Disc recording, as far as the consumer has been concerned, has always been read-only apart from a few machines that were marketed in the 30's using aluminum discs . Tape recording, by contrast, has always been read/write, allowing the user the choice of replaying commercially-made tapes or of making recordings from broadcast material . In theory, the tape user respects copy
right and never copies discs, but the coming of the cassette unit ensured that disc copying would be widespread, if only to allow music on disc to be played in a car . The copyright laws were never intended to cope with copying on such a scale, and in practice were used only against large-scale copying for gain (pirated tapes)
where breach of copyright could be proved.
Problems encountered with analog tape
Tape for consumer uses, whether open-reel or cassette, was initially analog, and as a consumer recording medium, using in cassette recorders almost exclusively, it presents many severe technical problems. These are :
1. The inherent non-linearity of the tape recording process, dictated by the shape of the magnetic hysteresis curve .
2. The high noise level of tape .
3. The poor 'headroom', meaning that severe distortion results when the tape is overloaded .
4. The poor frequency response of slow-moving tape.
All of these factors made the cassette totally unsuited to quality recording when the medium was first introduced; they also ensured that copies of tapes were of poor quality. Since the introduction of the cassette (more correctly, the compact cassette), improvements in tape coatings, record/replay heads and, particularly, circuit methods such as Dolby and dbx have made the use of analog cassettes for good quality recording achievable, and sales of 190 million cassette players each year worldwide confirm how thoroughly this medium has achieved success -- and how difficult it will be to replace.
Considering the unsatisfactory nature of analog magnetic tape recording, and particularly the narrow-tape formats like cassettes, it is surprising that the first consumer application of digital recording was to disc rather than to tape. It is all the more surprising when we consider that videocassette recording became available at about the time that studios were switching to digital tape recorders, so that the appearance of an audio digital tape system at that time might have been expected on technical grounds at least. Several audio enthusiasts bought one model of Sanyo videotape machine on the grounds that it permitted switching to audio recording, using the principles outlined in section 5. This allowed three hours of really high-class recording to be made on a cheap videotape, as opposed to an hour or so on an open-reel tape recorder using a $25 tape spool . The reasons for the emergence of CD rather than digital tape as the first consumer digital reproduction system lies more in established practice than in technology. Disc has always been the preferred medium for distribution of high-quality sound recordings. Discs have historically been of higher audio standard and much easier to mass produce than tapes, even cassette tapes.
Though sales of music on cassette have risen steadily in the past few years to overtake sales of disc, most of these cassettes are not of an audio quality that would appeal to the potential user of CD. As far as buying recorded high-quality music was concerned, the disc was the medium to use. This was the reason for concentrating on CD; and in any case the technical problems had already been researched during the development work on videodiscs . When analog tape systems are used, the worst effects are mitigated by the use of noise-reduction systems, such as Dolby-B, and results equivalent to LP quality can be obtained by using methods such as Dolby-C or dbx. An important advantage of tape is that the popular C90 size of cassette offers 45 minutes of playing per side, double the time of the LP. The tape cassette, in addition, is still the most convenient medium to carry around, to use in the car, and certainly the only one that can be recorded as well as being played. The videodisc never made much serious impact because the videocassette recorder was used primarily as a way of time-shifting transmitted programs, the ideal solution to the usual situation that the only two TV programs you want to watch in a week are always at the same time on the same evening. A player that offered replay-only of films at high prices was not particularly appealing - few films are worth seeing more than once, and hiring from the local video shop is usually cheaper and simpler than waiting for a film to become available on disc (or on cable or satellite) . Problems like that don't arise in audio - quite apart from anything else, not many listeners can receive NPR Music on FM with really good signal strength and the local radio that uses up all the money that might otherwise be used to improve WBUR fm is not exactly geared to music. In any case, if you were given a high-quality tape system, what would be worth recording - unless you already had a CD player?
The problem of copying
The factor that has, more than any other, delayed the appearance of digital tape has been the fear of undercutting the profitable market for discs . The manufacture of CD's requires enormous investment, and unless this can be recovered no manufacturer is likely to contemplate setting up a plant to make discs . If each disc can be transcribed to tape with no loss of quality, and more seriously, if that tape can be used to make another and so on, all with the quality of the original preserved, then the risk is obvious . Unlike analog tape recording, in which each copy is noticeably inferior to the recording from which it was made, digital tape offers the opportunity to make copies that are of equal quality, even if there have been hundreds of recording stages between master and copy. The sale of digital tape equipment in many countries has for some time been resisted until a method of limiting the extent to which CDs could be copied had been worked out. Such an agreement has, in the UK at least, been achieved by the incorporation of circuits into digital tape machines which allow a copy to be made of a CD, but do not permit that tape to be used to make a second generation copy. The system is described as the SCMS - Serial Copy Management System. This is a satisfactory solution which allows you to make a copy of each of your own CDs for your own use, but deters multiple-copying for commercial gain.
At the time of writing, then, the standards for DAT have been set and domestic machines are available, though at a price level very much higher then CD players . The specifications of a Denon machine will be examined later in this section. In addition we shall look at systems that are likely to provide severe competition for DAT, the Philips DCC system and the Sony Mini Disc.
The systems The two systems of digital audio tape (DAT) that have vied for consumer attention are the stationary head systems, DASH and S-DAT and the rotary head system, R-DAT. The stationary head systems have the considerable potential advantage of allowing simple tape-cut editing, something that is not possible for the domestic user of the R-DAT system, nor needed by most users . The advantage is potential only, because if data is recorded with interleaving, simple cut-and-join editing is not possible because the signals are not in their correct order. The DASH system is intended for professional use, and for a user who does not want to use editing or multi-channel recording, the S-DAT system has no compelling advantages, and is likely to be more costly. A brief look at DASH and S-DAT is useful if only to emphasize how different it is.
The DASH system The earlier DASH-1 recorders used ferrite heads with 8 digital tracks on 1/4" tape or 24 tracks on 1/2" tape . The later DASH-2 standard makes use of thin-film heads which allow double this number of tracks to be used. In addition to the audio tracks, the tape must allow for an auxiliary track for timecodes and other control signals along with cue tracks. Three tape speeds have been used, corresponding to three different sampling rates. These are 50.8 emfs (20"/s) speed with 32kHz sampling, 70 cm/s (27.56"/s) with sampling at 44.1kHz and 76.2 cm/s (30"/s) with sampling at 48kHz. These tape speeds are, of course, considerably higher than any used in domestic tape recording, even with open-reel machines.
There are also three options for track allocation which can affect tape speed. Using DASH-F, one audio channel is recorded on one tape track and the tape speed is maximum. For DASH-M two tape tracks share a single audio channel and half the tape speed of DASH-F. The DASH-S version uses four tape tracks for one audio track and can run at quarter the speed of the DASH-F version.
The S-DAT system The S-DAT cassette that has appeared to date measures 86 x 55 x 10 mm, and is slightly larger than the R-DAT type. The speeds that have been used are 4.76 emfs, identical to that for the conventional audio cassette, or the slightly slower speed of 4.37 emfs which gives longer recordings . The tape is considerably wider than the conventional audio cassette tape, however.
The density of data on the recorded tape is 64000 bits per inch, and this is achieved by using a multitrack head. This is a far cry from the type of multitrack heads that are used in conventional audio recorders, because these digital heads are made by thin-film techniques that allow a remarkable number of tracks to be laid down on a given width of tape. This is how the problem of tape speed has been overcome, since the head gaps that are currently used on high quality cassette machines are already as small as current technology allows . The R-DA T system The winner of the first round of technology for consumer use, however, is the R-DAT system, using a recorded wavelength of 0.7 microns . The basis of R-DAT is a cassette which uses tape that is 3.8 mm wide to match the tape of the ordinary compact cassette.
The cassette itself is 73 x 54 x 10.5 mm, slightly smaller than the established compact cassette (Figure 7.1). This tape is taken in a circular path across a guide, and scanned by two record/replay heads in the same way as is used for video recording. All similarity, however, ends there.
The original R-DA T designs were all intended to use recording that was compatible with video recording, using standard video tapes. This would have used a line and field structure that was identical to the US NTSC television signal, so that it could have been recorded on US recorders. Though there is no essential reason for objecting to this, the use of this standard would not permit the replay of such tapes using European video recorders, which use different video standards . We have already looked at the brief specification for the circuitry of such a recording system ...
Figure 7.1 Digital audio tape package.
... in section 5. The older VHS and Beta system tapes are too large for such uses, but the Sony Video-8 is a very attractive possibility, since the cassette is about the same size as a normal audio cassette . Sony Video-8 recorders and cameras allow these 8mm cassettes to be used for recording up to 18 hours of high-quality digital sound per cassette in place of their normal use for both video and sound tracks . Since audio digital tape for the consumer demands new equipment in any case, the agreed standard has changed considerably since the early days . Full details of the standards are available only to participating manufacturers, and the agreement of all 81 participating manufacturers is needed before technical information can be divulged . What follows, therefore, is based on such information as has been released and is freely available in the form of articles and lectures.
The standards that have emerged owe very little to previous suggestions, other than the use of rotating heads . In particular, there is absolutely no attempt to simulate a video signal, and this is a particularly important difference . Apart from anything else, it divorces rotating head audio tape technology from video tape technology, and puts an end to the idea that users might buy modified video recorders, or connect digital converters, to play their audio tapes. This does not, of course, prevent the manufacturers of video recorders from incorporating rotating audio heads into their machines, as a number have already done, particularly on Video-8 machines which now form a very substantial fraction of the video camcorders being sold.
The mechanics of the system R-DAT makes use of a 90 degree wrap of tape around its guides.
This is a much smaller angle than is used in video recorders, and it makes the tape path much simpler, avoiding the elaborate tape winding step which is needed when a video tape is loaded into a machine. In addition, this smaller angle greatly reduces tape wear and lessens the likelihood of tape breakage, since audio cassettes usually come in for rougher treatment than video cassettes. It also allows for the heads to maintain contact with the tape during fast wind at some 200 times normal speed so that indexing signals can be used to mark out one section of the tape from another . Two revolving heads are used, so that in one revolution of the head spindle, with a 90 degree wrap, the tape is being read or written by a head for only 50% of the time. This of course makes the time for which no head is in contact with the tape also 50% of the total, unlike the video recording systems in which there is an overlap period in which both heads will be in contact with the tape.
The smaller time of contact is possible because R-DAT signals make use of time-compression so that the signals are gathered up in a memory, and recorded only while a tape head is in contact with the tape. The amount of memory that has to be used for this purpose depends on the amount of data compression that is being used, and several standards exist.
The mechanical details of the system are as follows. The head drum is of 30 mm diameter, and its speed of rotation is 2000 rpm.
The speed of the tape itself is low, with three rates of 4.075 mm/s (half-speed), 8. 150 mm/s (normal speed) and 12.225 mm/s (wide track) standardized. The angle of the track laid down by the recording heads is 6°22'59.5", giving a track length of 23.501 mm in normal modes . In wide track mode, the angle is slightly greater, giving a track length of 23.471 mm. These half-speed and wide track modes are optional extras that allow for either extra long playing time at the expense of quality or extra quality at the expense of playing time respectively. The width (pitch) of each recorded track is 13.591 µm in normal mode and 20.41 µm in wide track mode. The head gap is at an azimuth angle of plus or minus 20°, alternating from one track to the next and recording over a width of 2.613mm. These mechanical layouts are summarized in the drawing of Figure 7.2. The term 'frame' is still used and defined as one pair of tracks, one in each azimuth direction.
Figure 7.2 R-DA T tape detail
The specification for R-DAT allows for four possible sampling rates, including 32kHz, 44. 1 kHz and 48kHz. The manufacturer of any piece of equipment can select a sampling rate, and it is likely that the 44. 1 kHz rate, identical to the rate used for CD, will be confined to play-only machines, so as to avoid the possibility of easily recording digital signals directly from CD on to DAT. The other rates are available for record/play machines, and it remains to be seen what the final choice for consumer equipment in the UK will be . The standards allow for six possible modes of use, combining these sampling rates with different bit numbers according to the use of the tapes.
The reason for the multiple standards is to allow for development . DAT is already being used as an (expensive) system for the backup for computer hard discs, and there would be little point in using a system that was intended for audio for such purposes . In addition, the DAT standards are intended to govern a whole range of DAT type recorders, including professional studio recorders, and copying machines, so that the existence of six modes does not mean that DAT players will come equipped with a six position selector switch, only that the standard that is used for domestic machines will use one of these modes, as agreed by the 81 participants.
The normal recording and playback sampling rate is 48kHz, providing a 16-bit word at each sample. The two channels of a conventional stereo signal are sampled at the same time, but the signals are then interleaved into the same type of sequence as is used for CD. The signal is considered as being made up of blocks of 45 bytes (360 bits), with each head laying down 196 blocks on a track. Of these blocks, only 128 blocks contain main signal, consisting of data, sync, identification, block address and parity bits. The rest of the track is used for automatic track finding (ATF) signals which are used to maintain the tape speed so that the heads are tracking correctly, subcode (number of channels, sampling frequency and copy protection codes), and marginal signals . Error checking As you might by now expect, the DAT system includes error checking, and the systems that were proposed for earlier versions have been abandoned in favor of the CIRC coding as used on CD. The main changes in the DAT specifications since the early efforts have, in fact, been to make the system much more akin to CD, so that the record companies can use the same signal processing equipment for most of the chain from microphone to recording medium, with less risk of problems arising from the use of different systems.
The raw sampled data consists of the two channels, each sampling sixteen bits at a sampling rate of 48kHz, giving a bit rate of 2 x 16 x 48kbit/s, or 1536 kbit/s, equal to 1.500Mbit/s . Two Reed Solomon coders are used, following which the bits are interleaved to reduce the effect of long path errors. The redundancy that the Reed Solomon coding adds, plus the effect of the addition of sub-data, increases the rate to 2. 77Mbit/s . In addition, there is track interleaving of blocks . One track in a pair will contain the even numbered blocks of the right hand channel and the odd numbered blocks of the left hand channel . The other track of a pair will contain the odd numbered blocks of the right hand channel and the even numbered blocks of the left hand channel . This additional interleaving allows for the complete loss of one track, so that data could be filled in from the neighboring track. The bunching of the signals by memory so that they can be recorded in bursts by the heads then increases the bit rate at the recording head to around 7.5Mbit/s.
Copy protection
The problem of copying CDs has been tackled by using the agreed Serial Copy Management System which adds copy protection codings to the subcode. This does not, of course, prevent anyone who is inclined to ignore copyright from copying the analog signal from a CD player to the audio input of a DAT recorder. What is does do, however, is to prevent digital copying of the digital signals . The difference is important, because copying digital signals results in perfect copies, whereas the conversion to analog and back will introduce noise and distortion in the analog stages so that the copy is not perfect . It is also likely that the different sampling rates would cause beat signals to appear on a tape created in such a way.
This is not likely to deter a professional audio pirate, of whom the Far East (and, alas, the near West) has more than its share, but it should fulfill the need to deter amateurs from mass copying. This, after all, was the original intention of the copyright acts, which were framed to prevent commercial copying at a time when no one foresaw the possibility of a domestic machine which could make recordings.
The first DAT machines are now available for sale, though travelers to the Far East have been able to buy both machines and tapes for some time now. The first applications for DAT in Europe were, in fact, for computer uses, and we shall probably see several enthusiasts trying to adapt these machines for audio signals . The prices of recorders are high at present, but once agreements have been reached for marketing DAT, then there is no reason why prices should not drop to about the same level as CD players . There is much less certainty about a supply of recorded tapes . Blank cassettes are already available from a number of suppliers, but no-one seems inclined to make any major commitment to releasing music on DAT. We may, in fact, see a circular argument developing here, with recorders unavailable because of the lack of tapes, and tapes unavailable because of the lack of recorders . The other question is whether there is any genuine demand for a high-cost high-quality tape system which is incompatible with all other tape systems, particularly when purchasing a video-8 camcorder at about the same price (or less, in many examples) as has been quoted for an R-DAT system will allow the use of 8mm cassettes for sound reproduction. On the face of it, the two appear to be mutually exclusive, and to many users it might seem logical to carry out all recording, audio and video, on the same 8mm cassettes.
We could also query whether the conventional analog tape is dead, because Dolby Laboratories, whose experience with both analog and digital signal processing is second to none, have · demonstrated the new Dolby-S analog coding system. Switching between a CD source and a Dolby-S recording made from the same CD has shown that listeners cannot reliably distinguish between the two even when very high-quality amplifiers and speakers are used. In addition, Dolby Laboratories have been working on tape system which use a form of delta modulation that allows a very considerable reduction in the bandwidth that is required for digital recording.
In addition, there are several other developments pending which might make R-DAT go the same way as domestic Beta system video recorders . One is the announcement by Philips of the Digital Compact Cassette (DCC) (see below) which uses the conventional audio cassette which can be recorded with either conventional audio tracks or a digitally recorded signal. The use of this format means that the new blank cassettes can be used either for analog or for digital recording, and the aim is that the digital recorders will also be able to play the analog recordings at normal analog quality.
The other development applies to analog recording and is a patent for a new tape bias system called contour biasing which avoids the compromises that always attend conventional AC bias systems . This system is claimed to reduce tape-hiss and permit better recording of high frequencies to such an extent that a tape recorded without any noise-reduction systems can be as free from noise as a CD. The very considerable advantage is that this could be applied to pre-recorded analog tapes and the benefits noticed even on simple replay systems . One point that does seem to be clear is that the copyright acts will have to be drastically revised. We can no longer pretend that people buy tape recorders in order to make tapes of baby's first words or to play poorly-produced and badly recorded commercial cassettes, just as we cannot pretend that owners of computers do not make copies of every disc that comes their way. Until it is possible for a private owner of domestic equipment legally to make copies from any source of any material for his/her own non
commercial use, the copyright law will be held in contempt and regarded as an unenforceable and antique provision. This will result in gains only for the commercial pirates.
The DCC system The principles of Digital Compact Cassette (DCC), a stationary head system, contain items of considerable importance for the future development of all recording systems, because they hinge on much more complex signal processing than has been used in the past for audio-only signals . When compact disc was developed, the conversion from analog to digital was total; the analog waveform was simply converted to a digital waveform with no attempt to alter its characteristics . This led automatically to the bandwidth requirements that have determined the standards for the CD system. This is still the standard as far as high quality sound is concerned, and there is likely to be furious argument over the use of 'doctored' sound in the new systems . Psychological factors The use of 8mm video tapes for sound recording (by Sony) started the quest for reduction of bandwidth. Two of the three factors which make such reduction possible are psychological rather than technical, which is why there will be arguments on the subjective sound quality for some considerable time to come.
The first psychological argument concerns the threshold of hearing. For each frequency in a sound, there is a minimum amplitude below which the sound is not heard. This amplitude is lowest for a frequency around 2.SkHz, highest in the extremes of the spectrum (20Hz and 20kHz) and though the effect varies from one person to another (and alters with age) an average value can be taken which applies to most of the human race. CD recording takes no notice of this, using valuable digits for recording sounds that cannot (in theory at least) be heard.
The second psychological argument is that loud sounds hide softer sounds. For years we have been told of the 'Cocktail-party effect' , in which the ear picks up softer conversations over a loud babble, but this now seems to be discarded as far as music is concerned. It is argued that the overall loudness of a piece of music allows softer sounds to be neglected - in effect this means that the threshold of hearing can be adjusted according to the overall amplitude, dispensing with the need to cater for the lower amplitudes in such a mixture . This latter argument seems to me to be on shaky ground. When a full orchestra is playing, can we ignore one oboe? If so, why not both? Are the bassoons required? Are the clarinets really heard? Conductors can certainly hear instruments whose contributions seem to be masked by the volume of sound, and other listeners might also find that the disappearance of 'insignificant' contributions made a noticeable change to sound quality: For many purposes, the saving in bandwidth might override the effects on the sound, but my feeling is that the full-bandwidth CD will still be the preferred medium for distribution of music for which high quality recording is important. This point is particularly applicable to the Sony Mini Disc system (see page 126), which some fear might replace CD. The third factor that allows reduction in bandwidth is the 'bit bank' use . Each digital number used in the coding of an analog signal can make use of a maximum number of bits (the bit-bank) , but never does so because only a white-noise signal ever occupies all of the frequency spectrum and requires each bit of a digital signal to be used. Normal music only ever uses a fraction of the total bits in each sampled unit, and the spare bits (which will not necessarily be the same bits in each sample) can be used for other purposes . By taking advantage of all three of these factors, it is possible to reduce the bit rate of a recording to less than one quarter of the rate used for a CD recording. Philips refer to this processing as PASC, Precision Adaptive Sub-band Coding, and has devised ICs to carry out the complex processing that is required, some aspects of which will sound familiar to anyone who has encountered the later versions of Dolby processing.
The audio frequency signal, following sampling into digital form, is split into 32 sub-bands, each with the same bandwidth. The important point here is that digital filters are being used. Such filtering would be impossibly elaborate with the analog signals, and could not achieve the separation between sub-bands that is possible with digital filters . For each sub-band, the signals are passed only if they are above threshold level, and the threshold level is varied according to the overall amplitude (a dynamic threshold level) . This is done by applying a set of numbers obtained from studies of the human ear and held in the memory of the processor . Since these numbers are stored in the processing chip, they can be altered by replacing the chip in order to modify the characteristics of the system if this should be required at any stage.
The use of the dynamic threshold filtering will alter the number of bits that each sub-band needs. A sub-band that makes use of only a few bits will be allocated what it requires, allowing other sub-bands to use more . In this way, a comparatively small number of bits can be more efficiently used, presenting the same overall effect as a system using a much larger number of bits . It is essential, of course, for the system to keep track of which bits are allocated for which band.
Once this coding has been completed, the bits are gathered, along with the usual Reed-Solomon error detecting and correcting codes, into eight channels of bits . The way in which bits in each channel have been allocated performs an action similar to that of interleaving, so that further interleaving is not needed . A ninth channel is devoted to control and display signals (and probably will have spare capacity) . For each channel, eight-to-ten modulation is used to avoid long runs of 0s or 1s. These nine channels are then recorded on to the tape, using a compact cassette format with chrome tape of videotape standard.
Mechanical features
The form of the cassette (Figure 7.3) is certainly not identical to the older type, though of the same dimensions. In particular, the DCC has holes for tape drive on one side only, and is intended for use only in systems using auto-reverse. This automatically excludes it from a large number of existing players (particularly the high quality ones), so that the market for the older type of cassette is not likely to suffer too soon. Philips believe that all cassette players will eventually use auto-reverse, eliminating the need to turn over a cassette.
The other striking difference is that the cassette has sliding covers for both tape and drive holes, using the methods pioneered by computer mini-discs and also by videotapes. This make a holder unnecessary, and with one side of the cassette flat and free of holes, the label can contain more information than was possible on the older analog cassette . Blank cassettes will have a tape length indicator hole and a record-protect arrangement . The head design allows both digital and analog tapes to be played (though not necessarily permitting analog recording) . The head can be rotated as part of the autoreverse action, and in each position, half of the head will be used for nine digital channels, the other half for two analog channels . The analog channels conform exactly to existing cassette standards, and when the head is turned, the positions of digital and analog sections are reversed. The head is constructed using thin-film wafers, a technique developed for S-DAT. Of the digital tracks, one is devoted to control and display information, and not all players will necessarily make use of this fully. Each digital track is 185 µm wide, but of this only 70 µm is needed for playback so that the tolerance of track width and for alignment of the tape is comparable with that of a normal cassette.
The use of a reduced set of bits and eight channels of signal allows the shortest recorded wavelength to be 0.99 µm, easily accommodated on chrome tape, and avoiding the need for pure iron tapes as are needed for 8mm videotapes and DAT tapes.
Performance
The performance that can be achieved is impressive on paper, allowing sampling frequencies of 32kHz, 44. 1kHz or 48kHz with corresponding frequency upper limits of 14.5kHz, 20kHz and 22kHz respectively. The lower limit of frequency is quoted as 5Hz.
The dynamic range, as would be expected of a digital system is high, 105dB. The audio bit rate is 384kbits per second, and the recording time is the usual C90 45 minutes per side - there is provision for the use of 2 x 60 minute cassettes. These cassettes will be designated as D45 and D60 respectively.
The Digital Mini Disc While Philips have concentrated on DCC (their agreement with Sony allowed access to Mini Disc technology, but they decided against it), Sony have been working on a radically different approach. The interaction between light and magnetism has been known since the 1820s (the Faraday effect), but only the development of small lasers and new magnetic material has made the effect one that could be used for low-cost read-write devices . A magnetic field will alter the plane of polarization of light, quite a different effect from the scattering that takes place at each pit of a CD. Since this effect is probably less-well known, a description appears in Appendix 3.
Mechanics
The basis of the Mini Disc (MD) technology is the reduction of the number of bits needed for coding, as described above for DCC and using the same basic ideas. The system used by Sony is called A TRAC (Adaptive Transform Acoustic Coding), and the reduction used for MD is IT\Ore severe, resulting in only one fifth of the current CD rate being required . This allows the use of a 2.5" diameter disc to store up to 74 minutes of music . The disc is in a protective cover with sliding shutters and the whole assembly is half the weight of an analog cassette (Figure 7.4) . The standard Serial Copy Management System (SCMS) is used, as for DAT, to prevent a copy being made from a copy to satisfy the requirements of copyright, allowing the user to make a copy but not to use such a copy to make others . Unlike any other disc medium, however, MD offers both read and write actions, and MD players are geared to cope with both . The methods that have been proposed are curiously mixed. The magneto-optical methods are ideally suited to recording and ...
Figure 7.3 Digital compact cassette .
Figure 7.4 Sony's Mini Disc
... playback by the user, allowing the equivalent of cassette use for recording. For pre-recorded discs, however, this method is not well suited to mass production, so it is assumed that pre-recorded discs would be manufactured by the well-established methods used for CDs, and making use of the same equipment. This cleverly avoids the need to set up a whole new production line for these discs . The effect on a laser-beam is not the same however, and MD players will incorporate two sets of detectors, one for the scattering effect of the conventional type of pits on the beam, the other for the polarized light signals from the surface of the magnetized discs . The correct detector can be switched into use automatically by detecting the magnetization of the disc surface.
Magneto-optical system
Magneto-optical recording and replay are not new, but in the past it was impossible to devise systems that could be manufactured at low cost, and the system has been used mainly for computer storage systems . In particular, it has always been difficult to erase old information in order to record new information, and Sony has solved this in order to make the whole process much simpler.
The basis of magneto-optical recording is the use of a magnetic coating on the disc which uses unusual materials, chemically classified as 'rare-earth elements' . The magnetic characteristics of these materials are unique - at normal temperatures they have very high coercivity so that their magnetism can be changed only by very strong fields. When the material is heated, however, it (like other magnetic materials) loses its existing magnetism and (unlike other materials) becomes much more easily magnetizable and will retain an imposed magnetic field when it cools . The heating of the material can be carried out by using a high-intensity laser beam, and the Sony system uses a magnetic head placed on the opposite side of the disc from the laser beam to alter the magnetization just following the instant when the material has been heated by the beam. This allows discs to be written by magnetic signals, and it make it unnecessary for any separate erase signal to be used, since the heat generated by the laser beam erases any old signals and the material has cooled to the extent that allows it to be remagnetized when it passed over the recording head.
The pickup system uses the laser beam at much lower intensity, together with two sets of analyzers . For a pre-recorded disc, the light from the laser will be reflected to varying degrees, as for a conventional CD system, and detected by the same form of photodiode array. For a magnetically recorded disc, the light will be reflected with an angle of polarization that will differ according to whether a 0 or a 1 is recorded, and the detector uses polarizing filters and photodiodes. The direction of polarization of the reflected light will determine which of the photodiodes generates more current, and by subtracting the signals, this will yield a 0 signal for one direction of polarization and a 1 signal for the other direction. In either case, the signals will then be processed in the usual way, but with the adaptations required for a reduced bit number form of signal, as outlined for CCD. Use on move Another novel feature of the Sony system allows the MD to be used in 'Walkman' applications. The pickup can read data from the disk (either coding system) at l.4Mbit per second, but the decoder can operate at only 0.3Mbit per second. This allows the use of a memory store which will provide bits if jolting the player has moved the pickup position. When the player starts, bits are placed at full speed into the memory and are read out at the lower rate.
The use of the pickup is suspended when the memory fills, so that the pickup is being used intermittently. If a severe jolt disturbs the pickup position, the amount of data in the memory allows for three seconds of playing, more than enough for the pickup to be returned to its correct position. The repositioning is done by temporarily recording pickup position every 13 milliseconds and using this to return the pickup to its position if it is moved.
Specifications
The specifications for the MD system are similar to those for DCC. The sampling rate, however, is fixed at 44. 1kHz, and the bandwidth at 5Hz to 20kHz. The dynamic range is 105dB and wow and flutter are unmeasurable . The disc speed (linear) is quoted as 1.2 to 1.4 meters per second, allowing a playing time of up to 74 minutes.
The future of DCC Sony envisage the MD format as a replacement for the cassette, placing it into head-on competition with the DCC for this market.
In this respect, the compatibility feature of DCC might prove to be irresistible for many audio enthusiasts who have already suffered the trauma of re-building a collection of vinyl discs . One fear that has been expressed is that record companies might find that they could produced pre-recorded MD (which are conventional plastic with no magnetic coating) cheaply enough to make them a rival to CD, eventually driving out CD and leaving no medium for distributing music of full range, free of bit-reducing technology. It appears the Philips and Sony will allow the consumers to decide, so that we might see a rerun of the battle of standards that bedeviled videotape recording for so long. Where this would leave DAT and other rotating head technologies is much easier to decide-- there appears to be virtually no place other than as a high-cost system for a few users . The Denon DTR-2000 The Denon DAT recorder-player (Figure 7.5) was the first domestic machine to be extensively marketed in the UK, though it was followed closely by a machine from Aiwa. The machine permits recording from Optical (CD), coaxial (other digital) or analog signal sources, and includes a digital fade-in and fade-out action for all inputs . The use of the sub-codes allows a large number of control features to be offered, such as Autostart ID to identify the start of each item, Skip ID to omit parts, Manual ID to insert a code, End ID to mark the end of a recording and Auto-renumber which will renumber all of the programs in order . One point about a DAT system is that it must include both A-D as well as D-A stages, because it will be necessary to provide for analog inputs that will be converted to digital for recording on the tape . The Denon uses sigma-delta modulation (see section 4) for its A-D stage, allowing a high sampling rate to be used . The D-A stage is Denon's Lambda SLC type (see section 4), rather than any form of bitstream.
Figure 7.5 The Denon DTR-2000.
The high-speed wind is at 400 times the normal tape speed, making the location of ID marks very rapid. There is also a Fine Cue function which slows the tape to half-speed for precisely placing an ID marker. The machine offers three sampling frequencies of 32kHz, 44. 1-khz and 48kHz. An amorphous head construction is used, and the claimed response is 2Hz to 22kHz ±0.5 dB, with a SIN of 90 dB. Total harmonic distortion is 0.008% and wow and flutter is unmeasurably low.