September 1997

Audio on the Internet:

Inevitability or oxymoron?

by Paul D. Lehrman

Thanks to my parents' discriminating taste in literature, I spent a large part of my childhood studying cartoons in The New Yorker. A lot of them were above my head, but that was actually part of their appeal. Finding out why a joke was funny that I didn't get at first was one of my favorite ways of learning about the world. One cartoon that I did understand, however, showed two guys in a basement with a printing press and several crates of dollar bills stacked around them. "Sure it's a lot of money," says one of them, "but let's not forget one thing: it's all counterfeit."

I think about this cartoon when anyone starts gushing about how the Internet is the future of audio. "Sure there are a lot of ways to do it, and some are pretty interesting," the voice in the back of my mind says. "But let's not forget one thing: they all suck."

There are millions of hours and dollars being spent on the Internet, and one of the things that seems to have the most kick for the folks who are really into it is how it can handle audio. Entire magazines are talking about how on-line delivery systems are going to revolutionize the music industry: we're going to buy records from Web sites, download new releases onto our hard disks, and attend live concerts through our modems. Independent musicians will have the same opportunities as major record labels, whose stranglehold on the marketing and distribution process will finally, forever be broken.

So how do we in the audio biz gear up for this brave new world? Well, I'd start out with great big yawn. Yes, I agree the Internet is changing the ways people look for information, and there are some interesting marketing and distribution opportunities there for artists and purveyors of entertainment. But it's going to require a major change in the way the world is wired, and the way people are wired, before we have to face the prospect of scrapping our current ways of bringing music to audiences.

First of all, let's look at the delivery systems people use to access the Internet. The vast majority of homes are still hooked up with twisted-pair copper, just the way they were 75 years ago (although in those days, yu needed three conductors for a phone line). The amount of stuff you can push through 18-gauge unshielded copper is limited by the immutable laws of physics. There have been remarkable strides–in one of the first articles I ever wrote about telecomputing, some 12 years ago, I quoted a source who solemnly said, "1200 baud represents the limit of what a telephone line can handle"–but I think now we really are at the end of that line.

"But soon," say the pundits, "the copper will be replaced by something much faster." Maybe Sprint and MCI can handle a few gigabytes of data per second on their latest trunk lines, but I'm not holding my breath for those little suckers to come snaking their way up my block. Cable? A few companies are testing the waters by passing up the opportunity to add yet another cubic-zirconia-shilling channel to make room for a fast downlink to customers (they still need phone lines for data from the customer), but they're going to get a hell of a rude awakening when they realize they're now in the Internet Service Provider business, and the stuff they're pumping down their lines actually matters to people. As bad as customer service is on most ISPs, can you imagine what would happen if the cable companies took over? Imagine Netscape has started scrambling your e-mail, but the only thing you can get out of your service provider is, "We'll send someone around to look at it in a week or so." No thanks.

The local telephone companies aren't going to spend billions digging up every sidewalk in every town to bring their fiber into your bedroom–not when the wireless services to whom the FCC just gave away huge amounts of spectrum are trying to sell you the same services, only their hardware investment has been just about zilch.

But now we have such fabulous compression schemes, that we don't even need all that bandwidth, right? Compression technology has gotten a lot cleverer in the last couple of years, but audio, unlike graphics and video, doesn't have a whole lot of redundant information (you don't see "the next 4,862 samples are at $45F6" unless you're dealing with pure square waves) and so there's a limit to how far that can take us. Not long ago this magazine ran a column of Stephen's in which some over-zealous spell-checker changed "lossy compression" to "lousy compression"–a lot of us recognized unintentional truth when we saw it. The newest "lossless" schemes save 50% or less of your bandwidth, and are not real-time. MiniDisc, the sound of which some people still can't stand, cuts your bandwidth needs by a factor of 5, which is fine for getting the stuff on and off poky mechanical media, but it's pretty useless for dial-ups: to pass a 44.1 kHz stereo 16-bit signal over a 56K line (assuming you and your ISP both are using the same kind of modem, which today is a one-in-two shot, and assuming your phone connection and the ISPs backbone connection are perfect, which is about a one-in-a-million shot), you have to compress it by a factor of 25 .

Real-time audio streaming works for some applications, like simple speech, quickie sound effects, and the most primitive of music. But the claims being made that high fidelity music can be delivered this way are simply ridiculous. "FM quality" or "Near CD quality" from a dial-up? Give me a break. And what the heck is (yes, I've really seen this) "Near Stereo quality"? 1.5 channels? And remember wrestling with SMPTE timecode and sample-rate clocks so that sound and picture would stay in sync? When your software and hardware are doing all they can to make sense out of super-compressed data, dropping and duplicating frames to give you a semblance of continuous motion, you can pretty much forget about syncing aything to anything.

Okay, so maybe we won't be listening to music on our computers the way we listen to the radio. But what about selling records? Even if the quality isn't so hot, isn't the Internet a great way to get people to listen to stuff, which they'll then go out and buy? It could be, if those people have the patience to wade through pages and pages of garbage to get to something they like. But record labels and their marketing departments are going to have to take the technology seriously, so their sites sound good enough to make listeners want more. Internet audio is actually far more finicky than CDs: at least with CDs you know what kind of D-to-A conversion the listener is going to use. But on the Net, you have absolutely no idea, and so you have to tailor your audio to sound as good as possible through an absurd variety of codecs and transmission protocols.

If you read Dan Daley's positively scary interview with quality-control engineer Cheryl Engels in the June Mix, you know that this is a disaster on its way to happen: if major record companies, with all their resources, won't pay for quality control on their CDs, a product into which they have invested many thousands of dollars in materials and manufacturing processes, which too often contain fatal, easily-correctable-if-someone-had-just-been-paying-attention flaws, what makes you think they're going to bother to check out all the possible ways the materials on their Web site–which cost them nothing–can screw up?

Well then, maybe this really is the chance for "alternative" marketers to strut their stuff. A small label with an attractive, well-functioning Web site can blow those lazy, clunky majors out of the water, right? Wouldst that it were so. Low-end promotion on the Internet is the rough equivalent of college radio: you know it's there, and maybe you swing over there once in a while, but a lot of it's really, really bad, and you haven't got the patience to listen for hours to inept and cooler-than-thou wannabe disk jockeys (translated: wade through pages and pages of slow-loading, ugly, unreadable, illiterate junk) in the hope that you'll stumble on something worthwhile. Don't get me wrong–I love college radio and it's a great breeding ground for new talent, but you have to want to listen to it. It's the same on the Internet: for someone to find you, they've got to want to. This, of course, costs money: print, radio, and television ads, or banners on other people's sites.

As I was writing this, a friend forwarded me a Usenet newsgroup posting that announced a new site where, for a fixed monthly fee, you can list your records in an on-line catalogue, and the site owners will handle all the publicity and credit card transactions. This sounds like a no-lose proposition, but this site is starting out with two strikes against it. First, people have to read the post, and since so much of the traffic in newsgroups is utter garbage, and everybody knows it, only a small percentage will. Second, once those few do get there, they'll find that the creators don't know how to design a site (downloading the buttons for the various musical catgories takes over a minute), have forgotten to include any content (most of the buttons don't do anything), and either don't speak English or don't know how to use a spell checker.

But you know, despite all the nasty things I have to say about it, all of this is going to happen. There's too much money, and ego, invested in the audio aspects of the Internet for it not to continue lurching forward. There will be a steady stream of marketing efforts, and new delivery systems will be announced. Some will succeed, due to brilliant strategies, brute force promotion, or a fluke in the nature of the audience that no one thought to exploit before. There will be a lot of nonsense that will vanish as quickly as it came, but there will also be genuine technological breakthroughs that will obliterate barriers we thought were impenetrable.

So we might as well take advantage of it. Those of us who have been doing well in the game business, or doing sound for CD-ROMs and other low-bandwidth media, should be able to find clients wanting audio for their Internet presence. Of course, first those clients are going to have get over the illusion that they can handle it themselves, which may take some time–tools that make Web design "easy" are proliferating like rabbits, which is why I often find myself yearning for the good old early days of desktop publishing, when at least the garish graphics and unreadable typefaces stood still.

Some of the most promising Internet audio technologies combine audio and MIDI: there are clever things being done by way of background music for Web pages, using downloaded samples and streaming MIDI control data. (Not very good, but clever.) So knowledge of those technologies is more important than ever.

But while the Internet marches on, like Sherman through Georgia, there are a couple of larger issues to think about. First, there's the idea, prevalent in the breathless prose of pop and sci-fi magazine writers, that personal computers are inevitably going to become mere appendages of the World Wide Waste of time. Now, I like being on the Net, but I also like being off it–and if you're like me, you get a lot more done when you're off it. It was liberation from the mainframes of Big Blue and their ilk that spurred the personal computer revolution in the first place, and now we're supposed to believe that the only road to the future is on a network? Sorry, but I consider my computer a creative tool for me, and being able to spend all my time looking at other peoples' stuff is not a feature I asked for. Our mothers told us to turn off the TV and read a book; now we need to turn off our modems and get back into our own minds and hard drives. The Internet must not be allowed to swallow us all up, distracting and titillating us constantly, never letting us think or create.

Second is the fact that it's so damned invasive, in ways that even commercial television, as dumb and offensive as it is, can't hope to be. Would you tolerate commercials running across the bottom of the screen while you watched your favorite TV programs? At least television makes a pretense, sometimes, of being entertaining. What's propelling the growth of the Internet, at least since it has become big business, is not the thirst for knowledge, the impulse to entertain, or the need for community: it's blatant and unsullied marketing–trying to bury you with as much hype and huck about goods and services as possible, while at the same time trying to find out as much as possible about you, so that the hucksters can sell that information to people who can exploit you more. The most successful sites on the 'Net don't even sell anything: they push brand recognition, and that's all, so that when you go into a real store, you'll run for those products, regardless of what else might be available that is better or cheaper. That the Internet, this wonderful knowledge machine, should be driven solely by marketing is about as pathetic as deciding that the primary goal of the interstate highway system was to have somewhere to put billboards–but that's precisely what's happening.

Finally, and more directly affecting our industry, I fear that the increasing use of the Internet to deliver entertainment could result in the splitting of the audio world into two: one consisting of crummy lo-fi on-line stuff, and the other "real" audio–records, concerts, radio and TV, and home theater. What's scary about this is that in our culture, especially when it comes to entertainment, the lowest common denominator tends to push out all competing forms–look at the ratio of the money and energy spent on rock and roll to that spent on "serious" concert music in this country, for just one example. If that principle holds true as people start confusing Internet audio with the real thing, those of us who care about fidelity, production values, art, and craft, are going to be very lonely.

Next month, I admit that there are some things the Internet is really good for. Stay tuned.


Paul D. Lehrman, composer, writer, and part-time college professor, surfs the 'Net from several undisclosed locations in Massachusetts. None of his music is on the Web, yet.

These materials copyright ©1997 by Paul D. Lehrman and Intertec Publishing