Professional Documents
Culture Documents
MODULE 11 AUDIO EXAMPLES
The
audio
examples
for
this
module
on
mastering
use
a
section
of
the
song
The
Woodcutter's
Wife,
by
Audio
Masterclass
Featured
Artists
Chandelle.
The
mix
is
purposely
quite
plain
and
simple
so
that
the
effects,
and
side‐effects,
of
the
mastering
process
are
very
clear.
First, the raw mix without any mastering processing...
Example file: unmastered.wav
Hopefully the mastered version will be better. But, as we shall see, mastering can have negative effects too.
If
you
intend
to
have
one
of
your
own
mixes
mastered,
then
it
should
be
completely
unmastered,
like
the
example
you
have
just
heard.
This
gives
the
mastering
engineer
most
options.
If
you
are
mastering
yourself
however,
your
next
step
would
be
to
normalize
the
track
so
that
it
peaks
at
full
scale,
at
least
at
one
point
in
the
song.
This
will
make
the
track
louder,
and
with
no
change
in
sound
quality...
Example file: normalized.wav
All
of
the
following
examples
are
processed
versions
of
the
normalized
mix,
which
should
be
your
point
of
reference
that
you
compare
back
to.
It
may
be
found
at
this
point
that
the
mix
is
a
little
on
the
dull
side,
which
this
mix
clearly
is.
If
so,
then
a
little
EQ
can
help.
This
is
just
a
simple
upper
midrange
and
high
frequency
boost...
Example file: eq.wav
You
could
easily
wonder
at
this
point
how
much
EQ
is
enough.
Should
you
be
expected
to
compensate
for
the
average
listener
having
dull
and
blurry
speakers
for
instance?
This
question
is
best
answered
by
the
technique
of
'benchmarking'.
To
do
this,
you
should
load
up
a
commercially
released
track
in
the
same
genre
and
style
into
your
digital
audio
workstation,
so
that
you
can
A/B
compare
it
with
your
mix.
If
the
track
has
sonic
characteristics
that
you
like,
and
it
has
proven
successful
in
the
marketplace,
then
if
you
can
give
your
track
the
same
characteristics
you
can
be
confident
that
you
are
heading
in
the
right
direction.
The
next
stage
in
the
mastering
process
is
compression.
There
is
always
the
question
of
whether
to
EQ
then
compress,
or
compress
then
EQ.
The
answer
to
this
is
to
consider
what
the
raw
track
would
benefit
from
most,
if
you
could
only
use
one
process.
If
you
consider
that
it
would
be
most
improved
by
EQ,
then
you
should
EQ
first.
If
its
EQ
balance
is
already
good,
then
compress
first,
and
perhaps
tweak
it
with
a
little
more
EQ
after
compression.
Try it both ways. When you have a good feel for these processes, you will do it almost by instinct.
Here is the EQed then compressed version...
Example file: eq‐compress.wav
So far, you have made the track sound better, and the compression process has made it subjectively a little
Module
11
Audio
Examples
Notes
louder.
You can stop now.
Or...
if
you
want
to
get
involved
in
the
'loudness
war'
and
make
your
track
sound
as
loud
subjectively
as
many
commercial
releases,
you
will
want
to
go
further.
One
fairly
harmless
process
is
limiting.
This
is
like
compression,
but
the
ratio
is
very
high.
So
the
peaks
of
the
signal
are
'clamped'
at
a
certain
level
that
they
can
never
exceed.
By
limiting
maybe
the
top
3
to
6
dB
of
the
peaks,
the
level
of
the
rest
of
the
signal
can
be
raised
by
that
amount,
thus
making
the
track
subjectively
louder,
even
though
the
peaks
can
never
be
higher
in
level
than
in
the
normalized
version.
Example file: eq‐compress‐limit.wav
Notice how so far each new version represents an additional process performed on the previous version.
Now we will take the first step into the danger zone...
It
is
quite
common
on
commercially
released
recordings
that
clipping
has
been
used
as
a
means
of
increasing
subjective
loudness.
Clipping
is
very
dangerous
as
it
can
make
the
sound
really
harsh.
In
this
example,
the
track
has
been
raised
3
decibels
into
clipping.
In
other
words,
it
is
already
peaking
at
full
scale,
then
it
is
raised
3
dB
and
the
tips
of
the
waveform
are
sheared
off.
Example file: eq‐compress‐limit‐clip‐3.wav
This
is
now
distinctly
'edgy'.
But
you
will
find
that
many
people
will
say
they
prefer
it
on
first
listening.
It
does
get
very
tiring
though
after
a
while,
so
it
is
very
difficult
to
say
that
it
is
a
good
thing
in
any
way.
Raising the level by another 3 dB, making 6 dB of clipping, proves the point...
Example file: eq‐compress‐limit‐clip‐6.wav
Fortunately
there
is
an
alternative
to
clipping,
and
that
is
'soft
clipping'.
Here
is
the
EQed
version
put
through
a
soft
clipping
processor...
Example file: softclip.wav
Clearly this is better than hard clipping, but the sound is still a little on the harsh side.
Another
option
is
to
use
a
processor
that
is
specifically
designed
to
increase
subjective
loudness,
in
this
case
the
Maxim
plug‐in...
Example file: maxim.wav
Clearly
this
is
also
an
improvement
on
hard
clipping.
But
it
changes
the
sound
quite
considerably.
This
next
example
shows
the
Maxim
file
with
the
normalized
version
subtracted
from
it.
In
other
words,
this
is
what
Maxim
does
to
the
signal...
Example file: maxim‐subtract.wav
It
would
not
be
entirely
correct
to
say
that
this
is
what
Maxim
is
adding
to
the
signal,
but
it
does
show
that
this
kind
of
processing
is
pretty
strong
stuff
and
needs
to
be
handled
carefully.
Multiband EQ
One further mastering technique is multiband EQ. This is also used in radio broadcasting to strengthen the
Module
11
Audio
Examples
Notes
signal
so
that
it
comes
in
loud
and
clear
over
a
wide
coverage
area.
The
multiband
compressor
splits
the
signal
into
several
frequency
bands,
typically
four.
It
then
applies
compression
in
each
band
individually.
So,
where
a
standard
compressor
latches
onto,
say,
a
high‐midrange
peak
in
the
snare
drum,
and
compresses
the
whole
of
the
signal
with
reference
to
that,
the
multiband
compressor
will
compress
the
low
frequencies
separately,
and
the
low‐mid
and
the
high
frequencies
separately
too.
This
means
that
each
band
of
frequencies
can
be
maximized
in
level
within
itself,
thus
'packing
out'
the
available
space
with
useful
waveform,
as
much
as
possible.
And
it
sounds
like
this...
Example file: multiband.wav
You
should
compare
this
with
the
normalized
version
normalized.wav.
You
will
notice
that
the
multiband
compressed
version
is
louder.
But
has
it
changed
the
sound
in
any
other
way?
For
convenience,
here
is
a
version
of
the
multiband
compressed
track
that
is
lowered
in
level
so
subjectively
it
sounds
about
the
same
level
as
normalized.wav.
Example file: multiband‐corrected.wav
What
you
will
notice
if
you
listen
carefully
is
that
the
frequency
balance
is
changed
slightly.
This
is
to
be
expected.
Multiband
compression
does
involve
the
risk
of
changing
the
balance
of
frequencies.
However,
kept
within
reasonable
limits,
it
does
not
have
the
harshness
of
clipping,
soft
clipping,
or
loudness
maximization
plug‐ins.
As
a
final
amusement,
here
are
the
individual
bands
of
the
multiband
compressed
version
‐
you
can
now
clearly
hear
what
the
multiband
compressor
is
doing...
Example
file:
multiband‐low.wav
Example
file:
multiband‐low‐mid.wav
Example
file:
multiband‐high‐mid.wav
Example
file:
multiband‐high.wav
In
summary,
mastering
is
a
powerful
process
that
can
easily
make
things
worse,
not
better.
If
possible,
you
should
have
your
work
mastered
by
someone
who
spends
their
entire
career
doing
this.
If
you
do
it
yourself,
use
the
benchmarking
technique,
and
at
all
times
ask
yourself
whether
you
are
making
improvements.
End of audio examples notes.
Module 11 Audio Examples Notes