I'm waving my 10 pulsing digits in the air-digital audio is here with a loud undistorted bang and total silence in between.
True, it has been arriving for quite some years, as this department rather obscurely pointed out several years back. But this is different. This is the year of professional digital impact.
(Home hi fi will have to wait a bit.) Pro digital tape recorder, not merely one but several, in different formats, all operational (if scarcely yet for sale)
and all, as usual, incompatible with each other. At the AES they set up a special meeting to see if it is already too late. Haven't heard the result.
Even more sensational, in the long run, and aimed at the home, was the first purely PCM disc, jointly developed by three important Japanese firms and surely the forerunner, at least, of the ultimate disc that will replace the aging LP x years from now. It jumps quanta ahead. Two of these firms had the disc on hand with literature and demos, Teac and Mitsubishi. The thing was so new that a special extra technical session had to be added at the AES convention, conducted through a three-way interpreting arrangement that was slow but did not, in the least, diminish the intense interest of the audience.
The earlier digital sessions at AES were jammed far beyond the doors. I do not remember seeing such crowds on hand for a technical session, no matter what the subject. All of which provided a vast concentration of info and demo which in due time will disseminate out to all of us. The Soundstream digital recorder, the upcoming 3M, the Mitsubishi--it uses quarter-inch tape at 15 ips-and that company's even newer digital cassette audio recorder, a machine that uses standard half-inch video cassettes for two hours of audio in one small package--every one of these was so incredibly removed from any previous analog recorder, even state-of-the-art, that one could only gasp in disbelief, then in due time, sigh with comprehension.
Yes, it can be done and is being done. Signal to noise at virtual infinity-there is no noise; 90 dB is the acceptable figure. Distortion down in the nil regions, NO IM, for instance (none that can be measured, anyhow). NO tape print-through (and is that a tremendous advance in a practical way). Even more important, due to the very nature of the digital coding, NO deterioration of signal from copy to copy to copy. If you can retrieve all of the pulses, all of the numbers in the code, you have the total signal-and not a thing else. So you can, in the theory, mix down 32 tracks in a dozen succeeding generations and come out with a crispy signal and no noise! Enough to make the pop people swoon. You see what I mean.
To say nothing of the elaborate new audio techniques that are available via the digital approach, since digital, as Soundstream puts it, "is not necessarily linked to a fixed time-base." Let's not go into that--it deserves a special article just in itself.
Finally, and significantly, the digital recorder can deliver a signal to your conventional cutting lathe that is in every respect as good as (or better than?) the direct-to-disc signals now rightly popular among hi-fi fans for their improved fidelity. All that and editing too.
Digital editing is tricky in the extreme and lends itself, typically, to the most advanced (and expensive) professional procedures. I really had to laugh when Soundstream demonstrated a perfect, single, classical "edit," two takes joined together between two trumpet notes, the precise micro-spot chosen via computerized calculation to a resolution of "less than 30 microseconds," matching wave forms for exact phase and continuity. Amazing, yes, but a wee bit costly, all in all. They used two machines and a disc-type storage; alternatively one could use three recorders, which would be just fine for budgets in the hundred thou area. I laughed, because in the time it took to demonstrate Soundstream's one, single splice I could have hand spliced a dozen of the same. Been doing it for years, and it doesn't cost a dime. (What do I do if my waveforms don't micro-match and there is a sonic bump? Put it back together and try, try again). Which is not to denigrate Soundstream's accomplishment, which without a doubt will have important usefulness in professional areas.
Conflicting Configurations
I am only mildly uneasy, in the face of all this, because of the disparity of approach and the incompatibility that has shown up in these operational digital systems. Nobody's matches anybody else's. I do hope that the AES meeting of minds explored modifications that just might bring these machines a bit closer to one another, but this is probably wishful thinking.
However, one thing must be understood at the beginning, which changes a lot.
True, there are four or five different configurations for digital master tapes, and thus we may have gross incompatibility as between machines, where in present analog masters there is track for track interchangeability (with a bit of equalization plus Dolby, dbx, and such). But we have a vast and saving digital grace-copying. Since there is no loss in digital copying (not counting dropouts, a special problem), it will be simple, relatively, to play any master tape on its own tape recorder and make an appropriate virgin perfect copy on some other machine, different specs. And if digital now debuts on two-inch, one-inch, half-inch-cassette, and quarter-inch tapes, at various speeds, scans, and basic digital coding systems (pulse coding being only one), then we also have a variety of analog tapes in as many sizes and speeds right now. All in all, not too bad a prognosis, within the pro area.
If & when the digital recorder gets to the consumer, it will be another story.
Please, gents, ONE system, this time! Remember stereo.
I was fascinated by the already well worked out solutions to the problem of tape dropouts in digital recording.
These can cause real chaos and would be doing so now, even with the best of tapes, if it were not for the ingenious correctives designed to cope with them. The details are for a technical article but I got the drift. It is a kind of redundancy, like making a safety tape as well as original, only this is built into a single channel, a lovely (and difficult) concept for the likes of me, involving "parity bits," extras carried along with the digital stream, a means whereby one or even several missing bits (dropouts) can be reconstructed for a perfect sequence. A scrambled order, in effect a micro time delay, seems to be a part of this technique if I get it right. (Phew, the things I have to learn these days, me a trained musician. Now, you ask me about augmented sixths and the Phrygian mode, I can whistle them to you....)
Disc
Finally, the biggest sensation, if the furthest out-the digital PCM audio disc, Teac and Mitsubishi, which those of us could stay on for the extra AES session actually got to hear. Stunning.
There is no better word. Once again, those high dB figures-here it was "better than 98 dB." That's the dynamic range. So you want an expander to expand that? There is no noise at all via this disc. Just signal. So you want a noise reduction circuit? There is no wow, no flutter, NONE. (Well, they say there is none and I didn't hear any.) As in all the digital machines, the disc speed is precisely crystal controlled. As for distortion, the specs sound just as crazy as those for the tape machines--preposterous, impossible. Did I say there was no IM? Probably. Nothing to inter-modulate.
The THD and noise (if any) combine to less than 0.1 percent, but this is just another one of those figures. Stop right there-all digital recording, any old kind, is this way when the designers want it to be.
Head room! Vast amounts of it in all technical directions. That is the thing we get via digital. Disc head room, too, just as in the digital light guide audio techniques of Ma Bell. The problem with the present 30-year-old LP is that we have systematically used up the last bits of its head room-indeed, in some aspects of four-channel sound we clearly went a dangerous step beyond. Now we see a new deal, a new disc generation, and head room suddenly opens up again. For audio it is nearly infinite in practical terms.
The PCM digital disc (not to be confused with the recent "PCM" LP records, cut from digital tape masters by Denon) is a laser beam record. The disc is plastic, 12-inch, inherently inexpensive to mass produce, and has no grooves. It uses the "pit" system, rows of tiny reflecting pits for the laser beam to scan digitally at high speed.
The disc turns silently at a whopping 1800 rpm but even so it runs a half hour in the present configuration, two channel stereo. Cryptic remarks in the fact sheet, however, imply that a much longer playing time is easily accomplished-"a whole set of symphonies" could go onto a single disc. And definitely there is multi-track capability -- no physical alteration, only a different coding -- up to 16 channels. How's that? Let's call it quad-squared, next time.
No stylus touches the playing surface. It isn't even on the surface. The pits are buried inside the record under a transparent protective layer. The back side is lightly mirrored. A four-way servo system does a number of improbable miracles, seemingly with success. 1) The speed is servo crystal controlled; 2) a tracking mirror system keeps the laser exactly on the middle of the endless spiral of tiny pits it is scanning, through the transparent surface; 3) the grosser radial motion, sidewise, is also, of course, servo controlled; and 4), a crucial focusing system is servo controlled to change the laser's point of sharp focus in accord with vertical irregularities in the disc itself. The beam thus follows the disc both sidewise and up-and-down, wherever it may wander.
Distortion? Not Me!
Ah-definitely important! One editor said that he heard a lot of shrillness at the demo, and felt that it could well be due to improper operation of the decode system. A slight warp, a deformation, and you are smearing your laser all over the lot. Or are you? A very big question. Frankly, I did not hear any distortion-type shrillness--but my ears are older.
Shrill or no, the laser beam disc demo was one of the most impressive I have ever heard, telling a concise story in a few minutes without a single word-just sound. Imagine it. The disc Player, size of an ordinary record player more or less (it'll get smaller), is turned on-and nothing happens.
Total silence. Is it working? Then some tiny little sounds, peeps, chirps, rustlings, and we gradually become aware of a faint woodsy scene in stereo. Far-off bird calls; you had to strain to hear them. Then suddenly WH0000000like a thunderclap in volume, the LOUDEST steam railroad whistle I have ever heard, enough to knock you silly. Followed by the engine itself, which clanked, or I should say ROARED by us with huge snorts of escaping steam, bangs, thumps, at a level that was on the edge of pain.
Distortion? I didn't hear any! Except one large loudspeaker that briefly bottomed into a death rattle. Now have you ever heard anything like that, from a disc? More--but space is out. The modern music was an amplified solo flute, played with enormous steam-like hisses and stranglings it is done quite a lot these days among the avant garde. Shrill-but distorted? I'd say no.
Then a percussion piece, marvelously chosen to show transients via rows of fast bounces of the sticks on the drumhead, each tap totally discrete. . . Now, do you want to hear the laser beam disc? N.B. Not a single word about analog to-digital converters, and back? Yes, they are vital but, at this point, mostly trade secrets. Questionings elicited a wide response of no comment, which I hereby pass on to you. There wouldn't seem to be major bottlenecks in this area, though, judging from results.
They'll tell us-later.
(Source: Audio magazine, Feb. 1978, Edward Tatnall Canby)
= = = =