References and Further info
Object Lessons: Compact Disc - Robert Barry (Bloomsbury, 2020, ISBN: 978-1-5013-4851-8)
The history of the CD - The CD family
Philips Honoured for Invention of Compact Disc - Frank Beijen (Trouw, 2009)
Compact Disc Story - Kees Schouhamer Immink (Journal of the Audio Engineering Society, May 1998)
Origins and Successors of the Compact Disc: Contributions of Philips to Optical Storage - J.B.H. Peek, J.W.M Bergmans, J. A. M. M. van Haaren, Frank Toolenaar, S.G. Stan (Springer Netherlands, 2009, ISBN: 9781402095535)
https://dutchaudioclassics.nl - Great website cataloguing Philips early work on the CD.
OPTICAL CD CODE - www.laesieworks.com
http://cdfs.com/index-cdfs.html
Anatomy of a Compact Disc
The 1960s saw two breakthrough ideas in sound recording. The idea of
using lasers instead of a needle to read recordings off a surface and recording
such sounds digitally via computer. A number of attempts were made in making a
laser-read record throughout the 1970s, with success finally achieved first
with the LaserDisc video format in 1978. But it wasn’t until 1982 that the
technology came in the form that set the template for later optical disc
formats – the Compact Disc. The 12cm diameter plastic disc proved to the be an
ideal size for a recording medium, to the point that anything smaller looked
“inferior.” Initially made to record high-quality digital sound, the CD became
an all-purpose data recording medium, storing (digitally) images, text,
computer programs, and (later) video. Over time their storage capacity
increased, until a point they became completely different formats. But their
basic structure and fundamental function remained the same. How data is
recorded on a CD is the same on a DVD, Blu-ray and other CD-based formats. So,
dissecting a CD covers most optical disk technology in a single nibble.
Born – 1st
October 1982
Price of first discs – 3,500-3,800 Yen
Size – 12cm wide
x 1.2mm thick
Rotation Speed –
586-228rpm
Audio Quality –
44,100Hz 16-bit stereo
Maximum Record Time (initially) – 74 minutes / 650MB
Laser Wavelength – 780 nm
Play Direction – centre to outside, anti-clockwise
What’s it made of?
A compact disc is a 12cm (5in) diameter, 1.2mm (1/16 in) thick, polycarbonate plastic disc, backed with reflective aluminium about 40-80 nm thick, coated with lacquer to protect it and to provide a surface for labelling.
Reflective Layer
Protective Lacquer
Injection-moulded Polycarbonate Substrate
Reflective Layer
The reflective layer on most CDs are made of aluminium, as its cheap. But its not ideal. Other metals can be used, such as silver or gold, which offer better reflectivity (reducing read errors) and is almost immune to “disc rot” (corrosion of the reflective layer). Aluminium also has the habit of getting hotter during use, which speeds up their “rot.”
Centre Hole
"The fastest decision in the development phase was about the diameter of the hole in the CD. I put a dubbeltje [a Dutch 10 cent coin] on the table and that was the size." – Joop Sinjou, head of Philips audio products (2009)
The 15mm hole, based on the dubbeltje, has since become a standard in optical disc technology, found DVDs, MiniDiscs, HD-DVDs and Blu-rays.
The “Groove”
The data on a CD is etched on its plastic surface in the form of a spiral groove, just like a record. But unlike a record, this groove is non-continuous, forming a collection of “pits” (bumps on the laser reading side) along the “land” (surface) of the disc. A laser reads this groove inside-to-out and the pits affect the laser, by “dimming” it when it hits a pit. Because of geometry, a CD is spun fast (586rpm) while playing initially, and slows down as it plays to the final track (to 228rpm).
This very thin set of four white lines mark (to
scale) four “grooves” on a CD
CD Audio
The sound recorded on CD is sampled 44,100 times per second as 16-bit long
binary numbers that record the loudness of a sound. This is recorded in two
channels for stereo sound, but the CD standard allowed for 4-channel
quadraphonic sound recordings in its first decade. A rare number of early CDs
were re-releases of quadraphonic records, recorded using methods that
compressed their four channels into two stereo tracks. With a special decoder,
a CD player can be made to play these 4-channel sounds as intended. It never
caught on and the 4-channel CD idea was abandoned by the mid-1990s. The two (or
four) channels are
compressed into one stream of data that are recorded as pits on the disc. This
is uncompressed again by electronics when the disc played.
Pit Sizes
The pits in a CD are around 100 nm deep and around 500 nm wide. They can vary in length from 833 nm to 3563 nm. The “grooves” are 1600 nm apart. The pits are made to be a quarter of the laser’s wavelength high. The idea is that when the laser’s light reflects off the CD’s land, the reflection takes a longer journey (by half a wavelength) to the sensor than light reflecting off a pit. Been half a wavelength slower, destructive interference with the initial light from the laser cancels it out, so no light reaches the sensor.
To allow enough intensity change to happen when the laser reads the disc, the space between pit/land boundaries is a minimum of two binary 0s wide and a maximum of 11 0s wide.
Recording Time
The first CDs had a total recording time of 74 minutes. This is due to (as the story goes) a Sony executive been a Beethoven fan. Originally 1600 nm, the “track pitch” of a CD can be reduced a bit to increase recording time. In 1988, a Mission of Burma compilation album was the first CD released with a record time of over 80 minutes. Most CD pressing plants are sicked to a maximum of 80 minutes, as any further may prove unreadable by some CD players. But some CD releases have pushed it to almost 90 minutes!
60 grooves on a CD can fit into one groove on a vinyl LP.
40 grooves on a CD equals the width of a human hair.
Wavelength of CD player laser (780 nm)
What is on an Audio CD?
A CD is not just a physical plastic disc. It is also the
data encoded on its surface. A CD is basically a record with a nanoscopically
thin groove that is read continuously by a laser. This groove is broken up to
form a pattern of “pits” and “lands” that encode a binary number sequence,
which encodes its digitized audio and additional information used by the CD
player. This is read in 588-bit long chunks, or “information frames.” Each of
these frames not only contain the data, but also additional information that identifies
what kind of data it is. That additional information breaks up the single
stream of binary numbers to form the multiple tracks that make up an album or
(later) individual files on a CD-ROM.
Information Frames
The data on CDs is recorded on the disc in the form of
588-bit long “information frames.” Each of these information frames is made up
of a “frame” of 33 17-bit “symbols,” which contains the audio (and one “control
word” symbol), and a 24-bit “sync word,” with three “merging bits.”
Sync Word
The sync word tells the CD player where each information
frame starts. Also, its frequency of 0s and 1s are used to tell the CD player’s
motor how fast to spin.
Symbols and Frames
The data recorded on a CD begins life as a collection of
8-bit long “symbols” which are then recorded in 32 symbol long “frames.” These
symbols are then converted into 14-bit long symbols before been recorded on the
disc. These 14-bit symbols then go through a process that converts them into
17-bit long symbols, which becomes the signal that is recorded as pits on the
disc. Why? It allows it to be encoded on the disc in a least error-prone way as
possible.
Control Words and Subcodes
An 8-bit “control word” is added to each frame. This control
word can encode up to eight channels of additional information, or “subcodes,”
that can be added to the data stored on a CD. Each subcode “word” is 98 bits
long, so to read one word, the player has to read 98 frames. These subcodes are
called P, Q, R, S, T, U, V, and W. Early audio CDs only used P and Q. Later on,
some CDs made use of the extra channels, storing text information about the
music (on CD-Text disks) or low-resolution graphics (on CD+G disks).
P Subcode
The P subcode is a simple code understood by the simplest of
CD players. It’s mostly a sting of 0s, except for two seconds worth of Is that
mark out the start of each music track. In the lead-out track, it becomes an
alternative 0 and 1 to create a 2Hz rhythm to tell the player the CD has ended.
Q Subcode
The Q subcode encodes what kind of audio it is recorded (is
it stereo?) and additional information about the audio, such as track number,
how long it is, etc. It comes in three modes.
Synchronizing Bits
These two bits are synchronizing bits, 0s and 1s. They tell
the decoder that this is a control word, not audio.
Control Bits
These 4 bits indicate the number of audio channels (two or
four) and if “pre-emphasis” (a noise reduction method) is used in the audio. This
was rarely used, as pre-emphasis was added as an option to the CD standard when
it was intended to use 14-bit audio, instead of the 16-bit it finally used. Later
on, this section is used to indicate if the information is data or audio and
tell if the CD is copy-protected.
0000 – 2 channel/ no pre-emphasis
1000 - 2 channel/with pre-emphasis
0001 - 4 channel/ no pre-emphasis
1001 - 4 channel/with pre-emphasis
Address Bits
These four bits tell the decoder which of three “modes” the
following 72 bits of data is in.
CRC Bits
The final 16 bits is Cyclic Redundancy Check (CRC) error
correction code.
Mode 1 - 0001
Mode 1 is the most important piece of subcode, as it is used
during the playing of audio CDs. At least nine of 20 consecutive subcode words
must be in mode 1. It is encoded in two forms, depending on if it’s in the lead-in
track or not.
Track Number
This is this track’s number, encoded in the form of 2 four-digit
binary numbers. Up to 99 tracks can be encoded on a CD, from “01” to “99.” But
if you include the possibility of hexadecimal digits, that can increase to 255
tracks. But hex digit “A” is reserved to indicate the first track with audio
(“A0”), the last track (“A1”) and the start of the lead-out track (“A2”) In the
Lead-in track, it is “00.”
Table of Contents
In the lead-in track, the following 8 bytes of data, especially
the final three, encode the CD’s “table of contents.” This subcode is what a CD
player reads first, to know where each track is on a disc. This information is
repeated three times. Then its repeated again and again, until the CD player is
satisfied it’s got the complete table of contents.
Point – indicates the track number, according to the table
of contents
Minute/Second/Frame – This is the running time of the
currant track, encoded by two 4-digit binary numbers, from “01” to “99.” This
clock ticks up from “00” as a track plays. But when it encounters a pause, the
clock ticks downward to “00,” which indicates the end of the pause. In the
lead-in and lead-out tracks, this clock ticks up from “00.”
Zero – Just eight 0 bits.
P Time – This is the starting time of track (recorded in
Point) in minutes, seconds and number of “frames” (blocks), encoded as two
4-digit binary numbers, for the table of contents.
A CD’s playing time is counted in blocks of subcode. 75
blocks (or 7,350 frames) = one second.
In music and the lead-out track, the following 8 bytes of
data encode the following…
Point – An index number within track, encoded by two 4-digit
binary numbers, from “01” to “99.” If Point equals “00” it indicates a pause in
between tracks.
Time – This is the running time of the currant track, in
minutes, seconds and number of “frames” (blocks), encoded by two 4-digit binary
numbers, from “01” to “99.” This clock ticks up from “00” as a track plays. But
when it encounters a pause, the clock ticks downward to “00”, which indicates
the end of the pause. In the lead-in and lead-out tracks, this clock ticks up
from “00.”
Zero – Just eight 0 bits.
A Time – This is the running time of the whole disc, in
minutes, seconds and number of “frames” (blocks), encoded by two 4-digit binary
numbers, from “01” to “99.” This clock is set to “00” at the start of the programme
area.
Mode 2 – 0010
Mode 2 is only used by CD manufacturers, as it encodes the
disc’s catalogue number, in the form of 13 4-digit binary numbers, recording
the number according to the UPC/EAN standard used in barcodes. After this
number there is a string of 12 0s. If used, at least one of 100 successive
subcode words must contain it.
Mode 3 – 0111
Mode 3, like Mode 2, if used, at least one of 100 successive
subcode words must contain it. Mode 3 is used to assign each selection with a
unique number, according to a 12-character International Standard Recording
Code (ISRC), defined by DIN-31-621. This is used to encode information about
the music on the track. This information can only change when the track number
is changed. After this code there is a string of four 0s. Mode 3 is not used in
lead-in and lead-out tracks.
The first five characters are encoded by five 6-bit binary
numbers.
Two 0s separate character 5 and 6.
The final seven characters are encoded by seven 4-bit binary
numbers.
Characters 1-2 – country of origin
Characters 3-5 – Record company
Character 6-7 – Year of recording
Characters 8-12 – Recording’s serial number
This graph helps explains the actions of the subchannels on a CD graphically. The following graph demonstrates the actions of the subchannels on an audio CD with four music tracks, where track three and four transitions by a fade, instead of a pause.
Point
These are index numbers within track. If Point equals “00”
it indicates a pause in between tracks. They can be used as a “Sub-Track
Number” so a listener can skip to a certain part of a track. This feature is
mostly used in professional equipment, like that used in radio stations, but
rarely found on consumer devices. Every track has an “index 01,” but …..
Pre-Gap and Hidden Tracks
…. Some CDs make use of “index 00” and use this section to
record data or “hidden tracks.” Computers can read this section, while normal CD
players ignore it, allowing “mixed mode” audio CDs to include some software.
The Enhanced CD format (introduced in 1995) was created to end such practise. If
it contains audio, it can be accessed by a CD player by playing track 1 and reverse
seek it back to the actual start of the track or letting the CD play without skipping
a track. Hidden tracks can also be marked out with “index 02,” “index 03,” and
so on.
Time
This is the running time of the currant track. This clock
ticks up from “00” as a track plays. But when it encounters a pause, the clock
ticks downward to “00”, which indicates the end of the pause. In the lead-in and
lead-out tracks, this clock ticks up from “00.”
A Time
This is the running time of the whole disc. This clock is
set to “00” at the start of the programme area.
Background
The background is made up of lines of 588 binary numbers,
representing the “information frames” that are read and processed off a CD. The
green section marks off the 24-bit “sync word.” The black line is the three “merging
bits” between the sync word and the “control word.” The red section marks off
the “control word” and altering shades of grey mark off each of the 32
“symbols” that encode the audio.
The blue covers half of the 196 lines of information frames that cover this spread, marking out the 98 frames that make up one subcode word.
Recording Sounds on a CD
Most explanations on how CD players work only talk about how
the pits and lands on the disc are read by the laser. That’s simple to
understand, because its similar to a stylus reading a record. Few cover what
happens next … or how sound is converted into the pits on the disc in the first
place. Most explanations describe it as follows. The pattern of light detected
by the photodiode is translated into a digital code. This code is processed and
translated into a varying electric current, which is then sent to the amplifier
and speaker(s). However, what happens to this digital code before it gets
converted into varying current is …. Complicated, but vital. A key part of this
code processing (and how the CD became a successful audio format) is an error
correction system called Cross-Interleave Reed-Solomon Coding (CIRC in short). The
CD was the first form of consumer technology to use such strong error
correction methods. CIRC can correct errors up to 4,000 bits long or insert a
correction in a string of errors 12,300 bits long. That translates to little
issue playing a CD with marks or scratches up to 2.5mm long, only really having
serious issues with obstructions over 7.7mm long.
To simulate how sound is recorded
on a CD we are going to simplify what is going on by recording the words
“sounds.” As CDs deal with stereo, we are going to record it twice, in
uppercase and lowercase letters, to simulate two channels of audio. The letters
are encoded in 8-bit ASCII. Actual audio would be sampled in 16 bits 44,100
times per second.
1.
Each 16-bit sample is split into two 8-bit
“symbols.” In our simulation, each 8-bit “sample” is split into two 4-bit
“symbols.
2.
The symbols are then fed through the CIRC
encoder and stored in some RAM memory, 24 symbols at a time (six 16-bit samples
in real life, remember).
2.1. In
the CIRC encoder, the first thing that happens to the six samples is
“scrambling.” Sample two, four, and six are delayed by a two-symbol long jam,
while the others got through without hinderance. The samples that do reach the
next step are then subject to a jumble of wires that scramble their binary
digits.
2.2. The
C2 encoder adds four 8-bit error-correcting “Q words” between the symbols, making
an output 28 symbols.
2.3. Each
of these 28 symbols are then subject to another delay. How long this delay is
depends on where in the sequence the symbol happens to be at. These 28 8-bit
delay lines are all of different lengths, designed to scramble the binary bits
that make up the 28 symbols in a determined pattern.
2.4. After
this second scrambling, the C1 encoder adds four 8-bit error-correcting “P
words” among the symbols, making an output of 32 symbols.
2.5. The 32 symbols go through another scrambling set up. Like what happened to the original six samples at the beginning, every other symbol is subject to a delay, while the others are not. Also, the P and Q words are inverted. The result is a 32 8-bit symbol long “frame.”
3. An 8-bit “control word” is added to each frame. (4 bits in this demonstration)
4. After the control words are added, each 8-bit symbol is
subject to eight-to-fourteen modulation. Basically, the 8-bit sequence is
converted to a 14-bit sequence. Why? To reduce data bandwidth during processing
(the CD was invented in 1982, remember), reduce its DC content, * and it adds additional
synchronization information to the data. The resulting 14-bit codewords
represent all 256 possible combinations of an 8-bit binary sequence. Why use
such a code? By using a code that uses low numbers to code for sequences with
fewer 0 to 1 transitions between long strings of the same bit, you can reduce
the bandwidth needed to process them. Also, with a limit to logically possible strings
of 1s, it also reduces DC content. *
5. After EFM, an additional three “merging bits” are added
to every 14-bit symbol. They further reduced DC content.* Their values depend
on the adjacent symbols.
6. he data bits go through a process that converts the
signal from a “non-return-to-zero” wave to a “non-return-to-zero inverted” signal,
where the signal only increases and decreases when there is a binary 1. The
resulting signal, made of 33 17-bit symbols, has a minimum length of three
clock periods and a maximum of 11 clock periods.
7. A “sync word” and its three merging bits are added,
creating a 588-bit “information frame.” Sync words indicate the start of each
frame and their frequency of changing from 0 to 1 is used to control the speed
of the CD player’s motor. The CD is spun fast when playing the first track,
then slows down as it plays along to the last track on the edge of the disc.
It’s these 588-bit long “information frames” that are
recorded on a CD, with each binary 1 marked by a transition from “pit” to
“land.”
How to Make a CD
CDs are basically made in a similar way to gramophone
records, except more precise and in a dust-free environment, similar to those
where microchips are made. This is how a pre-recorded CD is made, from
recording to packaging.
1.
Before recording, a glass disc is cleaned,
polished and coated with a photo-sensitive chemical and dried. The thickness of
this layer of chemical (and the roughness of the glass) determined how deep the
CD’s pits will become.
2.
The recording is burnt onto the glass’s coating
as tiny holes by a laser.
3.
The burnt disc is placed in developing fluid,
which dissolves burnt areas in the chemical coating, exposing the glass to acid,
etching in the pits that’ll be on the final discs.
4.
In a sealed vacuum, the etched side of the
“glass master” is given a metal coating to protect its pits, via an electrically-heated
piece of wire.
5.
In a nickel salt bath, the silvered glass master
is given a layer of nickel through electroforming.
6.
The resulting nickel layer is removed, creating
the “father” – a negative impression of the disc.
7.
The father is used to create several positive
impressions, or “mothers.”
8.
From the mothers are “sons” – negative
impressions of the disc.
9.
The sons are then cleaned through a rinse, then
dried, and given a protective coating.
10.
After processing, the sons have a centre hole
stamped and their edges trimmed, becoming the moulds for the final discs.
11.
The sons are placed in an injection moulding
machine. Molten polycarbonate is injected into the mould, making “green” discs.
12.
In a vacuum chamber, with a bit of argon gas, a
piece of metal alloy is given high voltage electric current, creating plasma,
which gives the CD it reflective coating.
13.
The disc is placed in a spinner as lacquer is
squirted onto it. The spinning ensures an even coat.
14.
After the lacquer dries, the label is
screen-printed.
15.
The CD is tested then finally packaged.
Comments
Post a Comment