SACD, DSD,
DTS, AC-3, DVD-Audio, SuperAudio
I suspect
many disc collectors these days feel
they are drowning in alphabet soup—SACD,
DSD, DTS, AC-3, DVD-Audio, SuperAudio
— and too many numbers ranging from
44.1 to 96 kHz, from "one" to 16 to
32 Bit. Let’s try to line up the letters
and numbers and make some sense.
Regular CD
format is 2.0 stereo, 44.1kHz 16Bit
digital sound. This means that the frequency
response of the two channels on the
disk is limited to 44,100 divided by
two or 22,050 Hz; an attempt to record
frequencies higher than this could result
in the production of subharmonics, so
they must be filtered out before mastering.
In practice, the frequency limit is
even lower than this, due to the published
technical standard. This would not be
so bad, except that to be sure everything
above 22,050 Hz is removed, the laws
of physics require that some distortion
and attenuation of frequencies below
this limit must necessarily occur. The
16Bit encoding of each sample restricts
the usable dynamic range before noise
and distortion become objectionable,
although clever ways (e.g. Sony’s "Super
Bit Map," Mobile Fidelity’s "Ultradisc
UHR," or Keith Johnson’s "HDCD"
[requiring a special player and recently
acquired by Microsoft]) have been found
to work around this. Some CDs are "matrix
encoded" with surround sound much
as were the old SQ Quadraphonic LPs,
and many of these are labelled "Dolby
Surround." Because the Dolby laboratories
charge a royalty for the use of the
technique and the name, and since not
all matrix encoding routines are licensed,
there are CDs, quite a few in fact,
which play back on your 5.1 Dolby decoder
player in realistic surround sound without
any notice on the packaging. If you
like surround sound (I do), this is
always worth a try.
The SACD format
appears to be bound to the DSD "Direct
Stream Digital" "one bit"
recording technique. Recordings originally
made in other formats must be dubbed
to DSD before the SACD can be mastered.
That "one bit" of course cannot
be directly compared to, say, "24Bit"
or other descriptions of digital resolution.
My genius recording engineer friends
deplore DSD as bad news all around,
although some of the DSD recordings
have sounded very good to my ears. It
has been suggested that Sony’s high
royalty fees have contributed to the
distaste with which some view DSD, and
others suggest that the system itself
is fine but economics have led to certain
compromises in practice.
SACD is a
popular format with manufacturers because
Sony has built in a copy protection
scheme which reportedly uses two lasers
during playback, i.e., there is NO WAY
you’re ever going to be able to play
it on a computer DVD drive. Sony and
Philips had not until recently licensed
other disc manufacturers, so manufacturing
capacity was very limited, although
obviously now increasing steadily.
Here is a
good case where the media giants’ lawyers
may have been conned, if you find that
cause to rejoice. Sony has sold everybody
on the copy protection features of SACD,
but as any engineer knows, a commercial
pirate operation would have no difficulty
whatever re-digitising the analogue
signal and producing pirate DVD-Audio’s
from SACD’s—and they’d sound pretty
good. However, if Sony, et al, can "win
the race" and make sure "nobody"
buys DVD-Audio equipment, then the market
for pirate copies would be so small
they wouldn’t be worth tooling up for.
So, they are currently engaged in a
vigorous and expensive campaign to manipulate
the market. Remember, these are the
people who brought us the El-Cassette,
the Betamax, Digital Compact Cassette
(DCC), and the Mini-disc. Not all innovations
result in improvements, but, of course,
without innovation we would have no
improvement.
Many digital
music master recordings had originally
been made using 48kHz sampling (24,000
Hertz high frequency response limit)
at 24Bit depth. If the original CD "Redbook"
standards had been established thus,
we could have started out with high
resolution sound. But the 44.1kHz/16Bit
format was dictated by the need to have
the disk player fit into a standard
automobile dashboard cut-out, which
required a 118mm (4.6 inch) disk and
to be sure that Beethoven’s Ninth
Symphony could fit on a single disk,
which was at first was limited to just
74 minutes of music. Of course in no
time the technology was stretched beyond
"Redbook" standards to allow
up to 83 minutes on a single disk, so
we could have had the 48kHz after all,
but by then it was too late.
With DVD-Audio,
a single 118mm inch disk can hold an
entire opera at 48/24 resolution, but
so far the commercial disks have instead
been partitioned to allow a regular
CD length program to be presented in
several different formats, perhaps in
imitation of the hybrid SACD, and sometimes
also including brief video programs
as well. Since no DVD can be made to
play on a regular CD player, DVD-Audio
disks are made to be "compatible"
with home DVD players instead, of which
there seem to be a sufficient number
now anyhow. And delay and controversy
over formats has allowed formats to
proliferate so that a dedicated DVD-Audio
player must test each disk for dozens
of data arrangements and can take up
to three minutes to initialise, whereas
my SACD player chirps merrily for three
seconds and is ready to go, no matter
how many buttons I’ve pushed.
DTS stands
for "Digital Theatre System"
and as the name implies was invented
for surround sound films. It offers
superior accuracy in surround sound
and is often the best sounding track
on any disk that offers it. Many home
DVD players are labelled "DTS"
but watch the fine print: if it says
"digital output" it means
that an accessory DTS decoder must be
connected to the digital output jack;
without the additional expense of that
decoder you get only silence if you
select that option.
Sometimes
the surround channel tracks on SACD’s
and DVD-Audio’s are actually derived
electronically from the two track recording.
Your player may be able to do this better
than the engineer at the studio. If
you are disappointed with the sound
quality in "Surround" setting,
try playing the two track version on
one of your player’s enhanced surround
settings. Sometimes this will result
in superior sound. Some players have
the licensed "Dolby Digital"
or AC-3 setting, others have it only
for DVD video playback, and some players
have proprietary systems which are to
all practical purposes identical to
the Dolby Surround system in all but
name. You won’t know ‘till you try.
There has
been talk of "watermarking"
high resolution audio disks, but despite
advertising claims no one has been able
to work out a means of achieving the
aims of watermarking without degrading
the sound quality unacceptably. At the
best, watermarking would not make piracy
impossible or even difficult, but would
make it a little easier to detect and
prosecute. At the worst all the advantage
gained by moving to high resolution
sound would be lost. Do not buy any
labelled watermarked audio disk without
listening to it first on high quality
playback equipment to be sure it is
listenable. If you buy a disk which
has degraded sound quality and you discover
or suspect it may be watermarked, return
it to the store at once and demand a
full refund of your money.
SACD’s generally
play nothing on your television screen
but the standard screen-saver and text
track information, but when playing
DVD-Audio’s you often get menus with
colourful scenes which change with the
selection, after you visit the audio
set-up menu to select your preferred
manner of playing. Most DVD-Audio disks
sound best on the DTS setting when played
on a standard DVD player, but there
are exceptions to that rule; try playing
any disk in all the modes it offers
to find the optimum presentation.
Paul
Shoemaker
25 March 2004
A lengthy Postscript
Feb 2005 S
When most CDs are labelled
ADD, that means that the manufacturing
of the CD has begun with the analogue
signal produced by the playback of the
original session master tape. This signal
began with the sound pressure wave in
the air at the recording session which
caused a microphone diaphragm to move
in a replica of this pressure wave. That
motion caused to be produced an electrical
signal, a replica of the movement of the
microphone diaphragm. That electrical
signal was amplified electronically (for
now we will assume that this amplification
was perfect and without flaw) and then
was used to produce a time-varying magnetic
field in a recording head. A mechanical
tape transport device presented a moving
magnetically susceptible tape to this
record head and received a replica pattern
of magnetisation corresponding to this
time-varying magnetic field in the record
head. To begin any further processing,
this tape is moved by a transport mechanism
by a playback head and the varying magnetic
field in the tape generates a varying
magnetic field in the armature of the
playback head which in turn generates
an electrical signal in the wire coils
of the playback head.
At this point, the original
sound wave pattern has been remapped six
times and subjected to two mechanical
processes. The laws of physics decree
that at each remapping there is an increase
in noise, a distortion of the wave form,
and a blunting of transient sounds. At
each mechanical process, wow and flutter
are added. That the resulting sound is
as good as it is is a tribute to much
hard work, many years of engineering successes,
some dramatic breakthroughs, some careful
finessing of myriad tiny details.
The subsequent digitisation
of the sound wave replica is accompanied
by some noise and distortions of different
kinds from the analogue processes (digital
information theory is as spooky as quantum
theory and cannot be understood by mere
mortals), but once this is accomplished,
there are no more significant distortions
between this point and the digital electrical
signal produced in the CD player. It has
been said that CDs potentially give us
a personal copy of the original master
tape recording, a potential which has
only been substantially realised by the
SACD and DVD-Audio formats.
(There are people who
believe that analogue recording processes,
even in succession, result in a Divinely
perfect living imprint of the original
sound wave, whereas digital sound processing
is a Satanic perversion which forever
kills the life of the music. We have freedom
of religion in Western nations, and this
is not the place for a discussion of this
opinion, or testimonials to Holy Martyrs
to the Faith. From the scientific point
of view, both digital and analogue sound
processing can be meaningfully described
in terms of noise level, resolution, and
distortion level. Neither is perfect,
both are flawed, and the result is as
exactly as good as the care and money
you invest in your project.)
In the case of recordings
restored to CD from manufactured disk
recordings, the journey is longer. The
signal from the playback of the master
tape recording first goes to a disc cutter
head wherein a varying magnetic field
is produced which causes an armature with
an attached cutting stylus to move in,
it is hoped, a replica of the original
sound pressure wave. A mechanical transport
device moves a vinyl surface by this vibrating
cutting stylus which digs a meandering
groove in the vinyl. Later this vinyl
disk, or a contact replica of it, is moved
under another stylus, this one a polished
diamond, and this stylus moves in a replica
of the sound pattern in the groove, generating
a varying magnetic field which in turn
generates a varying electrical signal.
Now, we are ready to begin digital processing
for the production of our CD, after six
more remappings of the original signal
as received from the master tape. Three
additional mechanical processes have further
increased the wow and flutter levels.
Even if the processing is digital from
hereon out, calling this disk an ADD is
probably not accurate. It probably should
be called an AADD disk.
During the high 78 rpm
era, the nineteen teens through forties,
records were routinely duplicated simply
by playing the disk into disk cutting
equipment. The quality of the result was
horrendously bad, but there was no other
way. In the analogue tape era, roughly
the fifties through the eighties, master
tapes were routinely duplicated onto other
tapes for various reasons, such as: 1.
Security copies in the event that the
original deteriorated, was lost, or damaged
in handling. 2. Early tape splicing tape
was only designed to last long enough
to duplicate the tape onto a new splice-free
"master" after which the first
tape, bearing the rapidly deteriorating
splices, was discarded. 3. Unscrupulous
employees would duplicate a master tape,
keep the original and leave the duplicate
in the vault. In a recent project to find
"original master" tapes, it
was found that due to careless storage
and careless labelling it was not always
possible to determine which of several
tapes was actually the original master.
Clearly, if they couldn’t tell, it shouldn’t
matter, but the result should honestly
be labelled "A[A?]DD."
One is always amazed
at the accuracy of these remappings of
the original signal, how low is the accumulating
noise level, how bright remain the transients,
considering what has been done to them.
In one instance I have directly compared
the wave form produced by an ADD CD transfer
made from the original master tape against
an "ADD" transfer made from
manufactured LP disks using modern restoration
methods and found no visible difference,
and no easily audible difference. That
it can be so is a marvel, a wonder, for
which we should be deeply grateful.
|
Error processing SSI file
|