Mp3s have embedded themselves into daily musical life so thoroughly that they’re taken for granted. They don’t call attention to themselves as aesthetic objects like records; instead, we interact with them as files, documents, icons. They’re infinitely duplicable and available in seemingly infinite numbers. At the same time, however, mp3s are intensely controversial. More than a decade after the government took down Napster, these digital objects remain at the center of intense debates about morality and musical value; they’ve eradicated the scarcity of music recordings, doing away with more than a century of economic and social realities about how, when, and where music can be accessed. And, if you listen to certain types of music or talk to certain people about them, mp3s sound quite horrible.
In his new book, Mp3: The Meaning of a Format, McGill University professor Jonathan Sterne exhaustively and eloquently traces the history of the mp3 from the initial hearing model developed in Bell Labs to the current debates about piracy. As the author argues, each time we rip a CD to our hard drives, we’re not only saving space in our living rooms or ensuring we have the appropriate gym soundtrack, but also reaffirming a fundamental idea about the limits of human perception.
“With an mp3, you’re not going to have the same enjoyment that vinyl collectors talk about in terms of physicality, though, for a lot of people, that pleasure has been transferred to computer electronics.”
Pitchfork: People often describe mp3s as “dematerialized” or “invisible” music. You argue that they’re actually “things” comparable with CDs, though.
JS: They’re just different kinds of things. You can hold an mp3 in your hand, you just need some kind of container for it; Apple developed an entire campaign about the thousands of songs you can hold in your hand. There’s a materiality to them because you run out of space on your hard drive, or you can or can’t stream them depending on bandwidth. But the important thing is that their materiality comes with completely different affordances than something like a CD.
With an mp3, you’re obviously not going to have the same enjoyment that vinyl collectors talk about in terms of the physicality of the medium. Although for a lot of people, that pleasure has been directly transferred to computer electronics. The whole fashionable-portable-audio-player phenomena is very much of a piece with the enjoyment of records, except that it’s not a subcultural phenomenon, and certainly doesn’t lend itself to the same kind of aestheticism that comes with being a record collector. But the flipside, in a way, is that mp3s are closer to what I would call the social demand for music– the desire to be with music, to move with it and share it. They’re much closer to that than a record is.
Pitchfork: One of the book’s best chapters is about the tests audio engineers ran in 1990 and 1991 to determine the standard for quality.
JS: Actually, the story of music testing comes from a longer history of industrial testing called hedonics, which I trace back to World War II. Hedonics was developed as a part of industrial psychology to try to measure and test pleasure and displeasure. For instance, if you’re packing string beans in cans for G.I.s, how disgusting do they have to be before people won’t eat them? Or on the flip side, a company like Frito Lay has chip testers at the ends of their production lines. The question isn’t, “How tasty is this potato chip?” But: “How does the taste of this potato chip compare with the standard potato chip?”
This was the mode of testing that was used with the mp3. So the first question is, “What’s the difference between one of several possible new coding schemes and CD-quality audio?” It’s always relative to other audio. Really, it’s just different flavors of processing, which is the defining characteristic of contemporary music aesthetics. So much stuff we hear now is so heavily processed. If you heard an unprocessed voice in a music recording, you would think it was wrong, because it didn’t have dynamic range compression, it isn’t heavily edited, there isn’t any artificial reverberation added. It would sound strange to most listeners.
The MPEG listening tests used a variety of recordings: some castanets, a Suzanne Vega or Tracy Chapman tune, Ornette Coleman, a solo bass guitar piece, examples of male and female speech. The testers would listen to each of these recordings, flip back and forth between different coding schemes and the original recording without knowing which was which, trying to determine the difference. They’d do it over and over for hours– the process is quite exhausting and unpleasant. If a tester was correct more often than they would be if they were guessing, the researchers would know there was a perceivable difference.
The question that immediately arises is: Would you get different results with different music and with different listeners? In the case of the MPEG test listeners, we’re talking about professional engineers, people who work for audio companies, or people who work for radio stations. If you look at the aesthetics of the recordings, they all conform to this sanctioned radio aesthetic in terms of how recordings are made and what they sound like. I wasn’t able to hear all the recordings, but if you listen to the commercial stuff, everything’s recorded in exactly the way you’d expect an engineer at a radio station in 1991 to classify as “well-recorded.” There isn’t anything too esoteric or strange.
So by choosing mainstream music as the tests for MPEG audio, they actually produced a format that doesn’t conform to universals of human hearing, but does conform very well to the record industry. In other words, a small number of mastering engineers have determined what good recordings will sound like.
Continue reading the rest of the story on Pitchfork