
b
a
n
d
s
h
o
ws
a
co
n
ce
n
tr
atio
n
o
f
o
v
e
r
5
0
%
o
f
en
er
g
y
,
wh
ile
th
e
[1
,
0
0
0
-
2
,
6
0
0
Hz]
b
an
d
co
n
tain
s
ab
o
u
t
1
0
%
o
f
en
e
r
g
y
,
a
n
d
t
h
e
[
3
,
200
-
4
,
2
0
0
Hz]
b
an
d
a
r
o
u
n
d
2
0
%.
T
h
e
o
t
h
er
b
an
d
s
h
av
e
lo
w
en
er
g
y
.
(
a)
(
b
)
Fig
u
r
e
5
.
E
n
er
g
y
d
is
tr
ib
u
tio
n
in
th
e
m
id
d
le
o
f
v
o
iced
c
o
n
s
o
n
an
ts
: (
a)
/ʕ/
an
d
(
b
)
/ɣ
/
3
.
1
.
4
.
T
he
/χ
/
co
ns
o
na
nt
Af
ter
ex
am
in
in
g
t
h
e
en
er
g
y
d
i
s
tr
ib
u
tio
n
g
r
ap
h
s
o
f
th
e
/χ
/
c
o
n
s
o
n
an
t
f
o
llo
we
d
b
y
th
e
/a
/
o
r
/i/
v
o
wels
(
Fig
u
r
e
6
(
b
)
)
,
it
was
o
b
s
er
v
e
d
th
at
t
h
e
h
i
g
h
f
r
eq
u
en
cies
[
4
,
2
0
0
-
8
,
0
0
0
Hz]
c
o
n
tain
ab
o
u
t
2
0
%
o
f
th
e
t
o
tal
en
er
g
y
,
wh
ile
th
e
[
1
0
0
-
4
0
0
Hz]
an
d
[
3
,
2
0
0
-
4
,
2
0
0
Hz]
b
a
n
d
s
co
n
tain
a
b
o
u
t
3
0
%
o
f
th
e
en
er
g
y
.
T
h
e
o
th
er
b
an
d
s
h
av
e
alm
o
s
t
n
o
en
er
g
y
.
Ho
w
ev
er
,
w
h
e
n
t
h
e
/
u
/
v
o
w
el
f
o
ll
o
ws
t
h
e
/
χ
/
c
o
n
s
o
n
a
n
t
,
a
s
i
g
n
i
f
ic
a
n
t in
c
r
ea
s
e
in
en
er
g
y
i
n
t
h
e
[
1
0
0
-
4
0
0
Hz
]
b
an
d
w
as
n
o
t
e
d
,
r
e
p
r
es
en
ti
n
g
o
v
e
r
6
0
%
o
f
t
h
e
t
o
tal
e
n
e
r
g
y
.
T
h
e
[
4
0
0
-
1
,
0
0
0
H
z]
an
d
[
4
,
2
0
0
-
8
,
0
0
0
Hz
]
b
an
d
s
co
n
tain
ab
o
u
t 1
0
% o
f
en
e
r
g
y
ea
ch
,
wh
ile
th
e
o
t
h
er
b
a
n
d
s
h
av
e
n
o
en
e
r
g
y
.
3
.
1
.
5
.
T
he
/h/
co
ns
o
na
nt
T
h
e
e
n
e
r
g
y
d
i
s
t
r
i
b
u
t
i
o
n
a
n
a
l
y
s
is
o
f
t
h
e
/
h
/
c
o
n
s
o
n
a
n
t
f
o
l
l
o
w
e
d
b
y
t
h
e
/
a
/
,
/i/
,
o
r
/
u
/
v
o
w
e
ls
(
F
ig
u
r
e
7
(
a
)
)
s
h
o
w
s
a
s
t
r
o
n
g
c
o
n
c
e
n
t
r
a
ti
o
n
o
f
e
n
e
r
g
y
i
n
t
h
e
[
1
0
0
-
4
0
0
H
z
]
b
a
n
d
.
T
h
e
o
t
h
e
r
b
a
n
d
s
h
a
v
e
l
o
w
e
n
e
r
g
y
.
T
h
e
d
o
m
i
n
a
n
t
l
o
w
-
f
r
e
q
u
e
n
c
y
p
r
e
s
en
c
e
h
i
g
h
l
i
g
h
t
s
t
h
e
a
c
o
u
s
ti
c
t
r
ai
ts
o
f
/
h
/
,
m
a
i
n
l
y
s
h
a
p
e
d
b
y
g
l
o
t
ta
l
a
i
r
f
l
o
w
.
3
.
1
.
6
.
T
he
/f
/
co
ns
o
na
nt
Af
ter
ex
am
in
in
g
th
e
g
r
a
p
h
s
o
f
n
o
r
m
alize
d
e
n
er
g
y
d
is
tr
ib
u
ti
o
n
f
o
r
t
h
e
/f/
co
n
s
o
n
an
t
f
o
llo
wed
b
y
th
e
/a
/
,
/i/
,
o
r
/u
/
v
o
wels
(
Fig
u
r
e
7
(
b
)
)
,
it
was
o
b
s
er
v
ed
th
at
th
e
[
1
0
0
-
4
0
0
Hz]
b
a
n
d
co
n
tain
s
ab
o
u
t
3
0
%
o
f
th
e
to
tal
en
er
g
y
.
T
h
e
[
3
,
2
0
0
-
4
,
2
0
0
Hz]
an
d
[
4
,
2
0
0
-
8
,
0
0
0
Hz]
b
an
d
s
ea
ch
co
n
tain
ap
p
r
o
x
i
m
ately
2
0
%
o
f
th
e
en
er
g
y
,
wh
ile
t
h
e
o
t
h
er
b
an
d
s
h
av
e
lo
w
en
er
g
y
.
Ho
wev
er
,
if
th
e
/u
/
v
o
wel
f
o
llo
ws
th
e
co
n
s
o
n
an
t,
an
in
cr
ea
s
e
in
en
er
g
y
i
n
th
e
[
1
0
0
-
4
0
0
Hz]
b
an
d
was n
o
ted
,
r
ep
r
esen
tin
g
o
v
er
4
0
% o
f
th
e
to
tal
en
er
g
y
.
T
h
e
o
th
er
b
an
d
s
ar
e
s
im
ilar
to
th
o
s
e
o
f
th
e
/
a
/
an
d
/
i/
v
o
wels.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
15
,
No
.
5
,
Octo
b
er
20
26
:
4
8
7
8
-
489
2
4886
(
a)
(
b
)
Fig
u
r
e
6
.
E
n
er
g
y
d
is
tr
ib
u
tio
n
in
th
e
m
id
d
le
o
f
u
n
v
o
iced
co
n
s
o
n
an
ts
: (
a)
/ћ
/
a
n
d
(
b
)
/χ
/
(
a)
(
b
)
Fig
u
r
e
7
.
E
n
er
g
y
d
is
tr
ib
u
tio
n
in
th
e
m
id
d
le
o
f
u
n
v
o
iced
co
n
s
o
n
an
ts
: (
a)
/h
/
a
n
d
(
b
)
/f/
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
E
xp
lo
r
in
g
A
r
a
b
ic
n
o
n
-
s
ib
ila
n
t fri
ca
tive
co
n
s
o
n
a
n
ts
:
a
co
mp
r
eh
en
s
ive
cla
s
s
ifica
tio
n
(
Yo
u
s
s
ef
E
lfa
h
m
)
4887
3
.
2
.
No
n
-
s
ibi
la
nt
f
rica
t
iv
e
co
ns
o
na
nt
cla
s
s
if
ica
t
io
n a
lg
o
rit
hm
s
3
.
2
.
1
.
Cla
s
s
if
ica
t
io
n o
f
no
n
-
s
ibi
la
nt
f
rica
t
iv
e
co
ns
o
na
nts int
o
v
o
iced/un
v
o
iced
ca
t
eg
o
ri
es
B
ased
o
n
th
e
r
esu
lts
o
f
ac
o
u
s
tic
an
aly
s
is
co
n
ce
r
n
in
g
th
e
d
is
tr
ib
u
tio
n
o
f
n
o
r
m
alize
d
e
n
er
g
y
,
th
e
alg
o
r
ith
m
f
o
r
class
if
y
in
g
n
o
n
-
s
ib
ilan
t
f
r
icativ
e
co
n
s
o
n
a
n
ts
in
to
v
o
iced
a
n
d
u
n
v
o
ic
ed
ca
teg
o
r
ies
was
d
ev
elo
p
e
d
.
Fig
u
r
e
8
illu
s
tr
ates
th
is
alg
o
r
ith
m
,
f
u
n
ctio
n
in
g
as
f
o
llo
ws:
I
f
th
e
e
n
er
g
y
in
t
h
e
f
r
eq
u
en
c
y
b
a
n
d
B
2
(
4
0
0
-
6
0
0
Hz)
e
x
ce
ed
s
5
%,
t
h
e
co
n
s
o
n
an
t
is
id
e
n
tifie
d
as
v
o
iced
.
Oth
er
wis
e,
th
e
c
o
n
s
o
n
an
t
is
r
ec
o
g
n
ized
as u
n
v
o
iced
.
Fig
u
r
e
8
.
Alg
o
r
ith
m
f
o
r
class
if
y
in
g
n
o
n
-
s
ib
ilan
t
f
r
icativ
es
3
.
2
.
2
.
Cla
s
s
if
ica
t
io
n o
f
v
o
ice
d
no
n
-
s
ibi
la
nt
f
rica
t
iv
e
co
ns
o
na
nts
Aco
u
s
tic
an
aly
s
is
o
f
v
o
iced
n
o
n
-
s
ib
ilan
t
f
r
icativ
e
c
o
n
s
o
n
an
ts
,
f
o
llo
wed
b
y
t
h
e
v
o
wels
/a
/
o
r
/i/
,
led
to
th
e
d
ev
elo
p
m
en
t
o
f
alg
o
r
ith
m
s
p
r
esen
ted
in
Fig
u
r
e
9
.
Fo
r
non
-
s
ib
ilan
t
f
r
icativ
es
f
o
llo
we
d
b
y
th
e
v
o
wels
/a
/
o
r
/i/
(
Fig
u
r
e
9
(
a)
)
,
if
t
h
e
en
e
r
g
y
in
b
an
d
B
3
(
6
0
0
-
1
,
0
0
0
H
z)
ex
ce
ed
s
5
%,
th
e
co
n
s
o
n
an
t
is
clas
s
if
ied
as
/ʕ/.
Oth
er
wis
e,
it
is
id
en
tifie
d
as
/ɣ
/
.
Fo
r
v
o
ice
d
n
o
n
-
s
ib
ilan
t
f
r
icativ
es
ac
co
m
p
an
ied
b
y
th
e
v
o
wel
/u
/
(
Fig
u
r
e
9
(
b)
)
,
th
e
alg
o
r
ith
m
o
p
er
ates
as
f
o
llo
ws:
if
th
e
en
er
g
y
in
b
an
d
B
2
(
4
0
0
-
6
0
0
Hz)
ex
ce
ed
s
2
0
%,
th
e
co
n
s
o
n
an
t is ca
teg
o
r
ized
as
/ʕ/
.
Oth
er
wis
e,
th
e
co
n
s
o
n
a
n
t is id
en
tifie
d
as
/ɣ
/
.
(
a)
(
b
)
Fig
u
r
e
9
.
Alg
o
r
ith
m
f
o
r
n
o
n
-
s
i
b
ilan
t
f
r
icativ
e
co
n
s
o
n
an
ts
f
o
ll
o
wed
b
y
(
a)
v
o
wels
/a
/
an
d
/i/
an
d
(
b
)
f
o
llo
we
d
b
y
th
e
v
o
wel
/u
/
3
.
2
.
3
.
Cla
s
s
if
ica
t
io
n o
f
un
v
o
i
ce
d
no
n
-
s
ibi
la
nt
f
rica
t
iv
es f
o
llo
wed
by
t
he
vo
wel
/a
/
T
h
e
class
if
icatio
n
o
f
u
n
v
o
ice
d
n
o
n
-
s
ib
ilan
t
f
r
icativ
es
f
o
llo
wed
b
y
th
e
v
o
wel
/a
/
was
c
ar
r
ied
o
u
t
b
ased
o
n
ac
o
u
s
tic
an
aly
s
is
o
f
th
e
co
n
s
o
n
a
n
ts
/f
,
χ
,
ћ
,
a
n
d
h/
,
l
ea
d
in
g
to
th
e
alg
o
r
ith
m
p
r
esen
ted
in
Fig
u
r
e
1
0
:
i
f
th
e
en
er
g
y
in
b
a
n
d
B
6
(
2
,
0
0
0
-
3
,
2
0
0
Hz)
is
g
r
ea
ter
th
a
n
1
0
%,
th
e
c
o
n
s
o
n
an
t
is
id
en
tifie
d
a
s
/ћ
/
.
Oth
er
wis
e,
if
th
e
en
er
g
y
in
b
an
d
B
4
(
1
,
000
-
2
,
0
0
0
Hz)
is
g
r
ea
ter
th
an
5
%
,
th
e
co
n
s
o
n
an
t
is
class
if
ied
a
s
/χ
/
.
I
f
th
e
s
u
m
o
f
en
er
g
y
i
n
b
an
d
s
B
2
an
d
B
3
(
4
0
0
-
1
,
0
0
0
Hz)
is
less
th
an
5
%,
th
e
co
n
s
o
n
a
n
t
is
id
en
tifie
d
as
/f/
.
Oth
er
wis
e,
th
e
co
n
s
o
n
an
t is r
ec
o
g
n
ized
as
/h
/
.
3
.
2
.
4
.
Cla
s
s
if
ica
t
io
n o
f
un
v
o
i
ce
d
no
n
-
s
ibi
la
nt
f
rica
t
iv
es f
o
llo
wed
by
t
he
v
o
wel
s
/i/
a
nd
/
u/
F
o
r
t
h
e
c
o
n
s
o
n
a
n
t
s
/f
,
χ
,
ћ
,
a
n
d
h/
f
o
l
l
o
w
ed
b
y
th
e
v
o
we
l
s
/i/
a
n
d
/u
/
,
c
l
a
s
s
i
f
i
c
a
t
io
n
a
l
g
o
r
i
t
h
m
s
w
e
r
e
e
s
t
a
b
l
i
s
h
e
d
(
i
n
F
i
g
u
r
e
1
1
)
b
a
s
e
d
o
n
o
b
s
er
v
ed
a
co
u
s
t
ic
c
h
ar
a
c
t
er
i
s
t
i
c
s
.
T
h
e
a
l
g
o
r
i
th
m
f
o
r
t
h
e
v
o
w
e
l
/u
/
i
s
p
r
e
s
e
n
t
ed
in
F
i
g
u
r
e
1
1
(
a
)
a
n
d
o
p
er
a
t
e
s
a
c
co
r
d
i
n
g
t
o
th
e
f
o
l
lo
w
i
n
g
c
r
i
t
er
i
a
:
if
t
h
e
en
e
r
g
y
in
b
an
d
B
9
(4
,
6
0
0
-
8
,
0
0
0
H
z
)
i
s
le
s
s
t
h
a
n
5
%
,
th
e
co
n
s
o
n
a
n
t
i
s
i
d
en
t
i
f
i
e
d
a
s
/h
/
.
O
th
e
r
w
i
s
e
,
i
f
th
e
e
n
e
r
g
y
in
b
a
n
d
B
5
(1
,
0
0
0
-
2
,
6
0
0
H
z
)
i
s
g
r
e
a
t
er
t
h
a
n
5
%,
t
h
e
co
n
s
o
n
a
n
t
i
s
c
l
a
s
s
i
f
i
e
d
a
s
/
χ
/
.
I
f
t
h
e
en
e
r
g
y
i
n
b
a
n
d
B
9
i
s
l
e
s
s
t
h
a
n
1
0
%
,
t
h
e
c
o
n
s
o
n
an
t
i
s
id
e
n
t
if
i
e
d
a
s
/ћ
/
;
o
th
e
r
w
i
s
e
,
i
t
i
s
r
e
co
g
n
iz
e
d
a
s
/f/
.
C
o
n
v
er
s
e
l
y
,
F
ig
u
r
e
1
1
(
b
)
p
r
e
s
e
n
t
s
t
h
e
a
lg
o
r
i
th
m
f
o
r
t
h
e
v
o
w
e
l
/
i/
:
i
f
t
h
e
e
n
er
g
y
i
n
b
an
d
B
7
(
3
,
200
-
4
,
2
0
0
Hz
)
i
s
l
e
s
s
t
h
an
5
%
,
t
h
e
co
n
s
o
n
an
t
i
s
i
d
e
n
t
if
i
e
d
a
s
/
h
/
.
O
th
e
r
w
i
s
e,
if
t
h
e
en
e
r
g
y
i
n
b
a
n
d
B
6
(
2
,
0
0
0
-
3
,
2
0
0
H
z)
i
s
g
r
e
a
t
er
t
h
an
1
5
%
,
t
h
e
c
o
n
s
o
n
an
t
Evaluation Warning : The document was created with Spire.PDF for Python.