Reprinted from the Journal of the AES, vol.
46, pp. 458-465, 1998. Click here for a pdf copy
of the original AES article.
Kees Immink presents a research engineer’s view
of the years leading up to the launch of the CD, and the various crucial
decisions that were made during the time, which could determine the technical
success or failure of the medium. From a personal perspective he comments on
the history of the first mass-market digital audio medium and delivers some
predictions for the future of optical media in data storage.
A bit of history
I should say at the outset that I am not a
historian but an engineer, and that what I will say comes from my personal
recollections of the meetings that took place over 15 years ago between
engineers at Philips and Sony. Even so, like the best historians, I shall give
you a brief summary of the 100 years of disk recording that led up to the
launch of the CD. The equivalent storage density of recording media in bits/mm2
grew gradually from that of the original Edison cylinder to the recent DVD. We
are now dealing with storage densities some one million times greater than
those available at the end of the last century. One must remember that it was
at least l0 years after Edison's cylinder, before Berliner and others came
along with a method of disk recording - in fact I visited Edison Labs and was
told that all recording devices up until about 1920 were effectively prototypes
and that there was no way of replicating media, so to have a recording 'hit' at
that time must have been quite an achievement. I think in the 50s, the stereo
LP record delivered a major step in sound quality, but as far as optical
recording is concerned everything before 1969 must be called prehistory.
Optics and Laservision
In 1969, Philips’ researcher Klaas Compaan
suggested that it might be worthwhile developing the use of optics for storing
pictures. They threw around the idea of a spiral track with bits or pits, and
from that time it took six years before Laservision videodisc was born. I
started in 1973 working on servo systems and electronics, and we considered
that if one could store video with audio information optically then one could
also manage audio only. The research management at the time dismissed this idea
saying that audio was far too simple, and not worth the effort, so we, at
research, left it alone for the time being. And then came the launch of
Laservision in 1975, which was a monumental flop. It flopped hugely, and of
some 400 players that were sold some 200 were returned because buyers had been
under the misapprehension that it would also record programmes - of course it could
not. After two years the Philips management decided to throw in the towel, and
withdraw it from the market.
Early digital audio
It is my view that digital audio was made possible at that point, and between
1970 and 1980 numerous meetings were held between Philips and Sony. In order to
understand the technical choices concerning sample resolution and such at that
time one must look to what was happening with studio equipment and with digital
audio in broadcasting. The BBC in the UK was one of the first to adopt digital
sound, and they were working with a 13-bit distribution system for the
transmitter network, sampled at 32 kHz. Stockham's experiments in 1972-3 with
the Soundstream system involved sample rates averaging around 40 kHz, and he
used computer tape as the means of storage, which made it practical to store
data at 16 bit resolution because of the byte-orientated nature of computer
data storage. 8 bits would have been too few, and 14 bits would have meant
inefficient use of the storage space, so 16 bits seemed logical. Computer tape
was really the only way of storing perhaps one hour of sound at the time.
Towards the end of the 70s, (PCM) adapters were developed which used ordinary
analogue video recorders as a means of storing digital audio data, since these
were the only widely available devices with sufficient bandwidth. This helps to
explain the choice of sampling frequency for the CD, because the number of
video lines, frame rate and bits per line end up dictating the sampling
frequency one can achieve if wanting to store 2 channels of audio. The sampling
frequencies of 44.1 and 44.056 kHz were thus the result of a need for
compatibility with the NTSC and PAL video formats used for audio storage at the
time.
An optical audio disc
From 1973 to 1976 two Philips engineers were
given a mandate to develop an audio disc based on optical videodisc technology,
and they started by experimenting with an analogue approach using wide-band
frequency modulation. The problem with this was that it was not really much
more immune to dirt and scratches than an analogue LP, although there was a
certain improvement in sound quality, so they decided to look for a digital
solution. In 1977-8 Philips and Sony both demonstrated the first prototypes of
a digital sound system using a laser videodisc, and, then, in 1979 a crucial
high-level decision was made to join forces in the development of a world audio
disc standard. Philips had lost the market for Laservision, but had
considerable optical expertise, as well as expertise in servo systems and
digital and analogue modulation systems. Sony's huge expertise digital
techniques, such as error correction, PCM adapters and channel coding would
complement this ideally. A reasonable summary would be that the contributions
were complementary: the videodisc 'physics' was provided by Philips and the
digital audio experience by Sony.
The Sony-Philips liaison
In 1979-80 a number of meetings were held in both Tokyo and Eindhoven. The
first, in August 1979 in Eindhoven, and the second in October 79 in Tokyo,
provided an opportunity for engineers to get to know about each other and to
discover the main strengths of each team. There was a great deal to learn from
each other. Both teams had working prototypes and decisions had to be made concerning
modulation and error correction systems. Parallel experiments were conducted in
both locations, and naturally there were numerous occasions on which each team
felt it had the best or most practical solution - either theirs could correct
the longest burst errors or give the longest playing time, and so on. The
delicate prototype electronic equipment had to be transported with the
engineers across the world, and consequently travelled first class in a
separate seat booked especially for it. The airline KLM really loved us for
that, because, as you know, boxes of electronics do not drink champagne or ask
for more food! By May 1980 almost everything was in order. The modulation
system was still a point of contention, with each party still claiming one was
better than the other, and then there was a phone call from the current
chairman of Sony, Mr Ohga, who told us that if we could not make a decision
within a week then management would make it for us. I was tempted to say that
if they were able to make a decision so easily then why did they not do so six
months ago? but refrained in the interests of diplomacy. So we moved quickly
and made all the decisions concerning the mechanical specification of the disk
and so forth.
Deciding the parameters
The disk diameter is a very basic parameter,
because it relates to playing time. All parameters then have to be traded off
to optimise playing time and reliability. The decision was made by the top
brass of Philips. 'Compact Cassette was a great success', they said, 'we don't
think CD should be much larger'.
As it was, we made CD 0.5 cm larger yielding 12
cm. (There were all sorts of stories about it having something to do with the
length of Beethoven's 9th Symphony and so on, but you should not believe them.)
The sampling frequency of 44.1 kHz was decided. In fact it was the only choice
we could make since there was, at that moment, no equipment available that
offered a different sampling frequency but 44.056 kHz. After a long
deliberation we decided 44.1 kHz as opposed to 44.056 kHz simply for the reason
that it was easier to remember - there was really no other reason. Sony made
the excellent choice of 16 bit resolution, although Philips had developed a 14
bit D/A converter at the time, leading Philips to argue initially that it was
impossible to redesign its converters for 16 bits in a short enough time. But
my colleague Karel Dijkmans said 'no problem, I know a small trick to turn a 14
bit converter into a 16 bit converter - it's called oversampling', so we managed
to solve that problem quite easily s it happened, and all the first Philips
players had oversampling converters in them as a result.
The Cross lnterleaved Reed-Solomon code (CIRC)
chosen was much better than that proposed by Philips, although extremely
complicated at the time. Sony was proposing using 16kbyte RAMS for interleaving
which would have cost around $50 to us, and added significantly to the
commercial cost of the players. 'Are they crazy, those Sony guys?' asked many
at Philips, but in fact the price of RAM and its capacity have changed so fast
that it is impossible to buy less than about 1 Mbytes RAMs today, and for less
money than 16k then - so fast have we moved. So that turned out to be a good
decision. The 8-14 bit channel code was agreed and all the specifications
between them led to a playing time of 75 minutes. Geometry and the other
physical parameters were decided, and a plastic layer covering the pits was
added (a basic paradigm of all optical disks) to protect the data surface from
damage and to ensure that dirt and scratches on the surface were well out of
focus for the laser pickup . (I should note that the recent DVD has lost a
small part of this paradigm, as the cover layer is only 0.6mm thick, and so the
optical pickup will be more vulnerable to dirt and scratches.)
The rise and rise of CD
Compact Disc was born, and the audio disk market shows a gradually declining
sale of LP and an exponential increase in sales of CD. The sale of players also
climbed sharply.
Personally, I was not at all sure that CD would
succeed, as I had seen the problems that arose with Laservision, but I am happy
to see that it is expected some 5 billion CDs will be sold in the in the world
in 1997, which makes it a very successful market. (I saw, walking to the
Convention Hall, that 100 billion hamburgers would be sold in the US this year,
so perhaps it is not so big after all...)
Now we have numerous other 'books' in the CD
standard, with CD-ROM players outselling CD Audio by about 70% to 30%, which is
something I never expected.
Looking forward
As a researcher I like to look forward as well
as back, and I should like to make some predictions for the future of optical
disk media, based on DVD. This new disk has a capacity some 7 times larger than
CD, and we need to look at this to see how the optical disk growth path will be
defined. A number of different characteristics have led to the increase in
capacity of the DVD compared with the CD. The wavelength of the laser and the
numerical aperture (NA) of the objective lens have both been altered, and new
research is making further advances possible in this area.
The greater NA of the lens is only possible if
the disk is made thinner, so future improvements here will only be possible if
the disk is made thinner and thinner. Blue and green lasers are already
possible. It will be appreciated that a large part of the potential increase in
capacity is purely dependent on physics, not on recording specialists. Other
advances in capacity have already been made for DVD. Some things can only be
done once - for example removing the 3rd layer, CIRC and removing subcode.
Reducing the track pitch and other physical margins has been possible because
we now know how to manufacture and read optical discs well. The main parameters
of the DVD compared with the CD can be seen in Table I. This leads me to make
some predictions to the year 2010, concerning the physical density of various
storage media. The recording density is compared with CD, which has a density
of 1 bits per square micron compared with the 6-7 bits per square micron of
DVD. High density forms of DVD will extend this further, so that in 2002
(probably in May!) we shall see a 20 Gbyte disk, and in 2006 (almost certainly
in June...) we shall see a 40 Gbyte disk. Meantime the growth in DRAM and
magnetic disk capacity continues to grow at an exponential rate, although the
price of DRAM is currently very much higher than that of optics. Whether
companies will market these products is another matter entirely - this is just
a prediction made by an engineer. So that is the end of my story, and of course
it has been biased.
I have told the story from an engineer's point of view. Many
of the decisions concerning the CD were made by very clever marketing people -
for example the idea of using a 'jewel case' for storing the disk. Those
decisions are not the responsibilities of engineers, for very good reasons.
