|
1983 - Bryan Adams - Cuts Like a Knife (A&M
CD-3288)
This represents the earliest stage of CD audio mastering
techniques. One of the big attractions of the CD
format was its very wide dynamic range and absence of
surface noise, so the first generation of CDs that were
released made full use of this. The CD's digital
audio format is inherently limited to a peak audio
amplitude that is referenced as "0 dB" or
"100%."
Unlike the variability of analog recording techniques,
this limit cannot be exceeded. If you try to,
you end up with a waveform that is horribly distorted, due
to the top and/or bottom of the wave being harshly limited
to the maximum level -- or "clipped." Early CDs
were mastered with full respect to this loudness limit and
did not use any clipping.
In this case, Bryan Adams' breakthrough album -- although
full of stereotypically "loud" rock & roll
music -- was mastered with a great amount of
"headroom"; the highest peak level on the entire
disc is only 74.8%, or -2.52 dB. This peak level
only occurs a single time on the disc, in Track 9,
whose waveform graph is shown here (with the stereo Left
channel on top, and Right channel on bottom):
Obviously this peak (which occurs at around 2:16 in the
song) is not even close to being clipped, and in fact the
loudness of
the whole CD could be increased by 2.5 dB without
sacrificing any quality or harming the content of the
recordings in any way. That single peak would simply be
raised up to just below the 100% limit. You can
think of this as a short person raising up the
height-adjustable driver's seat in their car. They
would sit at a higher level, with their head just below
the ceiling, but not touching it -- in this case, literal
"headroom."
1988 - Willie Nelson - What a Wonderful World (Columbia CK
44331)
This is an example of what I was just talking about.
With this CD, there is a single peak on the entire disc
which does reach the 100% limit, but yet is still
not "clipped", and every other track on the
album has a lower peak level. Here is the waveform
of Track 6, showing the single 100% (0 dB) peak near the
beginning:
And here is a highly "zoomed in" view of this
peak, to prove that it is not clipped:
This represents the extreme of the most conservative form
of CD mastering; where only a single peak, occurring only
in a
single track, reaches the maximum level, and everything
else on the disc is below it. However, as we shall
see, as the years went by and CDs became more and more
popular, this pristine form of CD audio production became
more and more rare!
1987 - George Michael - Faith
(Columbia CK 40867)
Yes, we are stepping back a year, in this case to show a
more contemporary pop music CD. On this particular
CD, most of the tracks contain multiple peaks which reach
the 100% limit, but are still not clipped. Here is
the waveform of the title track
"Faith."
So, there are multiple instances in the song where the
100% / 0 dB peak level is reached. But are these
peaks clipped? I looked at all of them to
check, and none of them were. Here is an example of
a few of these 100% peaks in the song, and as you can see,
the shape of the waveform is still fully preserved:
For the average public's listening tastes, this is
probably the ideal compromise between pristine, unharmed
audio quality, and the goal of having CDs sound a bit
"louder" when played -- and most CDs from the
late 1980s continued to respect this
compromise. But by the early 1990s, things had begun
to change -- for the worse.
1991 - Amy Grant - Heart In Motion (A&M 75021 5321 2)
Alas, in the highly competitive pop music world, something
had to give; who was first to do it may be lost to
history, but by this time, the trend towards the reduction
of the CD's quality and dynamic range had already begun.
In this particular case, not only do many songs on the CD
reach maximum peak level, a number of these peaks in each
song are also "clipped" -- an instance where the
top and/or bottom of the waveform has been
"flat-topped" or "hacked off" because
it ran into the brick wall known as the 100% / 0 dB limit.
This is evident by looking at the waveform graph of Track
3:
By zooming in on one of these peaks, I can illustrate just
what I have been talking about:
In this instance, seven samples (or "digital
snapshots" of the sound) in a row are shoved up
against the limit, and they have nowhere to go, so they
just form a flat line -- hence, our old friend known as
"clipping". A perfectly flat line like
this never occurs in audio recording unless the
audio is purposely clipped. What it is doing is
generating a very unnatural "square wave," and
when you play back this square wave, you get a burst of
distorted sound. CD digital audio runs at 44,100
samples per second. In this case, seven of these
samples formed a square wave. That's 7/44100th of a
second -- far too quick to be directly heard. But
when you add many of these clipped samples in each second
of the song, the effect becomes more and more noticeable
as "clipping distortion" -- a harsh, crackly
sound that is very unpleasant to the ear -- as we shall
see below.
1995 - The Rembrandts - L.P. (East West/Atlantic 61752-2)
Everybody loves "I'll Be There For You", the
theme song from the TV show "Friends", right?
Well, have you ever looked at its waveform graph?
Yikes!
Although this particular track only reaches a peak level
of 96.2% -- which is below the 100% limit and
theoretically should be free from "clipping" --
when you zoom in, you'll see that many of the MANY peaks
that are strictly limited to this level (causing the
"solid block of sound" effect) are clipped, like
this one:
The sad part is that when these peaks are clipped, or
"hacked off" as I like to call it, the original
content and quality of the sound is lost forever.
Although there are techniques to "guess" at the
missing part of the waveform and "re-draw" a
curved wave out of the part that was formerly a flat line,
this is not accurate to the original sound; it is merely
an "artist's reconstruction", so to speak.
Thus, when aggressive peak clipping is used, the record
company is DESTROYING part of the music. And
if you thought this song was bad enough, hold on and be
prepared for a very sickening sight below...
1999 - Ricky Martin (C2/Columbia CK 69891)
By now I don't think I have to explain what you see below.
And in this case, it probably will help explain why your
ears start to hurt after listening to "Livin' La Vida
Loca" even for just a short time!
And as we zoom in, the audio carnage is apparent:
The Radio Loudness Fallacy
Okay, so the end result of all this stuff shown above is
that CDs have gotten louder and louder over the years. But
that also makes them sound louder when played on the
radio, right?
WRONG! Every radio station uses "audio
processing" to accommodate a wide variety of audio
programming and deliver a consistent volume level, so that
you're not always reaching for the volume control to turn
up quiet songs and turn down loud songs. The goal of
this is to make every song come out of your radio at a
fairly equal volume level. The station's audio
processing does this by carefully compressing, limiting,
and yes, even clipping the audio. It also equalizes
the bass and treble levels so one song doesn't sound bassy
and muffled and the next doesn't sound tinny and shrill.
When properly adjusted, this audio processing can work
wonders, and gives radio stations just the type of sound
they're looking to provide their listeners with, be it a
soft and relaxing Classical music station, or a loud and
brash Heavy Metal rock station. But... this audio
processing relies on the peaks and valleys of the incoming
audio; it raises up the quiet spots and limits down
the
loud spots.
BUT WHEN EVERYTHING IS CONSISTENTLY LOUD
ALL THE TIME, LIKE THIS SENTENCE, the radio station's audio processor has nothing to do but
to limit down the volume level, and since the incoming
music contains virtually no variations in loudness, then
it stays consistently quiet when played on the air.
It's like walking along a street and passing a
construction zone where they have a jack-hammer going.
Due to the consistently loud noise, you have no choice but
to cover your ears until it stops.
Compare that example to a fireworks display. You may
cover you ears during the explosions, but otherwise,
things are much quieter and you could actually amplify the
sounds around you and not damage your hearing. This
is like the CDs of the 1980s shown above. The high
peaks get limited down so that the radio station doesn't
exceed the signal level which it can transmit,
but otherwise the rest of the audio actually gets raised
in level, so that it comes out louder on the air.
And you can get this same effect at home by simply TURNING
UP YOUR STEREO'S VOLUME CONTROL. You want your music
to be loud? You can make it loud yourself -- and the
full quality and dynamic range of the music is preserved.
Yes, even the Bryan Adams CD from 1983 can be played louder
than Ricky Martin's horribly produced CD -- the choice is
up to you. But when all of your CDs are recorded to
be loud right on the discs themselves, you don't have this
choice anymore; you no longer have a variety of
"loud" music and "quiet" music to
choose from and to play at a volume level that suits your
musical taste. The record companies are not only
filling your CDs with distorted, corrupted audio, they are
forcing you to listen to your music in a certain manner --
do you really want that?
Meanwhile, in Europe...
At least for some record companies "across the
pond", it took a lot longer for the "loudness
wars" to catch up with them than it did in the USA.
Take, for example, the original April 1999 CD single
release "Swear It Again" by Westlife (RCA/BMG
74321 66206 2) -- which was the first of a record-breaking
string of #1 U.K. hits for this Irish pop group:
This is just like the 1988 Willie Nelson CD shown far
above on this page. Only a single peak during the
entire song reaches maximum level -- in this case, 97.1%,
and it is not clipped:
But a funny thing happened to this very same song when it
ended up on Westlife's first full album (RCA/BMG 74321
713212) later in the year:
That may look like a drastic increase in loudness, and it
is, but notice that as compared to American pop music of
the 1990s, it still has a decent amount of "wiggle
room" for the peaks and valleys of the sound to
occur. And if we zoom in, only five peaks
during the entire song occur at the maximum level, and
NONE OF THEM ARE CLIPPED, as you can see from these two:
This is pretty much what I would consider to be the limit
of RESPONSIBLE CD PRODUCTION. It allows the
music to sound "loud" without DESTROYING the
audio through the use of peak clipping.
But, alas, the influence of American pop music became too
great, and the U.K. record companies felt they had to
respond by also resorting to the same tricks for their
music. Here is an example from 2000, the first track
from Westlife's second album "Coast To Coast"
(RCA/BMG 74321 824302):
What I don't get is that this song has its peak level set
unnaturally low, to only 88.1% out of the 100% possible --
and yet, even with over 1 dB of "headroom" to
work with, the audio is still clipped in numerous places
during the song:
You might just write this off as somebody setting the
recording level a bit low, but it is interesting to
compare it to the same
song on the Asian release of this album (RCA/BMG 74321
803772):
The peak level is now much closer to the CD's maximum
limit -- in this case 99.3% -- and at least some of
the peaks and valleys can still be seen amongst the highly
compressed audio. But, unfortunately, when zoomed
in, even this dynamically superior Asian release also used
clipping on some of the peaks:
Note that in each case above, when clipping is present, I
have picked the most clipped example out of the
entire song; that is, the instance which has the highest
number of "flat-topped" samples. So, while
this Asian version still contains clipping, it is less
drastically clipped than the U.K. release, while it
actually has a higher peak level. Are Asian ears
more sensitive to clipping distortion than British ears?
Maybe the record company knows something we don't....
|
