Using Digital For Fun and Profit Today

Production Issues

How is production in the digital realm different from analog production? For me, two things stand out.

In the analog realm, I always had to spend a great deal of time and effort making sure I didn’t degrade the signal, because the continuous nature of analog makes such signal erosion inevitable and irreversible. This was particularly true for tape recording, and maintaining tape deck alignment had a great deal in common with keeping a twelve-carburetor Ferrari in tune - it wanted constant gentle tweaking. In digital, the recording either works or it doesn’t. There are no tweaks except to set levels. This is to say that from a practical standpoint, signal handling and storage in the digital realm is a lot simpler and more robust.

The other big thing is that in the digital realm we are working with computers, as I mentioned above, and we have to accept their nature, which is definitely quirky. In return for their blend of flexibility, virtual integrity, etc., we have to be prepared to spend huge amounts of support time, a constantly changing hardware base, and really rude and insensitive software and support. Nothing makes me more furious than to buy a piece of digital hardware or software to do a specific job that it should do easily, according to the literature, only to run immediately into an operational problem (have any of you ever had problems getting the right SCSI cables, for instance?) and when I call the manufacturer for help have them condescendingly ask, “Why would you want to do that?”

The point is, computer manufacturers are building infinity in small boxes and there is no reasonable way they can anticipate what you want to use it for. As a result, things that are sweetly reasonable to you never occurred to them and they may (a) have not provided for it, (b) provided for it sort of, but not exactly they way you wanted, or (c) provided for it fine, except it conflicts with something else you also want to do with the same computer (Init Madness is an example of this problem).

You must be prepared to be patient, and spend a lot of time slogging through some really obscure, frustrating and confusing problems. There will be times when you think you have lost your mind, and there will be times when the image of a Louisville Slugger right through that smirking arrogant little blinking cursor is really a sublime and ecstatic fantasy.

In short, you’ve got to learn to work in virtual worlds as well as real ones, and that is an extra layer of stuff to know that is sometimes really frustrating. So, always book twice as much time as you think you will need, and then be mentally prepared to double that if necessary.

Like analog production, you should back up things compulsively, and document everything on paper or other software medium.

Finally keep in mind Moulton’s first real law: “As you get more sophisticated equipment, you get more sophisticated problems too.”

Digital Benefits

Editing, non-destructive

One of the biggest benefits of digital audio has been in the editing realm. In analog, of course, we take a razor blade and slice up the tape and then paste it back together and listen to it and then slice our wrists unless we’ve been lucky. This is called destructive editing. In digital, editing simply involves reordering of data files, which is a cinch computationally, and is mostly done as an edit decision list, which in turn is simply a set of instructions about the order in which the various contents of the data file are played back. The original audio data is never changed, so we can always go back!

Noise reduction

A variety of single-ended noise reduction schemes have come out in the digital realm that really work quite well. Typically, these are not real-time systems but involve an analysis of the noise floor of a signal and them some fairly complex computational work to reduce the amplitude of just the noise floor.

New EQ

In the analog realm, all filter topologies involve phase shift as a function of amplitude shift at a particular frequency. In the digital realm, a number of mathematical filter topologies are being written that don’t involve such phase shift. No doubt, EQs written with such topologies will sound different, with some loud disagreement about whether they sound better or not (“Just give me plain old analog Butterworth, please!”). For speaker crossover networks, such filters have really intriguing possibilities.

Flexibility

One of the major benefits of digital audio lies with its “virtual” quality -- a quality that permits us, as our craft grows, to do a whole raft of things that in the analog realm would end up being unbearably expensive and/or clunky. As your workstation grows, the range of things it can do grows exponentially, yielding a wonderful kind of adaptability and flexibility, plus economy of resource. The first $20 grand is the hardest!

Digital Oddities

The sonic controversy

There remains a pervasive sense that digital and analog “sound” different. Which sounds better depends on who you are talking to. Discussion usually drifts off into existential polemics about accuracy and truth vs. art and beauty, with some stuff about humanity and warmth vs. cold, sterile accuracy thrown in.

Whatever. Both digital and analog recordings are capable of providing a great deal of musical pleasure to the listeners. Which or what combination of digital and analog audio you use will probably be determined by more mundane considerations like flexibility, cost and ease of use. Concentrate on making music, not selling the medium!

Stereo collapsing to mono

This is an interesting one. For a number of years, I’ve heard some pretty respectable engineers talk about digital recordings that collapsed to mono when there were enough errors. Intuitively, this doesn’t make sense to me, but one day I heard it happen to one of my tapes. Rewind, switch tape machines, and everything was fine. Could grabbing good data from the other channel be part of the error correction for S/PDIF? In any case, with tape based systems it does happen occasionally and it means you got a heap o’ errors.

Low level distortion, artifacts

One of the oddities of digital audio is that distortion increases as the signal gets smaller (see dither). Further, at extremely low levels there are all kinds of funny things going on (my Sony D5 CD player, for instance, emits a nice 10K tone at about -97 dBu) that make analog tape hiss and noise seem positively wonderful. The point is, in digital as in analog, you want to keep your levels up so that you are working as close to the overload level as you can reasonably manage. Also, note that there is no gradual onset of distortion as you approach overload in digital (which may be part of why some folks prefer analog), and when you do hit overload, it isn’t quite Armageddon. It just sounds like good ol’ op-amp clipping.

In-out-in-out-in-out

As noted above, errors in digital primarily occur during conversion from analog to digital and vice versa. Because of this, it has become a truism that once you are in the digital realm, you should stay there in order to maintain signal purity and integrity (I love these semi-religious concepts applied to science!). In fact, you can happily flit back and forth from analog to digital to analog quite a few times (oh, say, ten or twenty) before you run into any kind of serious degradation. During a complex multitrack mixdown these days there are probably fifty such conversions going on, just so you know. There may be good reasons for going back and forth (just as there are sometimes good reasons for staying in one realm), so don’t get yourself into a digital panic about truth and beauty. Just use your ears.

Dave Moulton is the former Chairman of the Music Production and Engineering Department at Berklee College of Music. Alex Case attends to bits and pieces in the department and Peter Alhadeff is dithering about math at Berklee.