<s>
Binaural	B-Application
unmasking	I-Application
is	O
phenomenon	O
of	O
auditory	O
perception	O
discovered	O
by	O
Ira	O
Hirsh	O
.	O
</s>
<s>
In	O
binaural	B-Application
unmasking	I-Application
,	O
the	O
brain	O
combines	O
information	O
from	O
the	O
two	O
ears	O
in	O
order	O
to	O
improve	O
signal	O
detection	O
and	O
identification	O
in	O
noise	O
.	O
</s>
<s>
Binaural	B-Application
unmasking	I-Application
is	O
most	O
effective	O
at	O
low	O
frequencies	O
.	O
</s>
<s>
Binaural	B-Application
unmasking	I-Application
can	O
also	O
be	O
observed	O
for	O
narrowband	O
masking	O
noises	O
,	O
but	O
the	O
effect	O
behaves	O
differently	O
:	O
larger	O
BMLDs	O
can	O
be	O
observed	O
and	O
there	O
is	O
little	O
evidence	O
of	O
a	O
decline	O
with	O
increasing	O
frequency	O
.	O
</s>
<s>
The	O
fact	O
that	O
speech	O
can	O
be	O
unmasked	O
and	O
the	O
underlying	O
cues	O
vary	O
with	O
sound	O
direction	O
raised	O
the	O
possibility	O
that	O
binaural	B-Application
unmasking	I-Application
plays	O
a	O
role	O
in	O
the	O
cocktail	O
party	O
effect	O
.	O
</s>
<s>
Binaural	B-Application
unmasking	I-Application
has	O
two	O
main	O
explanatory	O
frameworks	O
.	O
</s>
<s>
Each	O
coincidence	O
detector	O
receives	O
a	O
stream	O
of	O
action	B-Algorithm
potentials	I-Algorithm
from	O
the	O
two	O
ears	O
via	O
a	O
network	O
of	O
axons	O
that	O
introduce	O
differential	O
transmission	O
delays	O
.	O
</s>
<s>
In	O
effect	O
,	O
the	O
coincidence	O
detectors	O
are	O
replaced	O
with	O
neurons	O
that	O
are	O
excited	O
by	O
action	B-Algorithm
potentials	I-Algorithm
from	O
one	O
ear	O
,	O
but	O
inhibited	O
by	O
action	B-Algorithm
potentials	I-Algorithm
from	O
the	O
other	O
.	O
</s>
<s>
For	O
a	O
binaural	B-Application
unmasking	I-Application
stimulus	O
,	O
the	O
differences	O
between	O
the	O
interaural	O
parameters	O
of	O
the	O
signal	O
and	O
noise	O
mean	O
that	O
there	O
will	O
be	O
a	O
different	O
vector	O
summation	O
at	O
each	O
ear	O
.	O
</s>
