by LEONARD FELDMAN
NO LONGER A PIPE DREAM, Thomson's magneto-optical Compact Disc recorder can be made right now-and for a reasonable cost! Test results are spectacular, but the recordings can't be played on conventional CD players.
--------------------
It looked for all the world like an ordinary CD player. Dietmar Uhde, Manager of the Physics and Chemistry R & D Labs of the German Thomson-Brandt Division of Thomson Consumer Electronics, carefully unwrapped the prototype magneto-optical disc (MOD) recorder and hooked it up in my lab. This was an exciting day for all of us. (Editor Gene Pitts accompanied Mr. Uhde, as did Friedhelm Zucker, another scientist from the Thomson-Brandt R & D lab in Villingen, West Germany.) There were, of course, no published specs, since this prototype is not in actual production as a commercial product. But as I was to learn in just a few hours of testing and listening, the DR 1000 MOD recorder is in every other way a real-make that, superb-feat of engineering. Its performance was so good, in fact, that I wouldn't mind owning it even in its present prototypical form.
Unfortunately, I wasn't allowed to keep the unit for the several days or more that I usually take to evaluate such a sophisticated component; my time with this remarkable product was limited to just a few hours. Before I tell you how well it performed as an ordinary CD player and as a CD recorder with unlimited erasability, you're probably wondering, as I was, just how the MOD works. To begin with, the MOD recorder is only partially compatible with CDs. Specifically, it has been designed to play ordinary CDs properly. However, a magneto-optical disc recorded on this Thomson unit cannot be played back on a conventional CD player-at least not on CD players as they are configured today.
----------
THOMSON SAYS THEIR data-compression system gives 16-bit sound quality from recordings made with four-bit quantization.
----------
(This is not to say that the optical assembly of some future CD players couldn't be fairly easily altered to read these recordable discs.) The problem has to do with the reflectivity of the information-bearing surface of the two media: Tv1ODs have a reflectivity of only 20%, while the reflectivity of CDs is almost 100%. The MOD recording medium is a thin layer of magnetic terbium-iron-cobalt alloy. (Terbium is a dark brown, rare-earth metallic element.) This magneto optical layer is applied by means of cathode-vapor, high-vacuum sputtering onto a 4.7-inch (12-cm) polycarbonate or glass disc. A percent spiral track guides the recording laser and already allows for the coded maximum playing time of 74 minutes. The first tracks, the lead-in tracks, are prerecorded onto the blank discs by their manufacturer. These tracks tell the DR 1000 MOD what sort of disc has been inserted and how the laser strength and magnetic field are to be set for recording.
Within the recorder, local heating of the disc's magneto-optical layer to the material's Curie point, 180° C, reduces its coercivity to a low enough level that an external magnetic field at the recording point can reverse the layer's polarity. For playback, the Kerr effect is employed: Areas whose magnetization is of opposite polarity reflect polarized or unidirectional light differently. The small change in direction of the polarized light reflected from the surface is a mere 0.5°, but this is sufficient to allow the system to differentiate between binary zeros and ones. The system also detects places where no change in polarization angle has taken place, indicating that no recording has been made.
Thomson's first generation of optical disc recorders made use of a permanent magnet, recording digital bits by means of rapid laser bursts. The disadvantage of this method was that erasure required a separate run or "pass," with the laser continuously at full recording power. In the current MOD, the laser stays on during recording, but a powerful modulator is brought to bear on the magnetic field. Its magnetic polarity is controlled by the digital data stream. Erasing and recording occur in the same sweep of the laser assembly.
A recent reduction in the thickness of the medium's recording layer substantially reduced the amount of heat required. Instead of a 60-mW laser used in earlier prototypes, the DR 1000 MOD that I tested uses 25 mW of laser power in recording, of which only 5.5 mW reach the disc. The laser power at the disc's surface is switched down to 0.8 mW for playback of magneto-optical discs and only 0.4 mW for playback of conventional CDs.
Fig. 1--Block diagram of the Thomson DR 1000 MOD recorder.
This unit can play CDs and re-recordable magneto-optical discs, and could be modified to play and record "write-once" discs.
Fig. 2--Physical layout of the optical system.
Fig. 3--Principles of Thomson's magneto-optical record/playback system.
The system shown is an earlier version, which used discs that could be recorded
only once.
Dietmar Uhde maintains that long-term stability of the stored information is no problem, nor does he foresee any limitation on the number of read/write cycles. The user lead-in area on each blank disc is a track directory to be recorded by the user. When the disc is loaded, this directory is read into a 64 kilobyte memory; if new material is then recorded, this information is updated and is recorded back onto the user lead-in track before the disc can be removed. As you'd expect, access to tracks and passages is about as fast as on good CD players.
A block diagram of the MOD recorder is shown in Fig. 1, while Fig. 2 shows the physical layout of the optical system for the MOD. Figure 3 illustrates the principles of magneto-optical recording and playback using polarity reversal of magnetic fields to change the reflectivity angle of polarized light.
Playback operating features of the DR 1000 MOD are very similar to those found on CD players. There's program repeat, A-B segment repeat, and full disc repeat. The display shows the track number and either the elapsed playing time of the current track, the time remaining in that track, or remaining time on the disc. Playback is programmable, with up to 39 program steps available.
As I mentioned earlier, at first glance the Thomson DR 1000 MOD recorder looks like an ordinary CD player. A power switch is at the lower left, and nearby are the usual transport pushbuttons plus a record/pause button.
The stop button also opens and closes the disc drawer, which is above this row of pushbuttons. A large display area is located to the right of the disc drawer, and below it are 10 numbered buttons for programming and accessing tracks. The row of controls below the number buttons handles display, level balance, headphone volume, memory (program storage), repeat functions, and the encoding of start IDs when making a recording. At the right end of the panel are the microphone and headphone jacks, a multifunction knob similar to those on Thomson Digital Line components (covered in Audio last issue), and several additional buttons for data reduction one of the heretofore undreamt-of capabilities of microprocessor control.
Fig. 4--Frequency response for CD playback.
Displacement between left channel (solid curve) and right channel (dashed curve) is due to slight differences in output levels.
Fig. 5 Deviation from linearity for undithered signals on CD.
Fig. 6 Low-level linearity for dithered signals on CD.
Fig. 7 Fade-to-noise test for CD playback.
Fig. 8--THD + N vs. frequency for CD playback.
Fig. 9 Frequency response, MOD record/play cycle.
Fig. 10 THD + N vs. frequency, record/play cycle.
Fig. 11 Record/play linearity, input vs. output; see text.
Data reduction uses a technique called Multiple Spectral Audio Coding, a joint project of Thomson and the University of Duisburg. This data-compression process uses four-bit quantization to quadruple the system's normal recording time! Mr. Uhde reported that experts on a listening panel were unable to distinguish between ordinary 16-bit music recordings and a music recording made using this sophisticated four-bit compression system. The digital recording of still photographs has already been demonstrated as well. With about 600 megabytes of storage capacity, MOD can become a totally interactive medium. As Mr. Uhde explained it, if a blue laser diode were substituted for the laser currently being used, the size of the heated area would shrink and tracks could be spaced even more closely together.
Storage capacity might then be quadrupled. The next goal, he further stated, would be to record one hour of video on a 12-cm (CD-sized) disc! The rear panel of the DR 1000 MOD has analog line in and line out jacks as well as optical and coaxial digital line in and line out jacks. This permits digital interconnection of a CD player and a DAT recorder with the Thomson DR 1000 MOD for recording onto or from the magneto-optical disc.
There was not nearly enough time to conduct all the lab and listening tests that I wanted to make. Since I had to ration my time, I decided to measure some fundamental performance characteristics of the unit as if it were a CD player, using my trusty CBS CD-1 test disc. I then made additional measurements using digitally generated test tones that I recorded onto a blank magneto-optical disc.
Figure 4 shows playback frequency response of the system in CD playback mode. Response was almost absolutely flat from below 20 Hz up to 20 kHz.
The output levels of the two channels differed by some 0.2 dB, which accounts for the +0.25 dB reading at 20 kHz on the right channel. In fact, that channel was also perfectly flat when referred to its own 1-kHz output level.
Figure 5 is the familiar plot of deviation from perfect linearity that I show when testing CD players. Using undithered signals, deviation at -80 dB was less than 2.0 dB for the left channel and just over 2.0 dB for the right. I repeated the test, this time using low-level dithered signals, and the error amounted to a bit more than 3 dB at -90 dB (Fig. 6). I've seen many CD players--some costing quite a bit--that didn't do as well at such low levels.
The results for the fade-to noise test, shown in Fig. 7, provided good correlation with the results obtained in Fig. 6 and also enabled me to see how the MOD recorder's EIA dynamic range compared with that of ordinary CD players. From the data plotted in Fig. 7, I calculated that the EIA dynamic range for this unit was between 105 and 110 dB in playback mode.
Figure 8 is a plot of THD + N versus frequency. At 1 kHz, THD + N was only 0.00336% on the left channel and 0.00398% on the right, both readings referred to maximum digital recorded level. Notice that, unlike some CD players I've tested, the DR 1000 MOD exhibited no sharp rise of THD + N as the high-frequency end of the spectrum was approached. Often, I see high readings here, due not so much to actual harmonic distortion as to out of-band "beats" caused by the D/A converter and other elements in the playback circuit. No such beats were in evidence with the Thomson unit in its CD playback mode.
There were many more measurements of CD playback I wanted to make, but only a short time remained to make my first-ever magneto-optical digital disc recording and play it back.
I think I know how Tom Edison must have felt when he yelled into that horn and made those first indentations in tinfoil 113 years ago! Rather than recite "Mary Had a Little Lamb," as Edison had done, I let my Audio Precision System One test gear generate 16-bit digital signals at 44.1 kHz. With the aid of the gentlemen from Thomson, I recorded the signals onto an MOD and then held my breath as I played back the first one-a sweep of frequencies from 20 Hz to 20 kHz. The response when the recording was played back, shown in Fig. 9, was as flat as that shown in Fig. 4, if not flatter. Again, the right channel's output was slightly different from the left channel's, but both were within 0.1 dB or less over the range from 20 Hz to 20 kHz! Distortion plus noise for the complete record/play cycle of a full-level (0-dB) signal, though not as low as for playback of the prerecorded test disc, nevertheless equaled that of many top-quality CD players. In Fig. 10, only one channel is shown, but both channels yielded practically identical readings at all frequencies. At 1 kHz, THD + N was 0.0072% for the complete record/play cycle.
Perhaps the most interesting result was obtained when I recorded a test signal which decreased in amplitude, in 2-dB steps, from maximum recorded level down to 120 dB below maximum.
This test signal was generated in the digital domain and recorded onto a blank section of the magneto-optical disc. During playback, a plot was made of output versus original input level, and as you can see from Fig. 11, linearity was superb. Admittedly, this time the test signal was dithered, but deviation from perfect linearity at-90 dB was not much more than it had been at-80 dB in the test shown in Fig. 5. Mr. Uhde smiled knowingly when I pointed this out to him. From his remarks, I gathered that the combination of A/D and D/A conversion used in the complete record/play cycle of the DR 1000 MOD was calibrated so that whatever slight linearity error is introduced during the record cycle is at least partly compensated for during playback. This is one of the tricks a manufacturer can perform when he has total control of a system, from input to record function and then to playback function.
In the little time that remained, I did manage to make a short voice recording, just to convince myself that the Thomson unit could digitally record more than mere test tones. As expected, the sound of my voice reproduced via a pair of headphones was as good as anything I had recorded and played back on my two DAT recorders. I'm not suggesting, certainly, that MOD will make DAT obsolete. For one thing, DAT will likely be with us long before we see commercial MOD recorders on dealers' shelves. But even if this should not be the case, each format has its advantages. For example, I can hardly envision a portable Walkman type of magneto-optical recorder, yet there are already DAT recorders smaller than the first portable analog cassette recorders were.
No one can know, at this point, whether or not the Thomson approach to recordable/erasable digital discs Will become the world standard. This much I am wiling to predict: We will have recordable CDs of one kind or another: sooner or later, during this new decade. Based on my brief experience with the Thomson DR 1000 MOD recorder, I hope it's sooner rather than later.
Also see: Technics SL-P8 Compact Disc Player (April 1984)
(Source: Audio magazine, Mar. 1990)
= = = =