<s>
Computer	B-General_Concept
audition	I-General_Concept
(	O
CA	O
)	O
or	O
machine	B-General_Concept
listening	I-General_Concept
is	O
the	O
general	O
field	O
of	O
study	O
of	O
algorithms	O
and	O
systems	O
for	O
audio	O
interpretation	O
by	O
machines	O
.	O
</s>
<s>
Since	O
the	O
notion	O
of	O
what	O
it	O
means	O
for	O
a	O
machine	O
to	O
"	O
hear	O
"	O
is	O
very	O
broad	O
and	O
somewhat	O
vague	O
,	O
computer	B-General_Concept
audition	I-General_Concept
attempts	O
to	O
bring	O
together	O
several	O
disciplines	O
that	O
originally	O
dealt	O
with	O
specific	O
problems	O
or	O
had	O
a	O
concrete	O
application	O
in	O
mind	O
.	O
</s>
<s>
Inspired	O
by	O
models	O
of	O
human	O
audition	O
,	O
CA	O
deals	O
with	O
questions	O
of	O
representation	O
,	O
transduction	B-General_Concept
,	O
grouping	O
,	O
use	O
of	O
musical	O
knowledge	O
and	O
general	O
sound	O
semantics	B-Application
for	O
the	O
purpose	O
of	O
performing	O
intelligent	O
operations	O
on	O
audio	O
and	O
music	O
signals	O
by	O
the	O
computer	O
.	O
</s>
<s>
Technically	O
this	O
requires	O
a	O
combination	O
of	O
methods	O
from	O
the	O
fields	O
of	O
signal	O
processing	O
,	O
auditory	O
modelling	O
,	O
music	O
perception	O
and	O
cognition	O
,	O
pattern	O
recognition	O
,	O
and	O
machine	O
learning	O
,	O
as	O
well	O
as	O
more	O
traditional	O
methods	O
of	O
artificial	B-Application
intelligence	I-Application
for	O
musical	O
knowledge	O
representation	O
.	O
</s>
<s>
Like	O
computer	B-Application
vision	I-Application
versus	O
image	O
processing	O
,	O
computer	B-General_Concept
audition	I-General_Concept
versus	O
audio	O
engineering	O
deals	O
with	O
understanding	O
of	O
audio	O
rather	O
than	O
processing	O
.	O
</s>
<s>
It	O
also	O
differs	O
from	O
problems	O
of	O
speech	B-Application
understanding	I-Application
by	I-Application
machine	I-Application
since	O
it	O
deals	O
with	O
general	O
audio	O
signals	O
,	O
such	O
as	O
natural	O
sounds	O
and	O
musical	O
recordings	O
.	O
</s>
<s>
Applications	O
of	O
computer	B-General_Concept
audition	I-General_Concept
are	O
widely	O
varying	O
,	O
and	O
include	O
search	B-Application
for	I-Application
sounds	I-Application
,	O
genre	O
recognition	O
,	O
acoustic	O
monitoring	O
,	O
music	O
transcription	O
,	O
score	O
following	O
,	O
audio	O
texture	O
,	O
music	O
improvisation	O
,	O
emotion	O
in	O
audio	O
and	O
so	O
on	O
.	O
</s>
<s>
Computer	B-General_Concept
Audition	I-General_Concept
overlaps	O
with	O
the	O
following	O
disciplines	O
:	O
</s>
<s>
Music	B-Application
information	I-Application
retrieval	I-Application
:	O
methods	O
for	O
search	O
and	O
analysis	O
of	O
similarity	O
between	O
music	O
signals	O
.	O
</s>
<s>
Auditory	B-Application
scene	I-Application
analysis	I-Application
:	O
understanding	O
and	O
description	O
of	O
audio	O
sources	O
and	O
events	O
.	O
</s>
<s>
Computer	B-Application
music	I-Application
:	O
use	O
of	O
computers	O
in	O
creative	O
musical	O
applications	O
.	O
</s>
<s>
Since	O
audio	O
signals	O
are	O
interpreted	O
by	O
the	O
human	O
ear	O
–	O
brain	O
system	O
,	O
that	O
complex	O
perceptual	O
mechanism	O
should	O
be	O
simulated	O
somehow	O
in	O
software	O
for	O
"	O
machine	B-General_Concept
listening	I-General_Concept
"	O
.	O
</s>
<s>
Analyzing	O
audio	O
accurately	O
involves	O
several	O
fields	O
:	O
electrical	O
engineering	O
(	O
spectrum	O
analysis	O
,	O
filtering	O
,	O
and	O
audio	O
transforms	O
)	O
;	O
artificial	B-Application
intelligence	I-Application
(	O
machine	O
learning	O
and	O
sound	O
classification	O
)	O
;	O
psychoacoustics	O
(	O
sound	O
perception	O
)	O
;	O
cognitive	O
sciences	O
(	O
neuroscience	O
and	O
artificial	B-Application
intelligence	I-Application
)	O
;	O
acoustics	O
(	O
physics	O
of	O
sound	O
production	O
)	O
;	O
and	O
music	O
(	O
harmony	B-Library
,	O
rhythm	O
,	O
and	O
timbre	O
)	O
.	O
</s>
<s>
For	O
best	O
results	O
,	O
these	O
transformations	O
require	O
perceptual	O
understanding	O
of	O
spectral	O
models	O
,	O
high-level	O
feature	B-Algorithm
extraction	I-Algorithm
,	O
and	O
sound	O
analysis/synthesis	O
.	O
</s>
<s>
Computational	O
models	O
of	O
music	O
and	O
sound	O
perception	O
and	O
cognition	O
can	O
lead	O
to	O
a	O
more	O
meaningful	O
representation	O
,	O
a	O
more	O
intuitive	O
digital	B-General_Concept
manipulation	O
and	O
generation	O
of	O
sound	O
and	O
music	O
in	O
musical	O
human-machine	O
interfaces	O
.	O
</s>
<s>
Feature	B-Algorithm
extraction	I-Algorithm
:	O
sound	O
descriptors	O
,	O
segmentation	O
,	O
onset	O
,	O
pitch	O
and	O
envelope	B-Algorithm
detection	I-Algorithm
,	O
chroma	O
,	O
and	O
auditory	O
representations	O
.	O
</s>
<s>
Musical	O
knowledge	O
structures	O
:	O
analysis	O
of	O
tonality	O
,	O
rhythm	O
,	O
and	O
harmonies	B-Library
.	O
</s>
<s>
Multi-modal	B-Application
analysis	O
:	O
finding	O
correspondences	O
between	O
textual	O
,	O
visual	O
,	O
and	O
audio	O
signals	O
.	O
</s>
<s>
Computer	B-General_Concept
audition	I-General_Concept
deals	O
with	O
audio	O
signals	O
that	O
can	O
be	O
represented	O
in	O
a	O
variety	O
of	O
fashions	O
,	O
from	O
direct	O
encoding	O
of	O
digital	B-General_Concept
audio	O
in	O
two	O
or	O
more	O
channels	O
to	O
symbolically	O
represented	O
synthesis	O
instructions	O
.	O
</s>
<s>
Audio	O
signals	O
are	O
usually	O
represented	O
in	O
terms	O
of	O
analogue	O
or	O
digital	B-General_Concept
recordings	O
.	O
</s>
<s>
Digital	B-General_Concept
recordings	O
are	O
samples	O
of	O
acoustic	O
waveform	O
or	O
parameters	O
of	O
audio	O
compression	O
algorithms	O
.	O
</s>
<s>
Additional	O
types	O
of	O
data	O
that	O
are	O
relevant	O
for	O
computer	B-General_Concept
audition	I-General_Concept
are	O
textual	O
descriptions	O
of	O
audio	O
contents	O
,	O
such	O
as	O
annotations	O
,	O
reviews	O
,	O
and	O
visual	O
information	O
in	O
the	O
case	O
of	O
audio-visual	O
recordings	O
.	O
</s>
<s>
Generally	O
speaking	O
,	O
one	O
could	O
divide	O
the	O
features	O
into	O
signal	O
or	O
mathematical	O
descriptors	O
such	O
as	O
energy	O
,	O
description	O
of	O
spectral	O
shape	O
etc.	O
,	O
statistical	O
characterization	O
such	O
as	O
change	O
or	O
novelty	O
detection	O
,	O
special	O
representations	O
that	O
are	O
better	O
adapted	O
to	O
the	O
nature	O
of	O
musical	O
signals	O
or	O
the	O
auditory	O
system	O
,	O
such	O
as	O
logarithmic	O
growth	O
of	O
sensitivity	O
(	O
bandwidth	B-Algorithm
)	O
in	O
frequency	O
or	O
octave	B-Language
invariance	O
(	O
chroma	O
)	O
.	O
</s>
<s>
Generally	O
,	O
this	O
creates	O
a	O
situation	O
where	O
computer	B-General_Concept
audition	I-General_Concept
can	O
not	O
rely	O
solely	O
on	O
detection	O
of	O
specific	O
features	O
or	O
sound	O
properties	O
and	O
has	O
to	O
come	O
up	O
with	O
general	O
methods	O
of	O
adapting	O
to	O
changing	O
auditory	O
environment	O
and	O
monitoring	O
its	O
structure	O
.	O
</s>
<s>
Computer	B-General_Concept
Audition	I-General_Concept
tries	O
to	O
find	O
relation	O
between	O
these	O
different	O
representations	O
in	O
order	O
to	O
provide	O
this	O
additional	O
understanding	O
of	O
the	O
audio	O
contents	O
.	O
</s>
