
		       CD-ROM Technical Summary
		   From Plastic Pits to "Fantasia"

			     Andy Poggio
			      March 1988


Abstract

This summary describes how information is encoded on Compact Disc (CD)
beginning with the physical pits and going up through higher levels of
data   encoding  to  the structured   multimedia  information that  is
possible with   programs like HyperCard.   This  discussion    is much
broader than any  single  standards document,  e.g. the CD-Audio   Red
Book,   while omitting  much of   the   detail needed  only  by  drive
manufacturers.


Salient Characteristics

1.    High information density --  With  the density achievable  using
optical encoding, the CD can contain some  540 megabytes  of data on a
disc less than five inches in diameter.

2.  Low unit cost -- Because  CDs are manufactured by a well-developed
process similar to that  used to stamp out   LP records, unit cost  in
large quantities is less than two dollars.

3.  Read only medium -- CD-ROM is  read only;  it cannot be written on
or erased.  It  is an electronic publishing, distribution,  and access
medium; it cannot replace magnetic disks.

4.  Modest random access performance -- Due to  optical read head mass
and data encoding methods, random access ("seek  time") performance of
CD is better than floppies but not as good as magnetic hard disks.

5.  Robust, removable medium -- The CD itself is comprised  mostly of,
and completely  coated by, durable  plastic.   This fact and  the data
encoding method allow the  CD to  be resistant  to scratches and other
handling damage.   Media lifetime is expected to  be long, well beyond
that  of magnetic media such as  tape.  In addition, the optical servo
scanning mechanism allows CDs to be removed from their drives.

6.  Multimedia storage -- Because all  CD data is stored digitally, it
is inherently multimedia in that it can  store text, images, graphics,
sound, and any other information expressed  in digital form.  Its only
limit in this area is  the rate  at which data  can  be read from  the
disc, currently about 150 KBytes/second.   This  is sufficient for all
but uncompressed, full motion color video.


CD  Data  Hierarchy

Storing data  on a CD  may be thought  of as  occurring through a data
encoding hierarchy with each level built  upon the previous  one.   At
the lowest level, data is physically stored as  pits on the disc.   It
is actually encoded by several  low-level mechanisms to  provide  high
storage  density and reliable data  recovery.   At the next level,  it
organized into tracks which may be digital audio  or CD-ROM.  The High
Sierra  specification  then defines  a  file system  built  on  CD-ROM
tracks.  Finally, applications like HyperCard specify a content format
for files.


The Physical Medium

The Compact Disc   itself  is a thin plastic  disk    some 12 cm.   in
diameter.  Information is  encoded in  a plastic-encased spiral  track
contained on the top of the disk.  The  spiral track is read optically
by a  noncontact  head which scans approximately  radially as the disk
spins  just above it.   The spiral is  scanned  at  a constant  linear
velocity thus assuring a constant  data rate.  This  requires the disc
to rotate  at a decreasing  rate  as the  spiral is  scanned from  its
beginning near  the center of  the  disc   to its  end near   the disc
circumference.

The spiral track contains   shallow  depressions,  called pits, in   a
reflective layer.   Binary information is  encoded by  the  lengths of
these pits  and  the lengths of  the areas  between them, called land.
During   reading,  a low  power  laser  beam from the optical  head is
focused on the spiral layer and is reflected back  into the head.  Due
to the optical characteristics of the  plastic disc and the wavelength
of  light used, the  quantity of  reflected  light varies depending on
whether  the beam is on  land or on a  pit.  The modulated,  reflected
light  is  converted  to  a  radio frequency, raw   data signal  by  a
photodetector in the optical head.


Low-level Data Encoding

To ensure accurate recovery, the disc data must be encoded to optimize
the analog-to-digital  conversion  process   that the radio  frequency
signal must undergo.  Goals of the low level data encoding include:

1.  High  information density.  This  requires encoding that makes the
best possible use  of  the high, but limited,  resolution of the laser
beam and read head optics.

2.  Minimum  intersymbol  interference.   This  requires making    the
minimum run length,  i.e. the  minimum number of consecutive zero bits
or one bits, as large as possible.

3.  Self-clocking.  To avoid a separate timing track,  the data should
be encoded so as to allow the clock signal to be  regenerated from the
data signal.  This requires limiting  the maximum   run length of  the
data so that data transitions will regenerate the clock.

4.  Low digital sum value (the number of one bits minus the  number of
zero  bits).  This minimizes the low  frequency  and DC content of the
data signal which permits optimal servo system operation.

A straightforward encoding would be to  simply to  encode zero bits as
land and one bits as pits.  However, this does  not  meet  goal (1) as
well as the  encoding scheme  actually  used.   The current CD  scheme
encodes one bits  as transitions from pit to  land  or land to pit and
zero bits as constant pit or constant land.

To  meet goals (2) to   (4), it  is not  possible  to encode arbitrary
binary data.  For example, the integer 0 expressed  as thirty-two bits
of zero  would have too long  a run length to  satisfy  goal  (3).  To
accommodate these goals, each eight-bit byte of actual data is encoded
as fourteen bits of channel data.  There are many more combinations of
fourteen bits (16,384) than there are of eight bits (256).   To encode
the eight-bit  combinations, 256 combinations  of  fourteen  bits  are
chosen   that  meet  the  goals.   This  encoding   is  referred to as
Eight-to-Fourteen Modulation (EFM) coding.

If  fourteen channel   bits  were  concatenated  with  another set  of
fourteen channel bits, once again the above goals may not be  met.  To
avoid this possibility, three merging  bits are included  between each
set of fourteen channel bits.  These merging bits carry no information
but are chosen to  limit run length, keep  data signal DC content low,
etc.  Thus, an eight bit byte of actual  data is  encoded into a total
of seventeen channel bits: fourteen EFM bits and three merging bits.

To achieve  a reliable  self-clocking system, periodic synchronization
is  necessary.   Thus, data is  broken up into  individual frames each
beginning with a synchronization  pattern.   Each frame  also contains
twenty-four data bytes,  eight error correction  bytes, a control  and
display byte  (carrying  the  subcoding  channels), and merging   bits
separating them all.  Each frame is arranged as follows:

   Sync Pattern 24 + 3 channel bits
   Control and Display byte 14 + 3
   Data bytes 12 * (14 + 3)
   Error Correction bytes 4 * (14 + 3)
   Data bytes 12 * (14 + 3)
   Error Correction bytes 4 * (14 + 3)

   TOTAL 588 channel bits

Thus, 192 actual data bits (24 bytes) are encoded as 588 channel bits.

Editorial: A CD physically  has a single  spiral  track about 3  miles
long.  CDs spin at about 500 RPM when reading near  the center down to
about 250 RPM when reading near the circumference.

Disc with a 'c' or disk  with a  'k'?  A usage  has  emerged for these
terms: disk  is used for  eraseable disks (e.g. magnetic  disks) while
disc is used for read-only (e.g. CD-ROM discs).  One  would presumably
call a frisbee a disc.


First Level Error Correction

Data  errors can  arise  from  production defects  in the disk itself,
defects arising from subsequent damage to the disk, or  jarring during
reading.  A  significant characteristic of these errors  is  that they
often  occur in long  bursts.  This could  be  due,  for example, to a
relatively wide mark on the disc that is opaque to the laser beam used
to read  the disc.  A  system with two  logical  components called the
Cross  Interleave Reed-Solomon  Coding  (CIRC)  is employed for  error
correction.   The cross interleave component breaks  up the long error
bursts into many short errors; the Reed-Solomon component provides the
error correction.

As each frame is read from the disc, it is first decoded from fourteen
channel bits (the three merging bits  are ignored) into eight-bit data
bytes.   Then, the bytes from each  frame  (twenty-four data bytes and
eight error correction  bytes) are passed   to the first  Reed-Solomon
decoder which uses four of the error correction  bytes and is  able to
correct   one byte in  error   out   of the    32.   If  there  are no
uncorrectable errors, the data  is simply passed along.  If  there are
errors, the  data  is  marked as  being  in error    at this stage  of
decoding.

The twenty-four data bytes and  four remaining error correction  bytes
are then  passed through unequal delays  before  going through another
Reed-Solomon decoder.  These  unequal delays result in an interleaving
of the data that spreads long error bursts among many different passes
through the second decoder.  The delays  are such that error bursts up
to   450  bytes  long    can   be completely    corrected.  The second
Reed-Solomon  decoder  uses the last four  error   correction bytes to
correct  any remaining errors in the  twenty-four data bytes.  At this
point, the data goes through a de-interleaving process to  restore the
correct byte order.


Subcoding Channels and Blocks

The  eight-bit control  and display byte   in each   frame carries the
subcoding channels.  A subcoding block consists of 98 subcoding bytes,
and  thus  98 of the  588-bit frames.  A block  then  can contain 2352
bytes of data.  Seventy-five blocks are  read each  second.  With this
information, it is now  straightforward to calculate  that the CD data
rate is in fact correct for CD digital audio (CD-DA):

Required CD digital  audio data rate: 44.1  K samples per second  * 16
bits per sample * 2 channels = 1,411,200 bits/sec.

CD data rate: 8 bits per byte * 24  bytes  per frame *  98  frames per
subcoding block * 75 subcoding blocks per second = 1,411,200 bits/sec.

The eight subcoding channels  are labeled P  through W and are encoded
one bit for each channel in a control and display  byte.  Channel P is
used as a simple music track separator.  Channel Q is used for control
purposes  and encodes information  like track  number, track type, and
location (minute, second, and frame number).  During the lead-in track
of the disc, channel Q encodes a table of contents for the disk giving
track number and starting location.  Standards have been proposed that
would use the remaining channels for line graphics and ASCII character
strings, but these are seldom used.


Track Types

Tracks  can have two  types as specified in the  control bit field  of
subchannel Q.  The first type is CD digital audio (CD-DA) tracks.  The
two-channel  audio is sampled   at  44.1  Khz with sixteen bit  linear
sampling encoded as twos complement numbers.   The sixteen bit samples
are separated into  two eight-bit bytes; the bytes  from  each channel
alternate  on   the  disc.   Variations  for   audio  tracks   include
pre-emphasis and four track recording.

The other type  of track specified  by  the subchannel Q  control  bit
field  is the  data track.  These  must conform to the CD-ROM standard
described below.   In  general, a disc  can  have a mix of  CD digital
audio tracks and a CD-ROM track, but the CD-ROM track must come first.

Editorial: This first  level error  correction (the only type used for
CD Audio data) is extremely powerful.  The CD specification allows for
discs  to have  up to 220 raw  errors per second.   Every one of these
errors is (almost always) perfectly corrected by the CIRC scheme for a
net error rate of zero.  For example, our  tests  using Apple's CD-ROM
drive (which also  plays audio) show  that raw error rates  are around
50-100  per  second  these  days.   Of  course,   these  are perfectly
corrected, meaning that the original data is  perfectly recovered.  We
have tested flawed  discs with raw  rates  up to 300  per second.  Net
errors on all of these discs?  Zero!  I would  expect  a typical audio
CD player to perform similarly.  Thus I expect this raw  error rate to
have no audible consequences.

So why  did  I say    "almost always"   corrected  above?   Because  a
sufficiently bad flaw  may produce  uncorrectable errors.  These  very
unusual  errors are "concealed" by  the player rather  than corrected.
Note that this concealment is likely to be less noticeable than even a
single scratch on an LP.  Such a flaw might be a  really opaque finger
smudge; CDs do  merit careful handling.  On the  two  (and  only  two)
occasions I  have found these,  I simply  sprayed  on  a little Windex
glass cleaner and  wiped  it off using radial strokes.   This restored
the CDs to zero net errors.

One can argue about the quality of the process of conversion of analog
music  to and from digital  representation, but  in the digital domain
CDs are really very, very good.


CD-ROM Data Tracks

Each CD-ROM data track is divided into individually addressable blocks
of 2352 data bytes, i.e. one subcoding  block or  98 frames.  A header
in each block contains the  block address and the  mode of the  block.
The block address is identical to the  encoding of minute, second, and
frame number in subcode channel  Q.  The  modes  defined in the CD-ROM
specification are:

   Mode 0 -- all data bytes are zero.

   Mode 1 -- (CD-ROM Data):
   Sync Field - 12 bytes
   Header Field - 4
   User Data Field - 2048
   Error Detection Code - 4
   Reserved - 8
   Error Correction - 276

   Mode 2 -- (CD Audio or Other Data):
   Sync Field - 12 bytes
   Header Field - 4
   User Data Field - 2048
   Auxiliary Data Field - 288

Thus, mode 1  defines  separately addressable, physical 2K   byte data
blocks making CD-ROM look at this level very similar  to other digital
mass storage devices.


Second Level Error Correction

An uncorrected error in audio data typically results in a brief, often
inaudible  click  during listening at worst.   An uncorrected error in
other kinds  of  data, for   example program code,  may   render  a CD
unusable.  For this  reason, CD-ROM defines  a second  level  of error
detection   and  error correction (EDC/ECC)  for   mode 1 data.    The
information for the EDC/ECC occupies most of the auxiliary data field.

The error detection  code is a cyclic  redundancy  check  (CRC) on the
sync, header, and user data.  It occupies the first  four bytes of the
auxiliary  data field   and  provides a  very  high  probability  that
uncorrected errors will be   detected.  The  error correction  code is
essentially the same   as the first level  error  correction in   that
interleaving and Reed-Solomon coding are  used.  It occupies the final
276 bytes of the auxiliary data field.

Editorial: This  extra level of error correction  for CD-ROM blocks is
one of the many reasons  that  CD-ROM drives are  much  more expensive
than consumer audio players.  To perform this error correction quickly
requires substantial  extra computing   power (sometimes  a  dedicated
microprocessor) in the drive.

This  is also  one reason   that consumer  players like the Magnavoxes
which claim to be CD-ROM compatible (with their digital output jack on
the back) are useless for  that purpose.  They have  no way of dealing
with the  CD-ROM error  correction.   They  also  have  no  way for  a
computer to tell them where to seek.

Another reason that CD-ROM drives are more  expensive is that they are
built to be a computer peripheral rather than a consumer  device, i.e.
like a combination race   car/truck rather  than a family  sedan.  One
story, probably apocryphal but not  far from the  truth, has it that a
major  Japanese  manufacturer  tested  some  consumer audio players to
simulate  computer use: they made them  seek  (move the optical  head)
from the inside of the  CD to the  outside  and back again.  These are
called  maximum seeks.  The story says  they  managed   to do this for
about 24 hours  before they broke down.   A CD-ROM  drive needs to  be
several orders of magnitude  more robust.  Fast  and strong don't come
cheap.


The High Sierra File System Standard

Built  on   top of   the   addressable   2K  blocks  that  the  CD-ROM
specification defines,  the  next higher level of  data encoding is  a
file system that  permits logical organization  of the data on the CD.
This can be a native file system like  the Macintosh Hierarchical File
System  (HFS).  Another alternative is  the High Sierra (also known as
the  ISO  9660)  file standard,   recently approved  by  the  National
Information  Standards  Organization  (NISO)   and  the  International
Standards Organization (ISO),  which defines a  file  system carefully
tuned to CD characteristics.  In particular:

1.  CDs  have modest seek time  and  high capacity.  As a  result, the
High Sierra standard  makes tradeoffs that reduce  the number of seeks
needed to read a file at the expense of space efficiency.

2.  CDs  are read-only.   Thus, concerns  like  space allocation, file
deletion, and the like are not addressed in the specification.

For High Sierra file systems, each individual CD is a volume.  Several
CDs may be grouped together in  a volume set and  there is a mechanism
for subsequent volumes in a set to update preceding ones.  Volumes can
contain standard file structures, coded character set  file structures
for character encoding  other  than  ASCII,   or boot records.    Boot
records can contain either data or program code  that may be needed by
systems or applications.


High Sierra Directories and Files

The file system is a hierarchical one in which directories may contain
files or other directories.  Each  volume has a  root directory  which
serves as an ancestor to all other directories or files in the volume.
This dictates an overall tree structure for the volume.

A typical disadvantage in hierarchical systems is that to  read a file
(which must be a leaf of the hierarchy tree) given its full path name,
it is necessary to begin at the root directory and search through each
of its  ancestral directories until  the entry for the file  is found.
For example, given the path name

   Wine Regions:America:California:Mendocino

three directories (the first three components of the  path name) would
need to be searched.  Typically, a separate seek would be required for
each directory.  This would result in relatively poor performance.

To avoid this, High Sierra specifies  that each  volume contain a path
table in addition  to its   directories  and files.   The  path  table
describes the directory hierarchy in a compact form that may be cached
in computer memory  for optimum performance.   The path table contains
entries  for   the  volume's directories  in a  breadth-first   order;
directories with a  common parent  are  listed in lexicographic order.
Each entry contains only the location  of the directory  it describes,
its name, and  the location in  the path  table of   its parent.  This
mechanism allows any  directory to be  accessed with  only a single CD
seek.

Directories  contain  more detailed  information than  the path table.
Each directory entry contains:

   Directory or file location.
   File length.
   Date and time of creation.
   Name of the file.
   Flags:
      Whether the entry is for a file or a directory.
      Whether or not it is an associated file.
      Whether or not it has records.
      Whether or not it has read protection.
      Whether or not it has subsequent extents.
      Interleave structure of the file.

Interleaving may be used, for  example,  to meet realtime requirements
for multiple  files  whose contents must  be presented simultaneously.
This would happen if a file containing graphic images were interleaved
with a file containing compressed sound that describes the images.

Files themselves are recorded in contiguous (or interleaved) blocks on
the  disc.   The  read-only  nature of   CD  permits this   contiguous
recording in a straightforward manner.  A file may also be recorded in
a  series of noncontiguous  extents with a  directory  entry  for each
extent.

The specification does not favor any particular computer architecture.
In particular  all significant, multibyte  numbers are recorded twice,
once with  the most significant  byte first  and  once  with the least
significant byte first.


Multimedia Information

Using  the file   system are applications   that create  and   portray
multimedia information.  While it is true that a CD can store anything
that a magnetic disk can store (and usually much more of it), CDs will
be used more for storing information than for storing programs.  It is
the  very large storage  capacity  of CDs coupled with their  low cost
that   opens   up   the  possibilities  for  interactive,   multimedia
information to be used in a multitude of ways.

Programs  like  HyperCard,  with it's   ease of  authoring   and broad
extensibility, are  very  useful for this purpose.   Hypercard stacks,
with related information such as color images and sound, can be easily
and inexpensively   stored on CDs  despite their  possibly  very large
size.

Editorial: The High Sierra file system gets its name from the location
of the first meeting on it: the High  Sierra Hotel  at Lake Tahoe.  It
is  much  more commonly  referred  to  as  ISO 9660,   though the  two
specifications are slightly different.

It has gotten  very easy  and inexpensive  to  make a CD-ROM disc  (or
audio CD).  For example, you can now take  a  Macintosh hard  disk and
send it with $1500 to one of several CD  pressers.  They will send you
back your hard  disk and 100  CDs  with  exactly the  same content  as
what's on your disk.  This is the easy way  to make CDs  with capacity
up to  the size of your hard  disk (Apple's  go up to  160 megabytes).
True, this is not a full CD but  CDs  don't  need to be  full.  If you
have just 10 megabytes and need 100 copies, CDs may be the best way to
go.

If you are buying a CD-ROM drive, there are  several factors you might
consider  in  making  your choice.   Two  factors NOT to consider  are
capacity   and data rate.   The  capacity  of  all  CD-ROM   drives is
determined solely by  the CD they are  reading.  Though you will see a
range of numbers in manufacturers' specs  (e.g. 540, 550, 600, and 650
Mbytes), any drive can read any disc and so they are all fundamentally
the same.  All  CD-ROM drives read data at   a net 150 Kbytes/sec  for
CD-ROM  data.   Other data   rates  you may  see    may include  error
correction data (not included in the net rate) or may be a mode 2 data
rate (faster than  mode 1).   All drives will  be  the same  in all of
these specs.


End of article.
 