Digital audio moves so fast that the mere lag between writing and publication can put us out of date--does everybody now know about low-cost digital editing? Maybe. Anyhow, herewith more on my own recent discoveries in that area.
I think we can assume that in a very short time the preparing of digital audio material will be as routine, in every price range, as it has long been in analog recording. Not quite as cheap as the razor blade, but priced within reason. It should be possible to work usefully in digital recording at every level, from the simplest one-man operation to L.A.'s fanciest mega-syncs.
Might eventually get down to consumer equipment, what with the remarkable possibilities in time coding.
I can be specific about the particular equipment that my friend Al Swanson in Seattle used to polish up my unedited digital Brahms, as described in this column last month. The units in his layout are not the only ones around, nor is his the only possible approach, but at least I can give you some idea of the reasoning--and cost effectiveness!--of his choice. He has four basic units in his studio, plus assorted computer elements including keyboard and the all-knowing CRT monitor that told us where we were and what we were doing at every moment. The most important item is the dbx 700 digital audio processor, a unit that created a noticeable sensation with its CPDM encoding when it first appeared, a while back. Using the dbx 700 implies, you understand, a basic choice of system, including VHS applied to audio recording. The next two of Swanson's units conform: Both are JVC recorders using VHS videocassettes.
It is possible to edit similarly in Sony Beta (PCM) format at low cost, but with restrictions. You can edit 14-bit PCM using home video equipment, but with less flexibility, and less accuracy, and with glitches, Swanson says. One could also avoid these drawbacks in PCM, but at far greater expense. So for budget editing, Swanson's system depends on dbx and VHS. The first step involved analog, but not analog recording. My unedited PCM tape of choral music was decoded into analog, and was then fed directly into the dbx 700, whose output (now digital, again) was taped onto a JVC BR6400 video recorder. This is not a special VCR, simply a standard "industrial" model, perhaps a bit more rugged and direct in application than the corresponding consumer models but otherwise quite similar. Here is one advantage of the dbx-VHS combination: The source player does not have to be an expensive editing type. (I gather that for similar work in the Sony Beta area, you must acquire two of the editing-type machines, thus adding considerably to equipment cost.) A second JVC video recorder in Swanson's studio, a BR8600 editing VCR, is indeed special. It costs more, but only one is required. The edited and assembled recording is taken down, piece by piece, on this "receiving" machine.
Finally, and vitally, these units are tied together with the all-essential editing controller, an updated equivalent of the familiar audio control unit, "boss" machine for a million hi-fi systems over the years. From a number of availabilities, Al chose the Convergence Systems VE-93 (that's a brand name). He gave me two good reasons for this choice, which I pass on. First, it was the least expensive unit available with time-code facilities. More important-and we can all understand this--the VE-93 comes ready-made with the drivers and interfaces for 'the JVC line of industrial video recorders, including the two units mentioned above. Half inch format, of course. Al sees no real need for the more rugged 3/4-inch formats (which may have fewer dropouts) unless you are (I add) one of those who Must Have the Best, at Any Price, in which case you wouldn't go for this economical approach to digital editing in the first place. There is plenty of audio bandwidth available via half-inch tape, the cost is less, and the playing time gratifyingly long-two hours per cassette. (Ah, but you must be sure you have a flying erase head, and you should also have frame servo in your half-inch receiving recorder-that's a Swanson quote. No need for these in the source machine, however.) So--you simply plug all this together, along with your own special peripherals, like speakers, amps, typewriters, computer parts, ashtrays and other studio appurtenances. Frankly, the result in Al's studio didn't look any too simple for my unaccustomed eyes. It was a cross between a very elaborate basement hi-fi installation and a typical working studio of the most casual sort, with everything surrounding Al's comfortable seating place. But from a pro viewpoint the configuration was aggressively practical, I admit-compared, that is, to what some audio people get themselves into, be it messy or neat as a pin. The hookup, you see, was easy and the working conditions are comfortable. Al is his own man.
Now, for those intelligent readers who are innocent and/or unknowing (I am certain there are many--and why not?), a brief description is now in order as to what Swanson was doing with my music as I looked over his shoulder. There are three basic, interdependent facets in digital tape editing (including this bottom-cost sort). They are remarkably different from all the analog snipping and cutting of these last 30+ years. The editing itself, the matching of sounds, the removal of unwanted material, is really not much changed. Only the method is revolutionized.
First, in digital one never cuts the tape (or we shouldn't, if we do!). We do not even touch it. We edit by copying onto a new tape, with automatically controlled joints and excisions that have been neatly programmed ahead of time in each case.
In analog, it is unwise to copy because of the quality loss involved.
There are too many copying jobs already being performed, notably from those gigantesque, multi-track master tapes and mix-downs--less in classical areas, but still too much. Now we have returned to a more pure, audiophile approach, with a minimum of analog copying. Nevertheless, to copy is to lose, no matter how perfected are the electronics, and so analog editing by the copy system is not a particularly good idea. As a billion used razor blades will attest.
But in digital, due to the very nature of the message, we can copy, and even base our whole editing technique on copying from one tape to another--hopefully, with no loss whatsoever in quality. If all goes well. It often does.
So the rock-bottom editing system copies from one VHS videocassette to another, assembling the material in the process. This we do and here's the second new facet--via time code, an audio-track signal applied directly to the tapes, video frame by frame.
(There are other variants in other types of editing.) The time code is read by the machines and used as the basis for an astonishing range of precise controls and movements that are automated, visible on the CRT, never mistaken, and-even with this relatively modest equipment-dependably accurate. By this method, using the VHS videocassette base, the resolution is to about one-sixtieth of a second (one video field, or half a frame) or better, but edit points can be shifted only in increments of about one-thirtieth of a second, or one video frame. At 15 ips in analog, that is one-half inch of tape.
(For specific sonic examples of this, see last month's column.) That's less exactitude than you can get with razor blades (even on audio cassette at 1% ips, one-thirtieth of a second is one sixteenth inch of tape), but good enough when dealing with the comparatively continuous sound of singing voices. With fast piano music, it might not do as well.
Now, anyone who has bumbled around an old-fashioned recorder with an "odometer" counter like an automobile's knows how frustratingly inaccurate any mechanical timing can be. I never used the things--you rarely got back to the same place twice in a row.
Anything that depends on a mechanical connection (even the fancier readouts) is no more than roughly accurate.
But the time code, an actual signal applied to the tape, is as good as a very fine set of gears, and always returns to the exact same place, within its level of tolerance. Just as important, the basic time coding in this marvelous era of digital operation allows for a most extraordinarily easy and exact automation of every imaginable and practical sort, smoothly and effortlessly, with the proverbial push of a button.
Believe me, if you haven't worked with such updated automation (or at least stood over the shoulder of someone who is), you have missed plenty.
Most people have an idea of it, of course. If you have mastered your own VCR, you have some idea of the feel of it (assuming you are getting what you want), and those who manipulate audio-cassette recordings have an idea, an inkling (my best word) as to what can be done. Same for disc jockeys who devise ingenious semi-automated ways to switch from one recording to another at precisely the right moment.
But, I say, none of this can match the smooth, effortless precision, the unfailing accuracy of location, the automatic procedures that go on while you sit and watch them happen, which can derive from an applied digital time code. That's what the magic letters SMPTE are all about when those video and movie people get into their own special coding.
I do not intend to go beyond my own outsider knowledge here, only to relate my own experience. I gather that time codes have not been much used in analog home audio; that the Nagra recorder (reported on in its first stereo form, many years ago, in this department) is a notable exception. (It's widely used for professional video and film soundtracks. –I.B.)
Perhaps there will be more-that is, a sort of "printed on" audio signal that can be used like the cogs on a fine gearing to control tape access minus slippage. Might be nice in some very amateur recording gear, yes? But somehow this kind of procedure is inherently better off in a digital environment, with the numerous technical advantages (in particular, copying integrity) which are a part of the digital process.
So we have, for budget digital editing, the copy system, the time code and, finally, what I am calling rehearsal. It is tied to the others and just as important. Until my experience this year I had not run into it-why should I have? But wow! What a difference. I simply marveled; I could scarcely believe. Even in this low-cost, bottom simplicity type of digital editing, you have full-fledged rehearsal. You could also call it tryout. In ways that are still somewhat mysterious to me (after only a few hours' experience), you may do a dry run on any sort of sonic joint or edit, ahead of time, and listen to it exactly as it will sound, when and if you make it final. No cutting and patching, which perhaps will destroy bits of priceless music when things don't work. Via the audio processor and the controller et al., you set up your splice and the machines play it for you, using the time code. Not quite right? You edge up a bit closer, a tiny fraction of a second, and try it again. And again, until you are satisfied. Then, and only then, do you do the actual copying splice onto the receiving final tape, automatically as per prior instructions.
If you aren't pleased, if you think another place in your music might be better (this is where experience counts, in analog or digital) you have lost nothing; your signal is intact no matter how many hairsbreadth joints you have rehearsed. Thus, you see, this kind of preparation opens up a whole range of hitherto risky experiment-as I know so well--in those "impossible" editing joints that sometimes miraculously work out. No priceless slivers lost on the cutting-room floor, or too small to reattach; no more un-removable thumps and bumps! Agony is the only word for old-style disasters of that sort, whether on a one-man project or at the expense of a major recording session. Agony no more, opportunity galore.
There is still one more highly positive aspect of budget digital recording, to further offset the marginal accuracy (from some points of view) that is achieved--those discrete one-thirtieth of a second "cogs" of location. It keeps track. What a bonanza! You can't get lost. You always know exactly where you are, and you can always go direct to some other place, unfailingly. The machine takes you there.
Never again a roomful of half-played reels, semi-identified or anonymous; no more hanging screens of short lengths of tape stuck to nearby surfaces like so much chewing gum awaiting possible re-use. No more snarls of tape accidentally reeled onto the floor and then stepped on--I've often enough reduced myself to tears in that sort of editing.
Yes, there are disadvantages. You can't cross-fade; it's butt to butt. You can't go slow and rock the tape (but this is not a problem when you can actually rehearse your proposed splice to find what it will sound like). You can't alter the final tape with second thoughts-the time code will be unjointed and all that follows will have to be done over again. (But your second thoughts are okay in rehearsal.) Most of all, there is that matter of accuracy, easily solved if you move up to much more costly equipment. But don't! So much depends on the ear, on experience, on the rapport between producer and engineer, on familiarization with new procedures. For some people it won't work. I myself had no trouble getting what I needed for low-cost digital editing. Couldn't you? Worth a try--whether you buy outright or use a Rent-a-Pro like Al Swanson.
(adapted from Audio magazine, Jul. 1985; EDWARD TATNALL CANBY)
= = = =