I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
p
ute
r
E
ng
in
ee
ring
(
I
J
E
CE
)
Vo
l.
7
,
No
.
5
,
Octo
b
e
r
2
0
1
7
,
p
p
.
2
51
4
~
2
5
1
9
I
SS
N:
2088
-
8708
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ec
e
.
v
7
i
5
.
pp
251
4
-
2
51
9
2514
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ia
e
s
jo
u
r
n
a
l.c
o
m/o
n
lin
e/in
d
ex
.
p
h
p
/I
JE
C
E
The Eff
ect
o
f
Ba
n
d
w
idth
o
n Spee
ch
Intelligibi
li
ty in
Alba
nia
n
La
ng
ua
g
e by Using
Multi
m
edia
A
pplica
tions
li
k
e
S
k
y
pe
and
Viber
Sa
brij
e
O
s
m
a
na
j
1
,
Alt
in Sh
a
la
2
,
B
lert
a
P
re
v
a
lla
3
1.
2
Un
iv
e
rsity
o
f
P
risti
n
a
,
De
p
a
rtm
e
n
t
o
f
El
e
c
tro
n
ics
,
P
risti
n
a
,
Ko
so
v
o
3
F
a
c
u
lt
y
o
f
Co
m
p
u
ter S
c
ien
c
e
,
AA
B
Un
iv
e
rsit
y
,
P
risti
n
a
,
K
o
so
v
o
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
u
l
15
,
2
0
1
6
R
ev
i
s
ed
Ma
y
1
7
,
2
0
1
7
A
cc
ep
ted
J
u
l
11
,
2
0
1
7
T
h
is
p
a
p
e
r
in
ten
d
s
to
a
n
a
ly
z
e
s
u
b
jec
ti
v
e
m
e
a
su
re
m
e
n
ts
o
f
in
telli
g
ib
il
it
y
o
f
sp
e
e
c
h
o
n
A
lb
a
n
ian
lan
g
u
a
g
e
d
u
rin
g
th
e
c
o
n
v
e
rsa
ti
o
n
b
e
tw
e
e
n
t
w
o
p
e
o
p
le
u
sin
g
a
p
p
li
c
a
ti
o
n
s
w
h
ich
to
d
a
y
a
re
v
e
r
y
u
se
d
f
o
r
c
o
m
m
u
n
ica
ti
o
n
su
c
h
a
s
S
k
y
p
e
a
n
d
V
i
b
e
r.
T
h
e
m
e
a
su
re
m
e
n
t
is
d
o
n
e
a
s
f
o
ll
o
w
s:
o
n
th
e
e
n
try
p
a
rt
o
f
th
e
tran
sm
issio
n
sy
ste
m
se
n
ten
c
e
s
o
r
w
o
rd
s
a
re
sp
o
k
e
n
o
r
j
u
st
sy
ll
a
b
les
w
h
il
e
o
n
re
c
e
iv
in
g
p
a
rt
is
re
c
o
rd
e
d
w
h
a
t
is
h
e
a
rd
;
th
e
p
e
rc
e
n
tag
e
o
f
w
o
rd
s,
se
n
ten
c
e
s
o
r
s
y
ll
a
b
les
c
o
rre
c
tl
y
r
e
c
e
iv
e
d
,
o
n
p
r
o
p
o
rti
o
n
to
t
h
o
se
i
m
p
o
se
d
o
n
th
e
e
n
try
o
f
th
e
s
y
ste
m
,
p
ro
v
id
in
g
th
e
p
e
rc
e
n
tag
e
o
f
in
telli
g
ib
il
it
y
(t
h
e
w
o
rd
s,
se
n
ten
c
e
s
o
r
s
y
ll
a
b
les
).
M
e
th
o
d
s
o
f
m
e
a
su
re
m
e
n
ts
a
r
e
m
a
d
e
a
t
d
iff
e
re
n
t
sp
e
e
d
o
f
th
e
In
tern
e
t
,
in
a
n
e
n
v
iro
n
m
e
n
t
w
it
h
o
u
t
n
o
ise
a
n
d
w
it
h
n
o
ise
,
in
o
rd
e
r
to
se
e
th
e
im
p
a
c
t
o
n
u
n
d
e
rsta
n
d
in
g
o
f
th
e
sp
e
e
c
h
w
it
h
d
if
fe
re
n
t
targ
e
t
p
a
ra
m
e
ters
.
K
ey
w
o
r
d
s
:
I
n
telli
g
ib
ilit
y
Me
asu
r
m
en
t
n
o
is
e
Mu
lti
m
ed
ia
ap
p
licatio
n
s
W
o
r
d
e
r
r
o
r
r
ate
Co
p
y
rig
h
t
©
2
0
1
7
In
stit
u
te o
f
A
d
v
a
n
c
e
d
E
n
g
i
n
e
e
rin
g
a
n
d
S
c
ien
c
e
.
Al
l
rig
h
ts
re
se
rv
e
d
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
1.
I
NT
RO
D
UCT
I
O
N
His
to
r
icall
y
,
s
p
ee
c
h
an
d
its
p
r
o
ce
s
s
in
g
is
h
a
n
d
led
in
d
if
f
er
en
t
w
a
y
s
in
co
m
p
u
ter
s
cien
ce
,
elec
tr
ical
en
g
i
n
ee
r
i
n
g
,
li
n
g
u
i
s
tics
,
a
n
d
p
s
y
ch
o
lo
g
y
.
T
h
e
f
ir
s
t
s
tep
s
o
f
th
e
d
ev
elo
p
m
e
n
t
o
f
t
h
ese
m
o
d
el
s
s
tar
ted
af
ter
Seco
n
d
W
o
r
ld
W
a
r
w
h
e
n
it
b
eg
an
t
h
e
d
is
co
v
er
y
o
f
co
m
p
u
t
er
,
s
o
,
in
th
e
p
er
io
d
f
r
o
m
t
h
e
1
9
4
0
s
u
n
til
th
e
la
te
1
9
5
0
s
in
ten
s
iv
e
l
y
w
as
w
o
r
k
ed
o
n
t
h
e
d
ev
elo
p
m
en
t
o
f
t
h
ese
m
o
d
els
an
d
s
p
ee
ch
in
t
ellig
ib
ili
t
y
.
Sp
ee
ch
r
ec
o
g
n
itio
n
,
as
an
id
ea
.
f
ir
s
tl
y
ap
p
ea
r
ed
s
e
v
er
al
d
ec
ad
es
a
g
o
at
s
cie
n
ti
f
ic
m
o
v
ies,
w
h
er
e
co
m
p
u
ter
s
r
ec
o
g
n
ized
s
p
ee
c
h
a
n
d
id
en
t
i
f
y
t
h
e
p
er
s
o
n
,
n
o
m
atter
h
o
w
f
a
s
t
a
n
d
w
h
at
la
n
g
u
a
g
e
h
e
s
p
o
k
e
[
1
]
.
B
u
t
ev
e
n
to
d
ay
,
in
r
ea
lit
y
it
is
n
o
t
m
a
n
ag
ed
to
d
esi
g
n
a
p
r
o
g
r
a
m
f
o
r
s
p
ee
ch
r
ec
o
g
n
itio
n
a
s
w
a
s
d
es
cr
ib
ed
in
s
cie
n
ti
f
ic
m
o
v
ie
s
.
Sk
y
p
e
a
n
d
Vib
er
ar
e
co
m
p
u
t
er
an
d
m
o
b
ile
ap
p
licatio
n
s
u
s
ed
as
a
te
s
ti
n
g
s
o
f
t
w
ar
e,
w
h
i
ch
p
r
o
v
id
e
co
m
m
u
n
icatio
n
t
h
r
o
u
g
h
s
p
e
ec
h
a
n
d
w
r
iti
n
g
.
S
k
y
p
e
a
n
d
Vib
er
u
s
e
Vo
I
P
s
ta
n
d
ar
d
th
at
m
ea
n
s
v
o
ice
co
m
m
u
n
icatio
n
th
r
o
u
g
h
t
h
e
I
n
ter
n
et
[
2
]
.
T
h
ese
ap
p
licatio
n
s
ar
e
in
c
lu
d
ed
i
n
t
h
e
s
o
-
ca
lle
d
w
eb
ap
p
licatio
n
s
th
at
f
u
n
ctio
n
o
n
l
y
b
y
h
a
v
i
n
g
a
cc
ess
to
I
n
ter
n
et.
St
u
d
ies
ab
o
u
t
th
e
i
n
telli
g
ib
ili
t
y
o
f
t
h
e
w
o
r
d
ca
n
b
e
p
er
f
o
r
m
ed
in
d
if
f
er
en
t
w
a
y
s
.
Sin
ce
t
h
e
in
v
e
n
tio
n
o
f
Alex
a
n
d
er
B
ell,
en
g
in
ee
r
s
an
d
s
cie
n
tis
ts
h
a
v
e
s
t
u
d
ied
th
e
p
h
e
n
o
m
en
o
n
o
f
s
p
ee
ch
co
m
m
u
n
icatio
n
a
s
a
co
m
m
u
n
icatio
n
b
et
w
ee
n
p
e
o
p
le
th
r
o
u
g
h
telec
o
m
m
u
n
icat
io
n
s
eq
u
ip
m
e
n
t
o
r
co
m
m
u
n
icatio
n
b
et
w
ee
n
m
a
n
an
d
a
m
ac
h
i
n
e.
Star
tin
g
f
r
o
m
th
e
6
0
'
s
d
ig
i
tal
s
i
g
n
al
p
r
o
ce
s
s
i
n
g
(
DSP
)
an
d
th
eir
p
r
esen
tatio
n
i
n
t
h
e
f
o
r
m
o
f
v
i
s
u
aliza
tio
n
an
d
m
at
h
e
m
atica
l
f
o
r
m
h
a
v
e
b
ee
n
s
o
m
e
w
h
at
ab
s
tr
ac
t
p
r
o
b
le
m
f
o
r
m
an
k
i
n
d
[
3
]
.
T
o
d
ay
,
m
o
d
er
n
tec
h
n
o
lo
g
y
h
as
n
o
t
y
e
t
b
ee
n
d
esi
g
n
ed
s
u
c
h
th
a
t
an
y
s
o
f
t
w
ar
e
o
n
t
h
e
d
atab
ase
h
av
e
p
lace
d
th
e
w
o
r
d
s
in
A
lb
a
n
ia
n
l
an
g
u
a
g
e,
ex
ce
p
t f
o
r
s
o
m
e
p
r
o
g
r
a
m
s
th
at
w
o
r
k
lo
ca
ll
y
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
E
C
E
I
SS
N:
2088
-
8708
Th
e
E
ffect
o
f B
a
n
d
w
id
th
o
n
S
p
ee
ch
I
n
tellig
ib
ilit
y
in
A
lb
a
n
ia
n
La
n
g
u
a
g
e
…
(
S
a
b
r
ije
Osma
n
a
j
)
2515
2.
RE
S
E
ARCH
M
E
T
H
O
D
2
.
1
.
T
he
ef
f
ec
t
o
f
I
nte
rnet
s
peed
o
n sp
ee
ch
inte
ll
ig
ibi
lity
by
u
s
ing
a
pp
lica
t
io
ns
t
ha
t
w
o
rk
w
it
h t
he
Vo
I
P
pla
t
f
o
r
m
.
So
m
e
i
n
ter
n
e
t
s
er
v
ice
p
r
o
v
id
e
r
s
o
f
f
er
I
n
ter
n
et
co
n
n
ec
tio
n
with
w
ir
ele
s
s
r
o
u
ter
s
,
s
o
,
i
n
t
h
i
s
c
ase
t
h
e
co
n
v
er
s
atio
n
s
ca
n
b
e
d
o
n
e
i
n
v
er
y
p
r
ac
tical,
co
n
v
e
n
ien
t
a
n
d
co
m
f
o
r
tab
le
w
a
y
u
s
i
n
g
S
k
y
p
e
a
n
d
Vib
er
.
I
t
s
h
o
u
ld
b
e
ta
k
en
i
n
to
co
n
s
id
e
r
atio
n
t
h
e
f
ac
t
th
at
i
n
ter
n
e
t
c
o
n
n
ec
tio
n
w
i
th
ca
b
le
co
n
n
ec
t
io
n
m
a
y
b
e
m
o
r
e
ef
f
icien
t
f
o
r
th
e
s
tab
ilit
y
o
f
t
h
e
co
n
v
er
s
atio
n
b
et
w
ee
n
p
eo
p
le
b
ec
au
s
e
th
e
s
p
ee
d
o
f
th
e
in
t
er
n
et
is
co
n
s
ta
n
t
a
n
d
at
w
ir
eles
s
co
n
n
ec
tio
n
s
it
m
a
y
v
ar
y
[
3
]
,
[
4
]
.
A
l
s
o
,
th
e
w
ir
ele
s
s
co
n
n
ec
t
io
n
m
a
y
h
a
v
e
a
n
i
m
p
ac
t
o
n
i
n
s
tab
ili
t
y
o
f
th
e
co
m
m
u
n
ica
tio
n
s
i
n
ce
w
ir
ele
s
s
co
n
n
ec
t
io
n
s
ar
e
o
f
f
e
r
ed
f
o
r
ac
ce
s
s
b
y
m
a
n
y
u
s
er
s
an
d
it
a
f
f
ec
ts
th
e
s
p
ee
d
o
f
t
h
e
in
ter
n
et,
t
h
e
n
at
co
n
n
ec
tio
n
s
w
it
h
o
u
t
w
ir
es
an
i
m
p
ac
t
ca
n
h
a
v
e
e
f
f
ec
ts
s
u
c
h
as
i
n
ter
f
er
e
n
ce
[
5
]
,
o
b
s
tacle
s
th
at
ar
is
e
d
u
e
to
th
e
f
r
eq
u
en
c
y
b
an
d
s
a
n
d
r
ef
lectio
n
o
f
w
a
v
es .
An
o
th
er
e
f
f
ec
t o
n
u
s
i
n
g
Vo
I
P
is
t
h
e
lo
s
s
o
f
p
ac
k
a
g
es a
n
d
as
a
r
esu
lt
is
lo
s
t a
p
ar
t o
f
t
h
e
co
n
v
er
s
atio
n
,
m
o
r
eo
v
er
an
i
n
tell
ig
ib
il
it
y
o
f
th
e
s
p
ee
c
h
is
lo
s
t
[
2
]
,
[
3
]
.
L
o
s
s
o
f
p
ac
k
e
ts
d
u
e
to
lo
ad
s
h
ed
d
in
g
a
n
d
as
a
co
n
s
eq
u
e
n
ce
th
e
p
ac
k
s
w
it
h
a
u
d
io
d
ata
m
a
y
r
e
m
a
in
o
n
t
h
e
n
et
w
o
r
k
lo
n
g
er
th
a
n
is
a
s
s
i
g
n
ed
to
a
f
r
am
e,
an
d
i
f
th
e
ti
m
e
ap
p
o
in
ted
is
p
a
s
s
ed
th
en
t
h
e
p
ac
k
a
g
e
i
s
lo
s
t,
w
h
i
ch
m
ea
n
s
t
h
at
th
er
e
is
n
o
d
e
s
tin
a
tio
n
u
n
til
th
at
m
o
m
e
n
t.
I
n
ter
n
et
s
er
v
ice
s
ar
e
d
iv
id
ed
in
to
class
es
ac
co
r
d
in
g
to
q
u
alit
y
o
f
f
er
ed
,
ie
th
e
b
est
o
f
th
e
clas
s
s
tar
ts
f
r
o
m
A
,
B
,
C
,
u
p
at
th
e
lo
w
es
t
q
u
alit
y
u
n
d
er
D
[
3
]
.
J
itter
(
v
ib
r
atio
n
)
-
r
ep
r
esen
t
s
s
tan
d
ar
d
d
ev
iatio
n
s
b
et
w
ee
n
p
a
ck
ets
o
r
f
r
a
m
es
o
f
d
ata.
T
h
e
s
ize
o
f
t
h
e
j
itter
r
ep
r
esen
ts
t
h
e
ev
a
lu
at
io
n
o
f
th
e
r
ip
p
le
d
ata.
W
h
en
tr
an
s
m
i
s
s
io
n
is
i
m
p
le
m
e
n
ted
w
it
h
o
p
tical
f
ib
er
o
r
o
th
er
m
ed
ia
p
ac
k
s
it
h
as
a
to
l
er
an
ce
f
o
r
a
d
ela
y
i
n
t
h
e
a
m
o
u
n
t
5
μ
s
/
k
m
[
6
]
.
I
T
U
r
ec
o
m
m
en
d
s
th
a
t
d
ela
y
s
o
f
th
e
j
itter
s
h
o
u
ld
n
o
t
b
e
g
r
ea
ter
th
an
1
5
0
m
s
f
o
r
m
o
s
t
ap
p
lica
tio
n
s
an
d
a
li
m
itatio
n
f
o
r
ap
p
li
ca
tio
n
s
w
it
h
v
o
ice
co
m
m
u
n
icatio
n
ab
o
u
t 4
0
0
m
S
[
2
]
,
[
6
]
.
T
h
e
to
tal
d
elay
in
t
h
e
s
y
s
te
m
i
s
co
n
s
i
s
ted
o
f
th
e
f
o
llo
w
i
n
g
co
m
p
o
n
en
ts
:
1
.
T
h
e
d
elay
i
n
t
h
e
p
r
o
ce
s
s
o
f
co
d
in
g
2
.
T
h
e
d
elay
d
u
e
to
w
ait
in
g
3
.
T
h
e
d
elay
o
f
t
h
e
tr
a
n
s
m
is
s
i
o
n
4
.
T
h
e
d
elay
d
u
e
to
t
h
e
d
ela
y
v
ar
iatio
n
a
n
d
i
m
p
r
o
v
e
m
en
t
s
o
f
th
e
d
ela
y
v
ar
iatio
n
s
5
.
T
h
e
d
elay
i
n
t
h
e
p
r
o
ce
s
s
o
f
d
ec
o
d
in
g
3.
RE
SU
L
T
S
A
ND
AN
AL
Y
SI
S
C
alcu
latio
n
o
f
W
o
r
d
E
r
r
o
r
R
ate
-
m
ea
s
u
r
e
m
en
t
p
ar
a
m
eter
f
o
r
m
ea
s
u
r
in
g
i
n
telli
g
ib
ili
t
y
o
f
s
p
ee
ch
i
s
th
e
w
o
r
d
er
r
o
r
r
ate.
Fo
r
s
o
m
e
s
i
m
p
le
r
ec
o
g
n
itio
n
s
y
s
te
m
s
(
s
u
c
h
as
f
o
r
e
x
a
m
p
le
th
e
is
o
lated
w
o
r
d
s
)
,
th
e
p
er
f
o
r
m
a
n
ce
is
s
i
m
p
l
y
th
e
p
er
ce
n
tag
e
o
f
lo
s
t
w
o
r
d
s
to
to
tal
w
o
r
d
s
.
Ho
w
e
v
er
,
th
i
s
m
ea
s
u
r
e
m
en
t
p
ar
a
m
eter
is
n
o
t
ef
f
ec
ti
v
e
b
ec
au
s
e
t
h
e
k
n
o
w
n
w
o
r
d
s
s
eq
u
e
n
ce
s
ca
n
co
n
t
ain
u
p
to
th
r
ee
t
y
p
e
s
o
f
er
r
o
r
s
[
9
]
.
Sim
ilar
ly
w
it
h
th
e
er
r
o
r
of
r
ec
o
g
n
itio
n
o
f
d
i
g
its
,
t
h
e
f
ir
s
t
er
r
o
r
k
n
o
w
n
as
r
ep
lace
m
e
n
t
o
f
w
o
r
d
s
,
o
cc
u
r
s
w
h
en
a
n
in
co
r
r
ec
t
w
o
r
d
is
ac
ce
p
ted
as
a
co
r
r
ec
t
w
o
r
d
.
T
h
e
s
ec
o
n
d
er
r
o
r
,
k
n
o
w
n
as
s
u
p
p
r
ess
io
n
o
f
w
o
r
d
s
,
o
c
cu
r
s
w
h
en
a
s
p
o
k
e
n
w
o
r
d
is
n
o
t
k
n
o
w
n
(
ie,
s
e
n
te
n
ce
s
h
a
v
e
n
o
t
r
ec
o
g
n
ized
t
h
e
s
p
o
k
en
w
o
r
d
at
t
h
e
e
n
tr
an
ce
)
[
7
]
.
A
n
d
t
h
e
th
ir
d
er
r
o
r
,
k
n
o
w
n
a
s
t
h
e
i
n
tr
o
d
u
c
tio
n
o
f
w
o
r
d
s
w
h
ile
p
r
o
ce
s
s
i
n
g
th
is
ca
s
e
h
ap
p
en
s
w
h
e
n
w
o
r
d
s
i
n
v
o
lv
ed
ar
e
ac
ce
p
ted
b
y
th
e
ir
k
n
o
w
led
g
e
(
ie,
th
e
s
e
n
ten
ce
is
r
ec
o
g
n
ized
an
d
ac
ce
p
ted
w
it
h
m
o
r
e
w
o
r
d
s
th
an
i
s
p
r
o
v
id
ed
in
en
tr
an
ce
s
u
c
h
as
n
o
is
e)
.
On
e
s
u
ch
e
x
a
m
p
le
w
o
u
ld
b
e:
Sp
o
k
en
s
e
n
te
n
ce
at
th
e
e
n
tr
an
ce
: G
o
o
d
ev
en
in
g
,
is
t
h
er
e
an
y
th
i
n
g
n
e
w
f
r
o
m
y
o
u
?
Sen
te
n
ce
u
n
d
er
s
to
o
d
an
d
ac
ce
p
ted
at
th
e
ex
it:
Go
o
d
ev
en
in
g
as
m
u
c
h
as
y
o
u
ar
e,
th
er
e
is
s
o
m
e
th
i
n
g
f
r
o
m
y
o
u
!
T
h
e
er
r
o
r
r
ate
is
d
ef
in
ed
as
th
e
p
er
ce
n
tag
e
o
f
th
e
w
o
r
d
s
in
c
o
r
r
ec
tly
ac
ce
p
ted
to
th
e
n
u
m
b
er
o
f
w
o
r
d
s
u
tter
ed
at
th
e
e
n
tr
an
ce
[
1
0
]
.
=
100%
∗
(
+
+
|
|
)
(
1
)
Su
b
s
ti
tu
t
io
n
s
-
R
ep
lace
m
e
n
t o
f
w
o
r
d
s
Dele
tio
n
s
-
T
er
m
i
n
atio
n
o
f
w
o
r
d
s
I
n
s
er
tio
n
s
–
I
n
s
er
tio
n
o
f
w
o
r
d
s
W
–
T
o
tal
n
u
m
b
er
o
f
w
o
r
d
s
=
100%
∗
(
)
(
2
)
T
o
u
n
d
er
s
tan
d
m
ea
s
u
r
e
m
e
n
t
s
b
etter
,
w
e
w
ill
g
iv
e
a
n
ex
a
m
p
l
e
o
f
a
co
n
v
er
s
atio
n
b
et
w
ee
n
a
tr
an
s
m
itter
a
n
d
a
r
ec
ip
ien
t u
s
in
g
Vib
er
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
J
E
C
E
Vo
l.
7
,
No
.
5
,
Octo
b
er
201
7
:
2
5
1
4
–
2
5
1
9
2516
-
A
n
e
x
a
m
p
le
o
f
h
o
w
ca
lc
u
lat
ed
W
E
R
,
b
etw
ee
n
t
h
e
s
p
ea
k
er
s
an
d
r
ec
eiv
er
u
s
in
g
Vib
er
.
R
E
F: i
*
*
*
*
*
UM
t
h
e
P
HONE
I
S
i L
E
FT
T
HE
p
o
r
ta
b
le
*
*
*
*
P
HONE
UP
ST
A
I
R
S
last
n
ig
h
t
HYP
: i
GOT
I
T
T
O
th
e
*
*
*
*
*
FU
L
L
E
ST
i L
OVE
T
O
p
o
r
tab
le
FOR
M
OF
ST
OR
E
S
last
n
ig
h
t
E
v
al:
I
I
S
D
S
S
S
I
S
S
B
y
ap
p
l
y
i
n
g
th
e
ab
o
v
e
e
x
p
r
ess
io
n
W
o
r
d
e
r
r
o
r
r
ate
(
W
E
R
)
=
100
6
+
3
+
1
13
=7
6
.
9
%
W
E
R
=7
6
.
9
%
[
7
]
,
[
8
]
Whil
e
inte
ll
ig
ibi
lity
is
co
un
t
e
d a
s
:
T
h
e
s
ca
le
o
f
in
telli
g
ib
ili
t
y
o
f
t
h
e
w
o
r
d
s
i
s
d
ef
i
n
ed
as th
e
p
er
ce
n
tag
e
o
f
co
r
r
ec
tl
y
r
ec
o
g
n
ize
d
w
o
r
d
s
to
th
e
n
u
m
b
er
o
f
w
o
r
d
s
u
tter
ed
at
th
e
en
tr
an
ce
.
Sp
ee
ch
u
n
d
er
s
tan
d
ab
ilit
y
=
∗
%
(
3
)
T
h
e
lev
el
o
f
s
ati
s
f
ac
tio
n
is
e
x
p
r
ess
ed
in
p
er
ce
n
ta
g
e
an
d
u
n
d
e
r
s
tan
d
in
g
h
as se
v
er
al
d
iv
i
s
io
n
s
ac
co
r
d
in
g
to
th
e
r
esu
lt
s
is
s
u
ed
[
1
]
.
T
ab
le
1
.
T
h
e
Scale
o
f
I
n
telli
g
i
b
ilit
y
o
f
Sp
ee
c
h
P
er
f
o
r
m
an
ce
Pe
r
f
o
r
m
a
n
c
e
l
e
v
e
l
Ex
c
e
l
l
e
n
t
G
o
o
d
En
o
u
g
h
W
e
a
k
In
t
e
l
l
i
g
i
b
i
l
i
t
y
7
6
-
1
0
0
6
6
-
75
6
1
-
65
3
0
–
60
T
h
e
d
er
iv
ed
r
esu
lts
ar
e
f
r
o
m
a
d
ialo
g
u
e
b
et
w
ee
n
t
w
o
m
al
e
p
er
s
o
n
s
.
So
at
th
e
tr
a
n
s
m
it
ter
,
at
th
e
en
tr
an
ce
o
f
th
e
s
y
s
te
m
th
e
te
x
t
i
s
r
ea
d
b
y
Dar
d
an
Me
h
m
et
i
ag
e
2
5
w
it
h
r
ea
d
ab
le
an
d
cl
ea
r
to
n
e,
w
h
ile
th
e
r
ec
eiv
er
o
r
at
th
e
o
u
tp
u
t
is
3
0
y
ea
r
s
o
ld
Alti
n
Sh
a
la
w
h
o
also
ass
es
s
ed
th
e
r
es
u
lts
o
f
th
ese
m
ea
s
u
r
e
m
e
n
t
s
.
R
eg
ar
d
in
g
t
h
e
m
e
th
o
d
s
a
n
d
m
an
n
er
o
f
m
ea
s
u
r
e
m
en
t
s
,
is
s
elec
ted
th
e
w
a
y
to
tes
t
d
if
f
er
e
n
t
te
x
ts
to
v
ie
w
an
d
co
m
p
ar
e
t
h
e
r
es
u
l
t
s
o
f
i
n
telli
g
ib
ilit
y
o
f
s
p
ee
ch
i
n
Alb
an
ia
n
l
an
g
u
a
g
e
d
ep
en
d
in
g
o
n
t
h
e
te
x
ts
t
h
at
ar
e
g
o
in
g
to
b
e
r
ea
d
f
o
r
test
in
g
a
n
d
i
n
o
th
er
lan
g
u
a
g
es
w
o
r
ld
w
id
e
an
d
d
if
f
er
en
t
m
o
d
els.
An
o
th
er
ac
h
iev
e
m
en
t if
w
e
s
ee
t
h
e
r
esu
lt
s
is
t
h
at
A
lb
a
n
ian
la
n
g
u
ag
e
i
n
ter
m
s
o
f
i
n
tell
ig
ib
il
it
y
an
d
ad
ap
tiv
it
y
f
o
r
co
m
m
u
n
ic
atio
n
,
ais
s
i
m
ilar
to
Ser
b
ian
an
d
C
r
o
atia
n
[
4
]
.
I
n
th
e
tab
le
2
is
p
r
esen
ted
a
W
E
R
s
tatis
tic
f
o
r
co
n
v
er
s
tat
io
n
d
o
n
e
b
y
d
i
f
f
er
e
n
t
ag
e.
A
ll
t
h
is
is
te
s
ted
in
E
n
g
li
s
h
la
n
g
u
a
g
e
b
y
r
ec
o
r
d
in
g
a
u
d
io
f
ile
s
an
d
th
e
n
te
s
ted
b
y
t
h
e
s
p
ea
k
er
s
.
T
ab
le
2
.
W
E
R
f
o
r
C
o
m
m
u
n
ica
tio
n
On
l
y
Me
n
an
d
O
n
l
y
W
o
m
en
Sto
o
d
A
s
id
e
a
n
d
Old
er
W
o
r
d
E
r
r
o
r
R
a
t
e
(
W
E
R
)
%
Y
o
u
n
g
Ol
d
D
i
f
f
e
r
e
n
c
e
O
v
e
r
a
l
l
3
0
.
4
4
0
.
4
1
0
.
0
M
a
l
e
s
3
0
.
1
3
8
.
8
8
.
7
F
e
mal
e
s
3
2
.
4
4
6
.
1
1
3
.
7
W
h
ile
w
e
ar
e
at
th
e
tex
ts
u
s
ed
f
o
r
r
ea
d
in
g
an
d
i
n
ter
p
etatio
n
it
’
s
m
u
ch
ea
s
ier
to
u
s
e
eq
u
i
v
ale
n
ce
o
f
titl
e
s
w
it
h
th
e
f
o
llo
w
i
n
g
ab
b
r
ev
iatio
n
s
:
T
ab
le
3
.
T
y
p
es o
f
T
ex
t Read
F1
Th
e
text
r
ec
o
g
n
iz
ed
b
y
th
e
r
ec
eive
r
F2
Un
r
ec
o
g
n
iz
ed
text
fo
r
th
e
r
ec
eive
r
F3
P
r
o
n
u
n
cia
tio
n
o
f 5
0
co
n
s
o
n
a
n
ts
w
ith
t
w
o
letter w
o
r
d
s
F4
P
r
o
n
u
n
cia
tio
n
o
f th
e
1
5
0
mo
s
t freq
u
en
tly
u
s
ed
w
o
r
d
s
in
A
lb
a
n
ia
n
la
n
g
u
a
g
e
F5
P
r
o
n
u
n
cia
tio
n
o
f 5
0
lo
n
g
er w
o
r
d
s
th
a
t a
r
e
less
u
s
ed
in
A
lb
a
n
ia
n
l
a
n
g
u
a
g
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
E
C
E
I
SS
N:
2088
-
8708
Th
e
E
ffect
o
f B
a
n
d
w
id
th
o
n
S
p
ee
ch
I
n
tellig
ib
ilit
y
in
A
lb
a
n
ia
n
La
n
g
u
a
g
e
…
(
S
a
b
r
ije
Osma
n
a
j
)
2517
M
e
a
s
ure
m
e
nts
do
ne
w
it
h S
ky
pe
–
Sp
ec
if
icatio
n
s
of
q
u
alit
y
o
f
th
e
i
n
ter
n
et
f
o
r
m
ea
s
u
r
m
en
ts
d
o
n
e
w
ith
t
h
e
Sk
y
p
e
ap
p
licatio
n
D
o
w
n
lo
ad
s
p
ee
d
-
1
0
0
Mb
p
s
U
p
lo
ad
s
p
ee
d
-
5
0
M
b
p
s
P
in
g
-
3
0
m
s
T
h
e
P
in
g
b
et
w
ee
n
th
e
S
k
y
p
e
s
er
v
er
an
d
th
e
P
C
w
e
h
av
e
m
a
d
e
th
e
m
ea
s
u
r
e
m
en
t
s
is
P
in
g
-
6
6
m
s
J
itter
-
2
1
m
s
L
i
n
e
o
f
s
i
te
q
u
alit
y
is
B
MO
S
-
4
:2
9
T
h
e
d
is
tan
ce
m
ea
s
u
r
ed
f
r
o
m
th
e
lo
ca
l
h
o
s
t
o
f
th
e
I
n
ter
n
et
p
r
o
v
id
er
Star
L
in
k
s
er
v
er
in
th
e
A
r
t
Mo
tio
n
P
r
is
tin
a.
As
w
e
i
n
d
icate
d
ab
o
v
e
f
o
r
t
h
e
u
s
e
o
f
te
x
t
ab
b
r
ev
iatio
n
s
,
at
t
h
e
b
eg
i
n
n
in
g
i
s
m
ea
s
u
r
ed
th
e
tex
t
f
o
r
F1
w
it
h
t
h
e
title
: a
m
p
li
f
ier
cir
cu
its
w
i
th
m
an
y
s
tag
e
s
,
a
to
tal
o
f
4
7
7
w
o
r
d
s
,
an
d
th
e
n
s
o
o
n
,
f
o
r
th
e
o
t
h
er
tex
ts
.
T
ab
le
4
.
Me
asu
r
e
m
en
t
R
es
u
lt
s
w
it
h
S
k
y
p
e
R
e
a
d
t
e
x
t
s
S
p
e
e
c
h
u
n
d
e
r
st
a
n
d
a
b
i
l
i
t
y
%
W
E
R
(
W
o
r
d
E
r
r
o
r
R
a
t
e
)
%
F1
9
7
%
3%
F2
9
5
.
5
%
4
.
5
%
F3
9
6
%
4%
F4
9
9
%
1%
F5
9
8
%
2%
I
n
th
e
d
iag
r
a
m
p
r
esen
ted
at
t
h
e
Fig
u
r
e
1
we
s
ee
t
h
e
d
ep
en
d
en
ce
o
f
t
h
e
er
r
o
r
o
f
w
o
r
d
s
in
r
el
atio
n
to
th
e
te
x
t
r
ea
d
.
Fig
u
r
e
1
.
G
r
ap
h
ical
Re
p
r
esen
t
atio
n
o
f
t
h
e
Me
as
u
r
e
m
e
n
t
s
w
i
t
h
S
k
y
p
e
M
e
a
s
ure
m
e
nts w
it
h Viber
a
pp
lica
t
io
n
Th
e
d
a
ta
a
b
o
u
t
q
u
a
lity
o
f in
tern
et
fo
r
V
ib
er a
p
p
me
a
s
u
r
eme
n
ts
in
a
clea
n
en
viro
n
men
t wi
th
o
u
t n
o
is
e.
•
Do
w
n
lo
ad
s
p
ee
d
-
4
Mb
p
s
•
Up
lo
ad
s
p
ee
d
-
1
.
8
Mb
p
s
•
P
in
g
-
4
8
m
s
•
J
itter
-
1
m
s
•
L
i
n
e
o
f
I
n
ter
n
et
q
u
ali
t
y
i
s
B
•
MO
S 4
.
1
•
T
h
e
d
is
tan
ce
o
f
m
ea
s
u
r
e
m
e
n
t
f
r
o
m
t
h
e
lo
ca
l h
o
s
t ser
v
er
I
p
k
o
in
Fra
n
k
f
u
r
t,
Ger
m
a
n
y
J
u
s
t li
k
e
in
S
k
y
p
e
w
e
h
a
v
e
u
s
e
d
th
e
s
a
m
e
m
et
h
o
d
f
o
r
m
ea
s
u
r
e
m
en
t
in
Vib
er
also
,
b
u
t
h
er
e
w
e
h
av
e
u
s
e
d
Vib
er
in
S
m
ar
t P
h
o
n
e
s
w
it
h
s
p
ec
if
icatio
n
s
o
u
t
lin
ed
ab
o
v
e.
8%
12%
6%
8%
2%
0%
20%
WE
R
R
ea
d
t
e
xt
s
M
easu
rme
n
t
wi
th
Skyp
e
F1
F2
F3
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
J
E
C
E
Vo
l.
7
,
No
.
5
,
Octo
b
er
201
7
:
2
5
1
4
–
2
5
1
9
2518
T
ab
le
5
.
Me
asu
r
e
m
en
t
R
es
u
lt
s
w
it
h
Vib
er
R
e
a
d
t
e
x
t
s
S
p
e
e
c
h
u
n
d
e
r
s
t
a
n
d
a
b
i
l
i
t
y
%
W
E
R
(
W
o
r
d
E
r
r
o
r
R
a
t
e
)
%
F1
9
2
%
8%
F2
8
8
%
1
2
%
F3
9
4
%
6%
F4
9
2
%
8%
F5
9
8
%
2%
Fro
m
th
e
s
e
d
ata
w
e
p
r
ese
n
t th
e
f
o
llo
w
in
g
d
iag
r
a
m
p
r
esen
ted
in
th
e
F
ig
u
r
e
2
.
Fig
u
r
e
2
.
G
r
ap
h
ical
Re
p
r
esen
t
atio
n
of
t
h
e
Me
as
u
r
e
m
e
n
t
s
w
i
t
h
Vib
er
B
elo
w
w
e
p
r
esen
t t
h
e
d
if
f
er
en
ce
b
et
w
ee
n
ap
p
licatio
n
s
S
k
y
p
e,
Vib
er
an
d
Z
o
ip
er
in
ter
m
s
o
f
in
telli
g
ib
ilit
y
o
f
s
p
ee
ch
in
e
n
v
ir
o
n
m
en
t
s
w
it
h
o
u
t n
o
i
s
e,
to
s
ee
it
as
a
s
u
m
m
ar
y
o
f
all
te
x
t
s
i
n
to
tal
m
ea
n
i
n
g
t
h
e
av
er
a
g
e
v
al
u
e
o
f
all
tex
t
s
.
T
ab
le
6
.
Fin
al
M
ea
s
u
r
e
m
e
n
t R
esu
lt
s
b
et
w
ee
n
Sk
y
p
e
an
d
Vib
er
U
sed Ap
p
s
S
p
e
e
c
h
u
n
d
e
r
st
a
n
d
a
b
i
l
i
t
y
%
(
A
v
e
r
a
g
e
v
a
l
u
e
)
S
k
y
p
e
9
7
.
1
%
V
i
b
e
r
9
1
.
6
%
W
h
er
ea
s
th
e
p
r
esen
tatio
n
h
as
a
ch
ar
t
lik
e
t
h
e
o
n
e
p
r
esen
ted
i
n
th
e
Fig
u
r
e
3
.
Fig
u
r
e
3
.
G
r
ap
h
ical
Re
p
r
esen
t
atio
n
o
f
Me
as
u
r
e
m
e
n
t
s
b
et
w
ee
n
w
it
h
V
i
b
er
an
d
Sk
y
p
e.
4.
CO
NCLU
SI
O
N
8%
12%
6%
8%
2%
0%
20%
WE
R
R
ea
d
t
e
xt
s
M
easu
rme
n
t
wi
th
Vib
er
F1
F2
F3
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
E
C
E
I
SS
N:
2088
-
8708
Th
e
E
ffect
o
f B
a
n
d
w
id
th
o
n
S
p
ee
ch
I
n
tellig
ib
ilit
y
in
A
lb
a
n
ia
n
La
n
g
u
a
g
e
…
(
S
a
b
r
ije
Osma
n
a
j
)
2519
R
es
u
lts
ac
h
iev
ed
f
o
r
th
e
i
n
t
ellig
ib
ili
t
y
o
f
s
p
ee
ch
ar
e
p
er
f
ec
t
f
o
r
all
ap
p
licatio
n
s
t
h
at
w
e
u
s
ed
,
m
ea
n
in
g
t
h
e
v
a
lu
e
s
ar
e
o
v
er
7
6
%
w
h
ich
i
s
co
n
s
id
er
ed
th
e
h
ig
h
e
s
t
d
eg
r
ee
.
Fro
m
t
h
is
we
co
n
clu
d
e
t
h
at
th
e
A
lb
a
n
ian
la
n
g
u
a
g
e
i
s
a
la
n
g
u
ag
e
t
h
at
i
s
ea
s
il
y
u
n
d
er
s
to
o
d
an
d
t
h
e
m
ain
r
ea
s
o
n
w
h
y
w
e
g
et
th
e
s
e
p
o
s
iti
v
e
r
esu
lt
s
ar
e
th
e
v
o
w
els
w
h
ic
h
g
iv
e
t
h
e
m
ea
n
i
n
g
o
f
th
e
w
o
r
d
s
.
E
ith
er
s
o
,
w
e
s
till
r
e
m
ain
r
eser
v
e
d
f
o
r
th
es
e
r
esu
lt
s
b
ec
au
s
e
m
ea
s
u
r
e
m
en
t
s
m
ad
e
i
n
th
i
s
r
ep
o
r
t
ar
e
o
n
l
y
b
et
w
ee
n
t
w
o
p
er
s
o
n
s
w
h
o
ar
e
f
a
m
iliar
w
it
h
ea
c
h
o
th
er
,
if
t
h
e
tes
tin
g
is
m
ad
e
w
it
h
m
o
r
e
p
eo
p
le
w
h
o
ar
e
u
n
f
a
m
iliar
b
et
w
ee
n
t
h
e
m
s
el
v
es,
is
e
x
p
ec
ted
to
h
a
v
e
lo
w
er
v
a
lu
e
s
o
f
i
n
tell
ig
ib
il
it
y
o
f
s
p
ee
ch
.
A
r
ed
u
c
tio
n
o
f
in
te
llig
ib
ili
t
y
ca
n
b
e
f
o
u
n
d
at
th
e
m
ea
s
u
r
e
m
e
n
t
s
d
o
n
e
b
et
w
ee
n
p
eo
p
le
w
h
o
co
m
m
u
n
icate
a
n
d
d
o
n
’
t
r
ec
o
g
n
ize
e
ac
h
o
t
h
er
w
h
ich
i
n
o
u
r
ca
s
e
m
ea
s
u
r
e
m
e
n
ts
ar
e
ca
r
r
ied
o
u
t
b
et
w
ee
n
p
er
s
o
n
s
w
h
o
ar
e
k
n
o
w
n
to
ea
c
h
o
t
h
er
an
d
t
h
is
is
o
n
e
o
f
th
e
r
ea
s
o
n
s
f
o
r
th
is
r
esu
lts
w
it
h
h
ig
h
v
alu
e
s
.
T
h
e
f
u
t
u
r
e
of
t
h
is
f
ield
i
s
i
n
th
at
t
h
er
e
i
s
m
u
c
h
to
b
e
d
o
n
e
an
d
r
eq
u
ir
ed
to
w
o
r
k
i
n
g
r
o
u
p
s
f
r
o
m
v
ar
io
u
s
f
iel
d
s
e
n
g
i
n
ee
r
in
g
,
p
r
o
g
r
a
m
m
i
n
g
,
li
n
g
u
i
s
tic
s
,
b
ec
au
s
e
is
eq
u
al
l
y
c
h
alle
n
g
in
g
a
n
d
als
o
v
er
y
n
ec
es
s
ar
y
f
o
r
th
i
s
m
o
d
er
n
ti
m
e.
T
r
u
l
y
,
t
h
e
f
o
r
ce
s
o
f
r
elev
a
n
t
n
atio
n
a
l
ex
p
er
ts
s
h
o
u
l
d
b
e
j
o
in
t
to
g
et
h
er
to
d
ig
italize
t
h
e
A
lb
a
n
ian
lan
g
u
ag
e
b
ec
au
s
e
t
h
is
w
i
ll
b
e
an
ad
v
a
n
ta
g
e
n
o
t
o
n
l
y
i
n
s
p
ee
ch
in
te
llig
ib
ilit
y
b
u
t
also
in
m
a
n
y
au
to
m
at
ic
s
y
s
te
m
s
s
u
c
h
as
r
o
b
o
tics
,
m
ed
icin
e
to
th
e
m
ac
h
in
e
s
t
h
at
w
o
r
k
w
it
h
s
ig
n
aliz
atio
n
o
f
th
e
v
o
ice
,
w
h
ic
h
also
ar
e
u
s
ed
in
o
u
r
co
u
n
tr
y
,
t
h
en
i
n
th
e
f
u
tu
r
e
th
e
ca
r
m
an
u
f
ac
tu
r
i
n
g
in
d
u
s
tr
y
o
f
i
n
f
o
r
m
atio
n
tech
n
o
lo
g
y
is
m
ea
n
in
g
le
s
s
to
d
ev
elo
p
ap
p
licatio
n
s
a
n
d
m
a
c
h
in
e
s
a
n
d
n
o
t
i
n
te
g
r
ate
A
lb
an
ian
lan
g
u
ag
e
o
n
th
e
m
.
RE
F
E
R
E
NC
E
S
[1
]
Ro
g
e
r
L
.
F
re
e
m
a
n
,
Fu
n
d
a
me
n
t
a
ls
o
f
T
e
lec
o
mm
u
n
ica
t
io
n
s
,
IEE
E
P
r
e
ss
,
W
il
le
y
In
ters
c
ien
c
e
,
2
0
0
9
.
[2
]
T
h
o
m
a
s F
.
Qu
a
ti
e
ri.
Disc
re
te
-
T
ime
S
p
e
e
c
h
S
i
g
n
a
l
Pro
c
e
ss
in
g
.
P
re
n
t
ice
-
Ha
ll
,
3
rd
e
d
it
io
n
.
1
9
9
6
.
[3
]
J.Ro
d
m
a
n
.
T
h
e
Ef
fec
t
o
f
Ba
n
d
w
i
d
th
o
n
S
p
e
e
c
h
In
telli
g
i
b
il
it
y
.
Co
m
m
o
n
w
e
a
lt
h
T
e
le
c
o
m
m
u
n
ica
ti
o
n
s
Org
a
n
iza
t
io
n
,
S
e
p
tem
b
e
r
2
0
0
6
.
[4
]
V
e
to
n
Kë
p
u
sk
a
.
V
On
W
a
k
e
-
Up
-
W
o
rd
S
p
e
e
c
h
Rec
o
g
n
it
io
n
T
a
sk
,
T
e
c
h
n
o
l
o
g
y
,
a
n
d
Eva
l
u
a
t
io
n
Res
u
lt
s
a
g
a
in
st
HTK
a
n
d
M
icr
o
so
f
t
S
DK
5
.
1
.
In
v
it
e
d
P
a
p
e
r:
W
o
rld
Co
n
g
re
ss
o
n
No
n
li
n
e
a
r
A
n
a
l
y
sts,
Orla
n
d
o
2
0
0
8
,
J
o
u
r
n
a
l
o
f
No
n
l
in
e
a
r
An
a
lys
is,
T
h
e
o
ry
,
M
e
th
o
d
s &
Ap
p
li
c
a
ti
o
n
s.
&
Klein
,
T
.
2
0
0
8
.
[5
]
H.K.
P
a
lo
,
M
ih
ir
Na
ra
y
a
n
M
o
h
a
n
ty
.
Cla
ss
if
ica
ti
o
n
o
f
Emo
ti
o
n
a
l
S
p
e
e
c
h
o
f
Ch
il
d
re
n
Us
in
g
Pro
b
a
b
il
ist
ic
Ne
u
ra
l
Ne
two
rk
.
In
tern
a
ti
o
n
a
l
Jo
u
r
n
a
l
o
f
El
e
c
tri
c
a
l
a
n
d
Co
m
p
u
ter
E
n
g
in
e
e
rin
g
(IJECE)
V
o
l.
5
,
N
o
.
2
,
A
p
ril
2
0
1
5
,
p
p
.
3
1
1
~
3
1
7
IS
S
N:
2
0
8
8
-
8
7
0
8
.
[6
]
D.
Ju
ra
f
sk
y
a
n
d
Ja
m
e
s
H.
M
a
rti
.
S
p
e
e
c
h
a
n
d
L
a
n
g
u
a
g
e
P
ro
c
e
ss
in
g
:
A
n
in
tro
d
u
c
ti
o
n
t
o
n
a
tu
ra
l
lan
g
u
a
g
e
p
ro
c
e
ss
in
g
,
c
o
m
p
u
tatio
n
a
l
li
n
g
u
isti
c
s,
a
n
d
s
p
e
e
c
h
re
c
o
g
n
it
io
n
Co
p
y
rig
h
t
2
0
0
6
,
d
ra
f
t
o
f
Ju
n
e
2
5
,
2
0
0
7
.
[7
]
R.
V
i
p
p
e
rla..
Au
t
o
ma
ti
c
S
p
e
e
c
h
R
e
c
o
g
n
it
i
o
n
fo
r
a
g
e
i
n
g
v
o
ice
s,
In
s
ti
tu
te
f
o
r
L
a
n
g
u
a
g
e
,
Co
g
n
it
io
n
a
n
d
C
o
m
p
u
tatio
n
S
c
h
o
o
l
o
f
In
f
o
rm
a
ti
c
s Un
iv
e
rsit
y
o
f
Ed
in
b
u
rg
h
2
0
1
1
.
[8
]
Ka
y
o
d
e
F
ra
n
c
is
A
k
in
g
b
a
d
e
,
O
k
o
k
o
M
k
p
o
u
t
o
Um
a
n
n
a
,
Isia
k
a
A
j
e
wa
le
A
li
m
i.
V
o
ice
-
Ba
se
d
Do
o
r
A
c
c
e
ss
Co
n
tro
l
S
y
st
e
m
Us
in
g
t
h
e
M
e
l
F
re
q
u
e
n
c
y
Ce
p
stru
m
Co
e
ff
icie
n
ts
a
n
d
G
a
u
ss
ian
M
ix
t
u
re
M
o
d
e
l
,
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
trica
l
a
n
d
C
o
mp
u
ter
En
g
in
e
e
rin
g
(
IJ
ECE
)
Vo
l.
4
,
No
.
5
,
Oc
t
o
b
e
r
2
0
1
4
,
p
p
.
6
4
3
~
6
4
7
IS
S
N:
2
0
8
8
-
8
7
0
8
.
[9
]
S
a
e
e
d
V
.
V
a
se
g
h
i.
M
u
l
ti
me
d
ia
S
i
g
n
a
l
Pro
c
e
ss
in
g
,
T
h
e
o
ry
a
n
d
Ap
p
li
c
a
ti
o
n
s i
n
S
p
e
e
c
h
,
M
u
sic
a
n
d
C
o
mm
u
n
ic
a
ti
o
n
s
,
De
p
a
rtme
n
t
o
f
El
e
c
tro
n
ics
,
S
c
h
o
o
l
o
f
En
g
in
e
e
rin
g
a
n
d
De
sig
n
Bru
n
e
l
Un
iv
e
rsit
y
,
UK
,
2
0
0
7
.
[1
0
]
L
.
L
iu
a
n
d
L
.
S
u
n
:
Per
fo
rm
a
n
c
e
An
a
lys
is
o
f
V
o
ice
Ca
ll
u
sin
g
S
k
y
p
e
,
Ce
n
tre f
o
r
S
e
c
u
rit
y
,
Co
m
m
u
n
ica
ti
o
n
s an
d
Ne
tw
o
r
k
R
e
se
a
r
c
h
P
ly
m
o
u
th
Un
i
v
e
rsit
y
,
Un
it
e
d
Kin
g
d
o
m
,
2
0
1
1
.
Evaluation Warning : The document was created with Spire.PDF for Python.