The View from 2005: The fundamentals of digital audio haven’t changed a bit. We now like to use higher resolution than we did in 1993 (96 kHz. Fs at 24 bits), but the sonic improvements from such usage is minor. At the same time, we have gotten into digital data compression in a big way (MP3, AAC, etc.), which have caused a significant erosion (to my ears, anyway) in sound quality (I usually refer to MP3 recordings as “Musik Lite”). Such compression isn’t dealt with at all in this article.

Why, for God’s sakes, should we be worried about this?

We got into the whole digital audio bag for a couple of pretty compelling reasons. First, digital data stores without degradation or data loss, and magnetic tape recorders were bumping into some pretty serious physical limits in this regard. Second, because digital audio is inherently mathematical, it enjoys mathematics’ freedom from context and it can be easily applied to a broad range of physical systems. Hey, it can even pretend to be a broad range of systems (we call them “virtual systems.”). Nothing comes for free, though: in order to make the discrete digital intervals of digital audio small enough to equal the resolution of our analog hearing, we have to generate immense amounts of digital data to represent audio: almost one million bits of data per second per audio channel (almost four billion bits per hour!).

This goes against the grain of much of the thrust of technology, which is to reduce data needs and costs. Techniques like synthesis, “sampling” and MIDI, not to mention the whole multitrack recording thing, are all really intended to reduce the resource, labor and data requirements (and therefore cost) of making music. Instead of hiring two guitar players, you overdub one. Instead of hiring a trumpet player for every brass sting you want, you buy a recording of a single brass note and play it back at whatever pitch you want. Instead of a whole band of synthesizer players, you use a simple control language to play all the synthesizers. As personal computers gain computing power and speed and musicians learn to use large-scale multiplexed digital data buses (such as the new one from Digidesign), even more effective reductions and simplification of data needs can be realized.

Back at the audio ranch, parallel efforts are underway to develop data-compression schemes and other techniques to tame the huge data requirements of digital audio. Such techniques include things like Sony’s Super Bit Mapping and other psychoacoustic schemes, as well as computational tricks like spectral encoding. Big efforts have been made to cut down the data required for both Digital Compact Cassette and MiniDisc. Nerds and nerdettes are hard at work on quantization schemes that more effectively correlate digital audio data to musically relevant data (right now, half of the linear PCM digital data stored is devoted to audio information between 10 and 20 kHz., which contains a very small proportion of the total musical information, for instance).

There’s another practical part of this too. I’m gonna say one word: computer. Computers are little gray boxes that increase productivity (snicker, snicker). You take a computer and spend six hours with it teaching yourself how to do a four minute task in three minutes, except when the computer crashes and you’ve got to redo thirty versions of the aforementioned three-minute task, which only happens just before deadlines (cf. Murphy).

These little gray boxes are general-purpose digital workstations. They save you money because they are so incredibly flexible and fast. And the lure of being able to simply do all of your music-making on the computer is a powerful, I said powerful lure indeed. Who can resist? Not me! The cost? You’ve got to learn to think like a computer, which is not so good for your social life, and it ain’t terribly musical either.

Thinking like a computer means, along with talking strangely and blinking in sync with the cursor even when its not there, that you should have some rudimentary understanding of computers’ value systems, as well as the way we’ve tackled the problem of getting music to happen in these really clunky gray and black and blue boxes with their dopey green, blue and gray screens and flashing LEDs, plus cursors and rodents (I hate rodents!).

Which, of course, is what this article is about. Avanti! And hang on! We’re gonna go through this fast!