IEEE Communications Magazine • January 201010 0163-6804/10/$25.00 © 2010 IEEE

INTRODUCTION

This article describes how the compactdisc digital audio system, also known asthe CD system, was the result of thesuccessful merger of two major existingtechnologies. They are first, opticalreadout of information stored on anoptical disc of 11.5 cm diameter, andsecond, the digital coding/decoding anddigital processing of audio signals. I willdescribe how a complete digital versionof this new audio system graduallyemerged, within Philips, during theperiod 1973–1979. Finally, this paperrecounts the partnership and collabora-tion between Philips and Sony thatresulted in a common CD standard inJune 1980 which soon after became theworld standard. More details on thehistory of the CD not discussed herecan be found in a lively, and accurate,doctoral thesis (in German) by J. Lang[1] and in a recent book [2].

A LOOK INTO A SUCCESSOR OF THERECORD PLAYER

In the 1970s the high-fidelity (Hi-Fi)record player and its LP (long play)vinyl record became increasingly theweakest product in the audio productrange. Whereas most audio equipmentwas downscaled in size, this was impos-sible for the Hi-Fi player with its largedisc with a 30 cm diameter. Further-more, the LP record was vulnerableand had to be handled with care. Ascratch or accumulation of dust couldspoil the enjoyment of the reproducedmusic.

Aware of these limitations, L.Ottens, the technical director of the

audio industry group within the PhilipsCorporation in Eindhoven, the Nether-lands, started to look for smaller andbetter performing alternatives for therecord player and its LP record. Thesmall size of a compact cassette was anattractive feature. Although its soundquality improved in the 1970s, mainlyby using better magnetic tape-coatingsand companding, it became no betterthan an LP record.

At that time, when Ottens was look-ing for a suitable successor of therecord player, the optics department,headed by P. Kramer, of the Philipsresearch laboratory in Eindhoven, alsoknown as the Nat.Lab. (NatuurkundigLaboratorium), was experimenting witha completely new system for recordingand playback of analog video color sig-nals.

THE VLP SYSTEM: A HARDLIMITEDANALOG PRECEDENT OF THE

DIGITAL AUDIO CD

The Video Long Play (VLP) system,presented to the press in 1972, used alaser to record a video color signal onan optical disc and another laser in aVLP player for playback. In 1974 theVLP was released to the market. Later,in 1978, this optical video system wasintroduced on the market in the UnitedStates, known as LaserVision.

In the VLP system a frequency-divi-sion multiplex signal consisting of a TVluminance and chrominance signal plusa sound (stereo) signal is frequencymodulated (FM) and then hardlimited,after which it is ready for recording onan optical master disc [3]. Production“stampers” made from this master disc

replicate the information as a consumerproduct by compression or injectionmolding of a transparent plastic disc of30 cm diameter. The informationappears on a spiral track consisting ofmicroscopic pits separated by flat areascalled lands. Finally, after receiving areflective coating over which a transpar-ent and protective lacquer is applied,the VLP disc is ready for playing. In aVLP player [2, Sec. 2.3, 4] the spiraltrack on the disc is scanned by a 1 mWHelium-Neon laser with a wavelength= 633 nm. The laser beam in the playeris focused by an objective lens on theinformation layer of the disc that liesunderneath a transparent protectivelayer. When the spot of the beam fallson a land, the light is almost totallyreflected and reaches a photodiode.However, when the spot falls on a pit,the depth of which is about a quarter ofthe wavelength of the laser light, inter-ference and extinction occur, whichcauses less light to be reflected andreach the photodiode. The output sig-nal of the photodiode is ideally a fairrepresentation of the original recordedsignal.

Unfortunately, there are severalsources of errors that can occur in oron an optical disc. First, small unwant-ed particles or air bubbles in the plasticmaterial, or pit damage, may occur inthe replication process. Second, finger-prints or scratches may appear on thedisc when handled. As a consequence,damage can occur not only to isolatedpits but also a sequence of pits andeven a substantial burst of pits, result-ing in dropout.

There are two reasons why thesebursts do not seriously affect the pic-

HISTORY OF COMMUNICATIONSEDITED BY MISCHA SCHWARTZ

The article following on the history of the development of theCD, written especially for this column by one of the engineerswho participated in the development effort, should be of interestto all readers of this magazine. As one of the reviewers of thearticle noted, “CDs and successor optical discs are so much apart of our lives, it is fascinating to read about their genesis.” Ashe goes on to note, “As an engineer, it is equally fascinating andinsightful to see how technologies which now seem so obviousand inevitable were once open to debate.” I venture to guess thatthis is true of almost all technologies, whether large systems ordevices within systems. This is what makes reading about the his-tory of a technology written by a participant in the developmentprocess so interesting. We learn not only of the successes in the

development process, but of the pitfalls and difficulties encoun-tered and eventually overcome before the system could bedeemed successful. As another reviewer noted, “ I also appreciat-ed the depiction of the human elements that are invariably partof these projects,” in this case the initial naming of the project orthe reasons for the choice of the final dimensions of the CD. Theability of two major companies, Philips and Sony, from two dif-ferent parts of the world, to collaborate and come up with such asuccessful product is another fascinating lesson taught by thisparticular history. I suggest you read on to see all of this for your-self.

—Mischa Schwartz

INTRODUCTION BY EDITOR

THE EMERGENCE OF THE COMPACT DISCHANS B. PEEK

Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next PageIEEE

Communications BA

M SaGEF

Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next PageIEEE

Communications BA

M SaGEF

Wideband Injection Transformer B-WIT 100

More at www.omicron-lab.com

Smart Measurement Solutions

Opti• mized for signal insertion intocontrol loops of any kind

Extremely wide frequency range • (1 Hz – 10 MHz)

Fully 600V CAT II compliant output •

Teams up perfectly with • OMICRON Lab’s Vector Network Analyzer Bode 100

The B-WIT 100 One injection transformer for all applications.

How stable is your switched mode power supply?

Vector Network Analyzer Bode 100 (1 Hz – 40 MHz)with Future.Pad Tablet PC from www.ibd-aut.com

IEEE Communications Magazine • January 2010 11

ture quality in the VLP system. Thefirst reason, generic to all optical stor-age systems, is that the diameter of thebeam at the surface of the disc is muchwider than the diameter of the spot atthe information layer. As a result, localdefects and imperfections at the discsurface effectively get blurred and de-emphasized in the readout signal. Thiseffect is inherent in reading out a discthrough a transparent substrate, andconstitutes one of the essential patentsof the CD system [5]. The second rea-son is that in a TV picture there is ahigh correlation between successivelines that causes a burst to be less visi-ble. Furthermore, there is a high corre-lation between successive TV frames,which in addition enhances the maskingof an error burst.

In the period 1972–1979 Philips wasnot the only company that was active inthe research and development of opti-cal recording and playback of video,audio, and data. For many an attractiveproperty of optical systems was alsothat there was no contact between thedisc and the pickup unit: thus, no wear

and tear. Another important featurewas that an optical system made ran-dom access possible. Several companiessuch as Sony, Mitsubishi, Hitachi, Pio-neer, Toshiba, TEAC Corporation,Tokyo Denka Corporation, and NipponColumbia investigated optical signalregistration and playback on 30 cmdiameter discs.

After 1972 most of these companiesstarted investigating analog signal opti-cal recording systems, but at the end ofthe 1970s some turned to digital signaloptical recording. Digital signals arethose for which both time and ampli-tude are discrete. The signal recordedon a VLP is discrete-amplitude but notdiscrete-time and is an analog signal.Already in 1965, J. T. Russell at theBattelle Northwest Laboratory in Rich-land, Washington, considered storage ofdigital information on an optical disc.In 1970 he was granted a patent on ananalog-to-digital photographic record-ing and playback system (U.S. Patent3,501,586). In 1969 M.D. Gregg wasgranted a patent on an optical videodisc [U.S. Patent 3,430,966]. It is not

easy to describe, briefly, the differencesbetween the systems described in thesetwo patents and the Philips-Sony CDstandard established in 1980. But let memention only one crucial difference.The recording and playback in the CDsystem of two digital audio signals dur-ing one hour and with for the most parta signal-to-noise ratio (SNR) of 97 dBwas impossible without using error cor-rection and error concealment. The sig-nal mentioned in Gregg’s patent isanalog, which excludes the use of anerror correcting code. Although the sig-nal described in Russell’s patent is digi-tal, no error correction is mentioned.

THE ALP PROJECTOttens, who was well informed aboutthe VLP technology, asked his engi-neers, T. van Alem and L. Boonstrafrom Nat.Lab., to investigate the possi-bilities for optical recording and play-back of four (quadro) Hi-Fi audiosignals stored on an optical disc with adiameter of 20 cm. Quadro sound wasseen as an enriched sound experiencecompared to stereo sound. After a pre-

HISTORY OF COMMUNICATIONS

Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next PageIEEE

Communications BA

M SaGEF

Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next PageIEEE

Communications BA

M SaGEF

liminary study, the Nat.Lab. was askedfor additional support in this project.

Consequently, a project group wasformed early in 1974. This project groupconsisted of engineers T. van Alem, C.Vos, and J. van Veerdonk, all fromAudio, and three researchers from theNat.Lab., one of whom, L. Boonstra,also functioned as project coordinatorand reporter. The other two were engi-neer L. Vries and his assistant M.Diepeveen, who were both members ofmy research department. In 1975 J.Mons from Audio joined the projectgroup. The project was named ALP(Audio Long Play), which indicates itsaffinity with VLP.

For L. Vries and myself there wasan important reason to prefer a digitalregistration of audio signals above ananalog one. Since we were aware of theexistence of dropouts on a VLP disc,we seriously doubted the possibility toplay back Hi-Fi sound using an analogregistration. Audio did not have thesame kind of redundancy as video. Withdigital registration, we knew that anerror correcting code could be used

that would correct random errors andburst errors to a certain extent.

Those errors that could not be cor-rected but only detected could be madeinaudible by using suitable masking(concealment) techniques. The humanauditory system as well as the humanvisual system can be deceived.

However, a digital ALP systemwould be more complex and expensiveto implement at that time. This wasprecisely the problem for Audio. Therestricted development capacity for dig-ital equipment at that time, and thelikely higher price for a digital player,prohibited consumer audio systemsfrom going digital. We agreed, however,that if an analog ALP would not func-tion properly, we would turn to a digitalALP. While the Audio industry groupstarted the development of a small ana-log player with a 20 cm optical disc, L.Vries and M. Diepeveen began bystudying, and later building, a modemfor the ALP. The following concept forthe modem was adopted. In the modu-lator, four Hi-Fi analog sound signalswere modulated in quadrature on two

subcarriers. This frequency-divisionmultiplex signal was then FM modulat-ed. Finally, this FM signal was hardlim-ited and thus made suitable forrecording on an optical master disc.Two different demodulators were builtand compared for their differentresponses to dropouts. One was a con-ventional time-delay demodulator andthe other a phase-locked loop.

At the end of 1975 the ALP playerand disc were ready for playing andtesting. It turned out that the dropoutscaused too many audible clicks andother disturbances, and neither of thetwo demodulators delivered Hi-Fisound. Thus, the analog ALP turnedout to be a dead end.

The next step the ALP project grouptook was an ALP system using deltamodulation (DM), but without applyingan error correcting code. Delta modula-tion, with slow companding, requiredalmost twice the bit rate of pulse codemodulation (PCM) but was easier toimplement. With slow companding, aDM signal did mask single and doublebit errors. However, with the nearinstantaneous companding required fora lower bit rate, the DM signal becamesensitive to single bit errors that causedaudible clicks. No masking of dropoutswas applied. Application of an errorcorrecting code was inevitable.

Nevertheless, a DM-ALP wasdemonstrated to the Philips Board ofManagement on 25 November, 1977.The Board of Management concludedthat the sound quality had to beimproved and must be made superiorto that of an LP record. They alsoemphasized that the directors of theAudio industry group must aim atachieving a world standard for this newoptical audio system. They did not wanta repetition of multiple standards ashad occurred with television, which hadthree standards: NTSC, PA,L andSECAM, and with the videocassetterecorder (VCR), which also had threestandards: VHS, Betamax, and V2000.

THE PHILIPS PROTOTYPECOMPACT DISC SYSTEM

With these directions of the Board ofManagement in mind, the directors ofthe Audio industry group decided toincrease their efforts. On a suggestionof J. van Tilburg, Audio’s general direc-tor, the name “ALP” was changed to“compact disc” or just CD. The word“compact” was in line with anotherPhilips product, the compact cassette.The diameter of a CD disc was set at11.5 cm, equal to the diagonal measure

HISTORY OF COMMUNICATIONS

12 IEEE Communications Magazine • January 2010

Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next PageIEEE

Communications BA

M SaGEF

Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next PageIEEE

Communications BA

M SaGEF

____________________

IEEE Communications Magazine • January 2010 13

of a compact cassette. The playing timeof the digital optical disc was limited toone hour. What’s more, the Audiodirectors decided to establish a Com-pact Disc Laboratory with the missionof creating a prototype of a CD playerand disc that could be shown anddemonstrated to the outside world. J.Sinjou, head of the CD Lab., had a 35-person team. He was also given the taskof coordinating the CD project withinPhilips.

At the start of 1978 discussions fol-lowed, both within the CD Lab. andbetween the CD Lab. and the Nat.Lab.,with the purpose to determine the tech-nical parameters of a CD system. TheNat.Lab. had excellent expertise notonly in optics but also in digital com-munications, digital signal processingand solid state D/A converters. Thisexpertise was essential for Audio inrealizing a CD system. After a briefperiod of discussions between the CDLab. and the Nat.Lab., a number oftechnical specifications and parameterswere determined, such as:• With PCM and 14 bits uniform

quantization for each of the twosound signals, a SNR of 85 dB foreach signal was guaranteed, suffi-cient to make the quantizationnoise inaudible.

• Specification of a recorded band-width of 0–20 kHz for each of thetwo sound signals with a samplingfrequency of 44.3 kHz.The error correction code should

ideally be based on the error statisticsof mass produced CD discs, but thatwas not yet possible. However, prelimi-nary measurements made by M. Caras-so of the Nat.Lab. optics departmenthad shown that most errors on a digitaloptical disc were single, double, ortriple bit errors [2, Sec.2.4, 8]. Therealso occurred now and then an errorburst (dropout) with a maximum lengthof about 200 bits. To correct theseerrors, L. Vries had developed a 2/3rate convolution code that was basedon ideas put forward by E. Berlekamp[6] and E. Preparata [7].

The input to the coder was a serialstream of blocks, each consisting of 14bits (one audio sample). The coderdivided a 14-bit block into seven sym-bols, each consisting of two bits. In thecoder a parity bit was added to each 2-bit symbol resulting in a 3-bit outputsymbol. This gave output blocks of 21bits, corresponding to a 2/3 rate code.This code was able to correct at mosttwo erroneous 3-bit symbols in two con-secutive output blocks. If more thantwo 3-bit symbol errors in an output

block of 21 bits occurred, the decoderdetected an erroneous output block.This erroneous block, correspondingwith one 14-bit audio sample, wasreplaced by a sample obtained by linearinterpolation from two adjacent reliablesamples. The remaining error wasinaudible.

The longer bursts with length up to200 bits were corrected by using inter-leaving in combination with the rate 2/3convolution code. By using interleavingbefore recording, a burst of errors inthe input stream of the decoder is, afterde-interleaving in a CD player, spreadout in time, allowing these now scat-tered errors to be corrected by the con-volutional code. In this way a maximumburst length of 210 bits was corrected[2, Sec. 2.4, 8]. If the interval betweentwo successive long bursts was toosmall, correction was impossible. Such asituation, however, could be detectedand muting of the audio signal applied.Muting can be inaudible provided thetime duration did not exceed 10 ms. B.Cardozo and G. Domburg observed in[9] that a dropout occurring in music

recorded on magnetic tape was inaudi-ble provided its duration was below 10ms. Muting can occur, or be obtained,in two ways. One way is to turn off theaudio signal, which takes it abruptly tozero and lets it later return abruptly tothe normal audio level. This way ofmuting is audible [9]. The second typeof muting as observed by dropouts inmagnetic recording was a different typeand inaudible. In that case, the signaldid not go to zero abruptly but gradual-ly, and later gradually returned to thenormal audio level. This type of mutingcan be used to suppress audible distur-bance caused by a scratch on a record[10], but it was also applied in the CDprototype when an error burst couldnot be corrected but was still detected.

The science of psychoacoustics eval-uates the perceptual implications of lin-ear interpolation, muting, companding,quantization noise, and many otherprocessing or environmental factorsaffecting audio signals. L. Vries and M.Diepeveen took this work into accountin building a codec, delivered to Audioin the summer of 1978, that included

HISTORY OF COMMUNICATIONS

Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next PageIEEE

Communications BA

M SaGEF

Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next PageIEEE

Communications BA

M SaGEF

___________________

linear interpolation and muting [2, Sec.2.4, 8].

After encoding for error correctionat the recording side in a CD system, itis necessary to modulation code thisoutput signal. There are several reasonswhy modulation coding (also known asline coding) is required. The first rea-son is that direct recording of the error

correction encoder output signal wouldseriously disturb the servo subsystem inthe player that keeps the beam on track.The disturbance is caused by the energypresent in the lower frequency range ofa binary signal. A first purpose of mod-ulation coding is to shape the spectrumof the signal to be recorded in such away that it has very little spectrum con-

tent at low frequencies. This require-ment is similar to that found in datatransmission and digital magneticrecording on tape or disc. Numerousmodulation codes have been proposedand used in the past. Many of thesewere aimed at a specific application.

Another reason for modulation cod-ing is that the light spot with which aCD in a player is scanned has finitedimensions. This causes intersymbolinterference (ISI). By imposing,through a modulation code, a lowerlimit on the length of a recorded run ofonly ones or zeros, the ISI can be mini-mized.

Finally, the binary signal transferredto the master disc must be such that itis possible to regenerate the bit clockfrequency from the signal picked upfrom a CD disc. This requirement canbe met by imposing an upper limit onthe length of a run of all ones or zeros.

A modulation code called M3,invented by M. Carasso from Nat.Lab.,and W. Kleuters and J. Mons from theAudio industry group, was used in theCD prototype. This M3 code is a modi-fied Miller code. Although this codewas not described in a journal or con-ference paper, it is covered in a U.S.patent [11]. If T is the bit period at theinput of the M3 modulator, the mini-mum run length at the output of theM3 modulator is T, while the maximumrun length is 3T. Furthermore, the M3code is DC free.

A further significant contribution ofthe Nat.Lab. was a completely mono-lithic 14-bit digital-to-analog (D/A) con-verter, described in [2, Sec. 2.5, 12] byR. van de Plassche and D. Goedhart.Two of these D/A converters were usedin the prototype system. In 1978 it wasthe only monolithic one available onthe market with that resolution.

Just before Christmas in 1978, thesmall player was finished, and J. Sinjoushowed it to the community of the CDLab.. He said “this here is our smallboy, shall we give him a name.” One ofhis engineers, J. Mons, shouted “hisname is Pinkeltje.” Pinkeltje was adwarf and the central figure in a Dutchfairy tale written by D. Laan. On manyevenings Mons read this book to hischildren.

On February 6, 1979 the CD proto-type system, or just Pinkeltje, with 11.5cm discs made at the CD Lab., wasdemonstrated to audio experts (in Ger-man called Tonmeisters ) from Poly-Gram in Baarn, the Netherlands.PolyGram was founded in 1972 by amerger of the Philips and Siemens LPrecord production units. They had also

HISTORY OF COMMUNICATIONS

14 IEEE Communications Magazine • January 2010

Figure 1. From left to right: J. Vig, President of the IEEE, at the Milestone Award cere-mony on March 6, 2009, and R. Harwig, CTO of Royal Philips Electronics.

IMDEA Networks is an independent, non-profit research institute whose multinational teamis engaged in cutting-edge fundamental science in all areas of networking. As a growing,English-speaking institute located in Madrid, Spain, IMDEA Networks offers a unique oppor-tunity for scientists to develop their pioneering ideas and to shape the future of networkingover the coming years.

We are looking for excellent researchers at all levels of their professional career, rang-ing from those who have recently achieved their Ph.D. (Staff and Post-Doc Researchers)to tenured Senior Researchers and very distinguished Chief Researchers.

SUCCESSFUL CANDIDATES SHOULD HAVE:

· Ph.D. in Electrical Engineering or Computer Science· High level of English (please note: Spanish is not required)· Research experience in the field of networking· Significant scientific record backed by high impact

international publications· Ability to undertake academic research at the highest level· Substantial experience in leading research groups

(in the case of applicants for Chief and Senior Researcher positions)

OPEN CALL FOR RESEARCHERS

Apply now through our online form at

www.networks.iMdea.org

Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next PageIEEE

Communications BA

M SaGEF

Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next PageIEEE

Communications BA

M SaGEF

___

IEEE Communications Magazine • January 2010 15

HISTORY

produced VLP records. Before decidingwhether or not to produce CD discs,they wanted to determine if the soundquality of this new audio system was upto their standard. During the listeningtests that day, the audio expertsswitched many times between the musicfrom the CD player and that from theoriginal analog master tapes. Theseaudio experts could not determine anydifference in sound quality. I noticedthat the muting operation in the playerwas frequently activated, demonstratingthat it was effective.

In line with their strategic plan, thedirectors of the Audio business divisiondecided that the time had come to gopublic with the CD system. Thus, onMarch 8, 1979, at Philips in Eindhoven,the prototype CD system and disc werepresented and demonstrated to an audi-ence of about 300 journalists by J. Sin-jou. The text of his presentation isprinted in [2, Sec. 2.2]. Referring to thisdemonstration, R. Bernard noted in hispaper [13] that “demonstration systemshave been impressive, and the total lackof background noise of any kind duringpauses in musical passages is particular-ly dramatic.”

The sequence and purpose of thesuccessive signal processing operationsin the CD prototype system, from thestudio to the home of a listener, did notchange when Philips and Sony estab-lished a joint standard in June 1980.What changed, however, in the Philips-Sony standard was that the successivesignal processing operations becamemore powerful and effective. These suc-cessive signal processing operationswere: A/D conversion, error correctingencoding, modulation coding, modula-tion decoding, error correction, con-cealment of detected errors, and finallyD/A conversion.

THE PHILIPS-SONY COOPERATIONAfter the CD prototype system hadbeen shown, demonstrated, and madepublic on March 8, 1979, it was essen-tial for the directors of Philips Audio tofind a strong industrial partner thatwould be interested in cooperating inattaining a common CD system stan-dard. Consequently, the Audio direc-tors approached several Japanesecompanies to ask if they would receivea delegation so that the Philips CD pro-totype could be shown and demonstrat-ed. There were many positive responses,and the following companies were visit-ed in succession: JVC, Sony, Pioneer,Hitachi, and MEI (Matsushita). TheDAD (Digital Audio Disc Committee)was also visited. The Japanese Ministry

Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next PageIEEE

Communications BA

M SaGEF

Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next PageIEEE

Communications BA

M SaGEF

__________

___________

_______________

IEEE Communications Magazine • January 201016

of Industry and Trade had assignedDAD the task to evaluate various digi-tal audio disc systems that were beingdeveloped and to recommend a worldstandard.

The period of the Philips Audio del-egation visit was March 14–23, 1979.On the last day of their visit, J. vanTilburg received a phone call from A.Morita, the chairman of Sony. Moritasaid that, after consulting the manage-ment of Sony, he had decided to coop-erate with Philips. Soon, the vicechairman of Sony, N. Ohga, wouldcome to Eindhoven to discuss the con-tract. Sony was an ideal partner forPhilips. It not only had an excellentposition in products related to digitalrecording of audio on magnetic tape,but had also developed a prototypeoptical digital audio player and disc.

On September 2, 1977, T. vanKessel, also a department head at theNat.Lab. , and I visited Sony’s researchlaboratory in Tokyo to observe the stateof the art in their digital audio record-ing. In the morning T. Doi showed andexplained to us two digital audio record-ing products of Sony. The first was aconsumer product, the “PCM-1” thatused 13-bit nonuniform quantization.The second was a recorder for the pro-fessional market, the “PCM-1600.” Thisrecorder had been developed togetherwith NHK (Nippon Hoso Kyokai, theJapanese Broadcasting Corporation)and used 16-bit quantization and errorcorrection, and later in the 1980s turnedout to be essential for mastering CDdiscs. Doi said that Sony was studyingoptical digital audio systems.

Later that year at the “Tokyo AudioFair,” Sony and Hitachi exhibited pro-totypes of optical digital audio playersand discs. The diameter of both opticaldiscs, however, was 30 cm, much largerthan the 11.5 cm of the Philips disc. Atthe 58th AES Convention in November1977, three authors from three otherJapanese companies presented a paper[14] on an optical digital audio systemwith a 30 cm diameter disc. These threecompanies were TEAC Corporation,Mitsubishi Electric Company, and theTokyo Denka Company. Sony’s digitalaudio optical disc system with a playingtime of two and a half hours that usederror correction and a 30 cm diameterdisc is described in [15].

To establish a common standard forthe CD system, Philips and Sony agreedto a sequence of meetings to be heldalternately in Eindhoven and Tokyo.During these meetings, the technicalexperts from Philips and Sony workedout issues including the playing time of

a disc, its diameter, the audio samplingfrequency, the signal quantization (bitsper sample), and the signal format tobe used. The signal format specifies thepurpose of each bit in a recorded frameon a disc.

To fix the signal format, Philips andSony also had to agree on the modula-tion code and error correction code tobe used. It turned out that these lattertwo issues would require the most timeto settle. The reason for this is likelythat the Philips-Sony team wasinformed about several decades of com-munications theory and practice; thus,there was a large variety of optionsfrom which to choose.

A period of intense discussions, cor-respondence, computer simulations,and performance measurements of discsfollowed. At the first of six meetings, onAugust 27 and 28, 1979 in Eindhoven,the two companies had different pro-posals not only for error correction butalso for modulation coding. A detaileddescription of how Philips and Sonyestablished, during their six meetings,the final error correction and modula-tion codes can be found in [2, Sec.3.1].This description explains that these twofinal codes had superior performance tothe initial proposals of Philips and Sony.In other words, these were the result oftrue interaction and cooperation. Thestandard for error correction was called“cross interleaved Reed Solomon code”(CIRC), whereas the standard for mod-ulation coding was called “eight-to-fourteen modulation” (EFM).The lastmeeting in Tokyo was on June 17 and18, 1980.

Philips and Sony decided, in June1980, to apply for two patents, one onCIRC and the other on EFM. Thesetwo patents were later granted by theU.S. Patent office and are registered as:• K. Odaka, Y. Sako, I. Iwamoto, T.

Doi (all from Sony), L. Vries(Philips), “Error Correctable DataTransmission Method,” U.S.Patent 4,413,340, 1983.

• K. Immink, J. Nijboer (both fromPhilips), H. Ogawa, K. Odaka(both from Sony), “Method of

Coding Binary Data,” U.S. Patent4,501,000, 1985.Together with the patent of P.

Kramer [5], these three patents areessential for the CD system.

Furthermore, the following impor-tant parameters were fixed during thePhilips-Sony meetings: a 44.1 kHz sam-pling frequency, a 16-bit uniform signalquantization, a playing time of approxi-mately 60 minutes, and a disc with adiameter of 12 cm.

The full Philips-Sony CD standard isdescribed in the so-called “Red Book”but is also available in a publication ofthe International Electrotechnical Com-mission (IEC) [16].

According to J. Sinjou, the diameterof 11.5 cm of the prototype disc was, bya personal wish of N. Ohga, changed to12 cm. Using a 12 cm disc, a particularperformance of Beethoven’s Ninth Sym-phony, a favorite of N. Ohga with alength of 74 minutes, could be recorded.Another anecdote is how the size of thecentral hole in a disc, which had to bestandardized as well, was determined.When this issue turned up during one ofthe Philips-Sony meetings, J. Sinjou puta Dutch coin, a dime, on the table. Allagreed that this was a fine size for thehole. Compared with other lengthy dis-cussions, this was a piece of cake.

FINAL REMARKSIn 1982, CD players and CDs came onthe market. Soon after that the musicloving world embraced this new way oflistening to music. The advent of theCD system was followed by severaldecades of development of digital opti-cal disc systems for reproduction andrecording of computer data, music, andvideo. To name only a few, they includeCD-ROM, CD-R, CD-RW, DVD, andrecently Blu-ray [2, chapters 4–6]. Thisevolutionary process was set in motionon March 8, 1979, when a prototype ofa CD system, with a small disc, waspublicly presented and demonstrated.

To commemorate this breakthrough,Royal Philips Electronics recentlyreceived a prestigious IEEE MilestoneAward in Eindhoven. On March 6,2009, IEEE President J. Vig presentedthis award to R. Harwig, chief technolo-gy officer of Philips Electronics, at aceremony that was hosted by Eind-hoven’s Technical University andPhilips. At the ceremony, a presenta-tion on the history of the CD was givenby J. Sinjou, and H. Ogawa, a memberof the Sony team in 1979–1980, gave apresentation on Sony’s developmentsrelated to the CD system.

During the award ceremony, I pre-

HISTORY OF COMMUNICATIONS

J. SINJOU PUT A DUTCH COIN,A DIME, ON THE TABLE.

ALL AGREED THAT THIS WAS

A FINE SIZE FOR THE HOLE.COMPARED WITH OTHER LENGTHY

DISCUSSIONS, THIS WAS

A PIECE OF CAKE.

Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next PageIEEE

Communications BA

M SaGEF

Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next PageIEEE

Communications BA

M SaGEF

Are You Ready for a Growing Worldwide Wireless Market?

Wireless Communication Engineering Technologies

(WCET) Certification Program

Application DeadlineApplication Deadline: 26 March 2010

WWW.IEEE-WCET.ORG

ni

IEEE Communications Magazine • January 2010 17

sented the first copy of the MilestoneBook with the title Origins and Succes-sors of the Compact Disc [2] to J. Sin-jou. This book provides much moredetail, including reprints of key papers,on the origins of the CD system and thesubsequent evolution of digital opticalstorage, with a focus on the contribu-tions of Philips.

ACKNOWLEDGMENTI want to thank J. Sinjou for the discus-sions we had on the history of the com-pact disc and for providing me detailson that history. Thanks also to S. B.Weinstein for his careful reading of adraft, and for helpful comments andsuggestions.

REFERENCES[1] J. K. Lang, Das Compact Disc Digital Audio

System, 1996, RWTH, Aachen.[2] J. B. H. Peek et al., Origins and Successors of

the Compact Disc, Springer, 2009, PhilipsResearch Book Series, vol. 11.

[3] W. van den Bussche, A. H. Hoogendijk and J.W. Wessels, “Signal Processing in the Philips‘VLP’ System,” Philips Tech. Rev. 33, 1973,pp. 181–85.

[4] K. Compaan and P. Kramer, “The Philips

‘VLP’ System,” Philips Tech. Rev. 33, 1973,pp. 178–80.

[5] P. Kramer, “Reflective Optical Record Carri-er,” U.S. Patent 5,068,846, 1991.

[6] E. R. Berlekamp, “Notes on Recurrent Codes,”IEEE Trans. Info. Theory, 1964, pp. 257–58.

[7] E. P. Preparata, “Systematic Construction ofOptimal Linear Recurrent Codes for BurstError Correction,” Calcolo (Pisa), 1964, pp.147–53.

[8] L. B. Vries, “The Error Control System of thePhilips Compact Disc,” 64th AES Convention,Nov. 2–5, 1979, New York City, Preprint1548 (G-8).

[9] B. Cardozo and G. Domburg, “An Estimationof Annoyance caused by Dropouts in Mag-netically Recorded Music,” J. AudioEng. Soc., Oct. 1968, vol. 16, no. 4, pp.426–29.

[10] J. B. H. Peek and J. M. Schmidt, “Circuit forSuppressing Noise caused by Scratches on aPhonograph Record,” U.S. Patent 4,208,634,1980.

[11] M. G. Carasso, W. J. Kleuters and J. J.Mons, “Method of Coding Data Bits on aRecording Medium, Arrangement for Puttingthe Method into Effect and Recording Medi-um Having an Information Structure,” U.S.Patent 4,410,877, 1983.

[12] R. J. van de Plassche and D. Goedhart, “AMonolithic 14-bit D/A converter,” IEEE J.Solid-State Circuits, SC-14, no. 3, 1979, pp.552–56.

[13] R. Bernard, “Higher-Fi by Digits,” IEEE Spec-trum, Dec. 1979, pp. 28–32.

[14] T. Iwasana, T. Sato, and A. Ogawa, “Devel-

opment of the PCM Laser Sound Disc andPlayer,” 58th AES Convention, New York,NY, Paper 1309, Nov. 1977.

[15] T. Doi, T. Itoh, and H. Ogawa, “A Long-PlayDigital Audio Disk System,” 62th AES Con-vention, Mar. 13–16, 1979, Brussels, Bel-gium, Paper 1442.

[16] IEC, “The Standard of the ‘Audio Record-ing-Compact Disc Digital Audio System,’”Geneva, doc. 60908, 2nd ed. 1999.

BIOGRAPHYHANS B. PEEK [SM’75, F’87, LF’ 06] received hisdegree in electrical engineering from the Tech-nical University of Delft, the Netherlands ,in1959, and his Ph.D. degree from the TechnicalUniversity of Eindhoven, the Netherlands, in1967. In 1961 he joined Philips Research Labo-ratories, Eindhoven, where he was involved inthe areas of digital signal processing and com-munications, and was a department headfrom 1969 to 1987. From 1987 to1991 he waschief scientist at the same laboratories, andfrom 1987 to 2001 he was a part-time profes-sor at Nijmegen University, the Netherlands.He was the recipient of various awards, includ-ing the Gold Medal from the Dutch VederFoundation in 1967, the IEEE CommunicationsMagazine Best Paper Award in 1985 for hispaper, “Communications Aspects of the Com-pact Disc Digital Audio System,” and the IEEEMillennium Medal 2000. He is a member ofEta Kappa Nu.

HISTORY OF COMMUNICATIONS

Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next PageIEEE

Communications BA

M SaGEF

Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next PageIEEE

Communications BA

M SaGEF

_____________________