The fundamental unit of electronic popular music is the loop. This puts it at odds with the Western art music tradition, which typically favors linear structures with a narrative arc. Repetition has mostly appeared in classical music at the macro level of phrases and sections. While shorter repetitive cells do appear in classical music, they are not always welcome. The term ostinato, from the Italian “obstinate,” does not connote approval. Popular music (and some minimalist classical) of the twentieth century has been significantly more repetitive, deriving its harmony from western Europe but its rhythms and circular loop-based structures from Africa and the Caribbean. The advent of synthesizers, drum machines and computers has strongly encouraged the trend toward cyclic repetition, since the default output of such devices is the endless loop.

Björk produced relatively conventional dance music early in her solo career, but her use of loops has become more sophisticated and complex over the course of her career. Her 2004 album Medúlla is comprised entirely from vocals, aside from the occasional synthesizer. Some of the songs are traditional songs and choral works, but most are built from vocals that have been heavily edited, sampled and looped in Pro Tools.

Malawey’s article analyzes three songs from Medúlla in depth: “Öll Birtan,” “Who Is It” and “Triumph of a Heart.” Malawey is primarily interested in the circularity of these songs’ chord progressions. She explains their two-chord or one-chord structures as creating a sense of oscillation or stasis, rather than the source-path-destination schema of western classical tradition. While Björk’s harmonies are colorful, they are hardly groundbreaking. What makes Malawey’s analysis valuable for my purposes is that her vocabulary for harmonic analysis is also applicable to the sonic and emotional quality of looped samples.

“Öll Birtan” is built from layers of Björk’s voice singing highly repetitive figures on a single mixolydian scale. There is no beat per se, but there is an implicit pulse. The piece is less a pop song, and more an art piece reminiscent of Steve Reich. “Who Is It” and “Triumph of a Heart” are closer to dance music convention in form, but they use sampled beatboxing and mouth sounds in place of drum machines and breakbeats. Similarly, sampled vocal tones fill the role of basslines and keyboards. These videos use different mixes than the ones on the album, but they’re so delightful you should watch them anyway.

Malaway’s central thesis is that harmonic stasis creates a feeling of timelessness. She quotes Kofi Agawu:

Repetition enables and stabilizes; it facilitates adventure while guaranteeing, not the outcomes as such, but the meaningfulness of adventure. If repetition of a harmonic progression seventy-five times can keep listeners and dancers interested, then there is a power to repetition that suggests not mindlessness or a false sense of security (as some critics have proposed) but a fascination with grounded musical adventures. Repetition, in short, is the lifeblood of music.

Despite the fact that Agawu is dealing primarily with repetition in African music, much of what he says in this passage applies directly to all kinds of music that are based in repetitive processes. Furthermore, his quote explains the appeal of the more literal sonic repetition of looped samples.

When you listen to repetitive music, time is progressing in its usual linear way, but the cyclic sounds evoke timelessness and eternity. Malawey describes harmonic ostinatos as having a feeling of alternating repose and tension. I’ve found this to be true of looped samples as well; the first and third repetitions will have a “call” feeling that is “answered” by repetitions two and four. The continuing reversal of call and response, of front and back halves of a phrase, can evoke other image schemas as well. Malawey lists swinging, fluttering, quivering, jiggling, hovering and flickering as appropriate image schemas for repetitive music.

In other songs not discussed by Malawey, the feeling of stasis brought on by sampled vocals is even more prominent. The atmospheric haze that dominates “Desired Constellation” was created from a sample of Björk singing the phrase “I’m not sure what to do with it” from “Hidden Place” on her previous album, Vespertine. The vocal loop that opens “Mouth’s Cradle” is so severely processed as to be unrecognizable as a voice; it sounds more like a synthesizer sequenced using MIDI, and the bass notes are pitch-shifted to an unearthly depth. Together with the animalistic backing vocals forming the “percussion,” the collective effect is otherworldly.

Malawey does point out the connection between Björk’s radically simple chord progressions and her unearthly timbres. “[N]on-teleological harmonic processes make room for the vocal textures to carry the music forward.” By building a predictable musical foundation under her sonic experiments, Björk leaves the listener with enough attentional bandwidth to absorb the complexities of her vocal manipulation and looping. The familiarity of the voice combined with the “future shock” of radical digital editing challenges the listener, but the repetition makes the challenge a surmountable one, eventually leading to deep musical gratification.

Update: this post is featured on Synthtopia’s Facebook wall, and there’s a lively discussion happening over there. Check it out!

It’s hard for us now to imagine a time when recorded sound was a wondrous technological novelty.

Those gathered around the phonograph were experiencing music in ways unimaginable not so many years before. They were hearing performers they could not see and music they could not normally bring into their homes. They could listen to the same pieces over and again without change. And they ultimately decided what they were to hear, and when, where, and with whom.

Before the phonograph, music was an intrinsically social activity. The only way to hear it was to be in the physical presence of the performers, necessarily placing the experience into a strong social context. Recordings suddenly divorced the sounds from their original context. Now that recordings flit effortlessly to and fro on the web, the original context of their making is even further removed, and often not even recoverable. Katz gives the usual Walter Benjamin quote about how reproduction of art destroys its aura.

“Even the most perfect reproduction of a work of art,” [Benjamin] maintained, “is lacking in one element: its presence in time and space, its unique existence at the place where it happens to be.” Reproductions, therefore, lack what Benjamin called the “aura” of the artwork. From Benjamin’s standpoint this absence is to be lamented. He speaks of the withering of the aura, the depreciation of the artwork, the loss of authenticity, and the shattering of tradition. Benjamin, however, missed half of the equation. True, mass-reproduced art does lack temporal and physical uniqueness, yet reproductions, no longer bound to the circumstances of their creation, may encourage new experiences and generate new traditions, wherever they happen to be… [W]hile recorded music is often decoupled from its origins in space and time, this “loss” begets a contextual promiscuity that allows music to accrue new, rich, and unexpected meanings.

All this freedom can be disorienting. It’s particularly weird to not be able to see the performers’ faces and bodies. Katz ascribes the popularity of music videos to our discomfort with the lack of visual emotional cues in recordings. The problem is that the videos have so little to do with actual music making that they usually contribute to the alienation. The found videos in Auto-tune The News have an authenticity to them that traditional music videos don’t. The Auto-tune on Antoine Dodson’s voice might be fake, but the anger in his face is very real.

When musicians record, their invisibility to listeners removes an important channel of communication, for performers express themselves not only through the sound of their voices or instruments but with their faces and bodies. In concert, these gestures color the audience’s understanding of the music. As Igor Stravinsky rightly explained, “The sight of the gestures and movements of the various parts of the body producing the music is fundamentally necessary if it is to be grasped in all its fullness.” The violinist Itzhak Perlman, for example, is effective in concert in part because his face registers and reinforces every expressive nuance in the music. Perlman himself once remarked that “people only half listen to you when you play—the other half is watching.”

Recording artists have also reacted to the fact that they cannot see their audiences. For many, the task of performing to unseen listeners, with recording equipment as their proxy, can be both daunting and depressing. In her memoirs, French soprano Régine Crespin registered her dismay at the artificiality of performing in the studio:

“Fear of an audience is healthy; it stimulates you. The people are there in front of you. With them there can be mutual lovefests. But how can you fall in love with a microphone? First of all, a microphone is ugly. It’s a cold, steel, impersonal thing, suspended above your head or resting on a pole just in front of your nose. And it defies you, like HAL the computer in Stanley Kubrick’s film 2001: A Space Odyssey, although at least he talked. No, the microphone waits, unpitying, insensitive and ultrasensitive at the same time, and when it speaks, it’s to repeat everything you’ve said word for word. The beast.”

Recordings have created the possibility that a musical experience can be exactly repeated.

Live performances are unique, while recordings are repeatable… [A]ny orchestra can play Beethoven’s Fifth Symphony many times; each performance, however, will necessarily be different.

For listeners, repetition raises expectations. This is true in live performance; once we’ve heard Beethoven’s Fifth in concert, we assume it will start with the same famous four notes the next time we hear it. But with recordings, we can also come to expect features that are unique to a particular performance—that a certain note will be out of tune, say. With sufficient repetition, listeners may normalize interpretive features of a performance or even mistakes, regarding them as integral not only to the performance but to the music. In other words, listeners may come to think of an interpretation as the work itself.

The repeatability of recorded sound has affected listeners’ expectations on a much broader scope as well. When the phonograph was invented, the goal for any recording was to simulate a live performance, to approach reality as closely as possible. Over the decades, expectations have changed. For many—perhaps most—listeners, music is now primarily a technologically mediated experience. Concerts must therefore live up to recordings. Given that live music had for millennia been the only type of music, it is amazing to see how quickly it has been supplanted as model and ideal.

I’ve played in a lot of bands, and one of the biggest obstacles they all faced was that most of the time, people would prefer to hear recordings than to hear us play. I’ve backed some very good jazz singers, but why would you want to listen to them when you could hear Billie Holiday and Ella Fitzgerald? Live music is an invaluable art experience but if you just want to dance or be entertained, DJs beat bands nine times out of ten.

A lot of musicians I know would read the previous paragraph and bitterly remark on the harm that recording has done to musical culture. Maybe they’re right, but recordings have given us a lot too. I think that the experience of repeat listening to recordings has made Western listeners more accepting of repetition as a feature of the music itself. Westerners have historically resisted the loop-based, chant-oriented forms in other world cultures. But American popular music has been getting more and more loop-based, especially in my lifetime. Tape loops and samplers have both created the tools to make loop-based recordings, and fueled the demand for them.

Recording has changed music education radically, putting improvisation and other non-notated forms on an equal scholarly footings with written scores.

With recordings, performers can study, emulate, or imitate performances in a way never before possible. In the early days of recording, this possibility was trumpeted as a gift to all musicians, who could learn from the world’s great masters by studying their discs. For performers of popular music, recordings have been especially valuable learning aids. The available scores do not always represent performances adequately, and they cannot easily indicate the timbres and sonic effects that musicians seek to develop.

Audio editing tools also make it possible to build recordings without an actual “performance” ever taking place.

Listen to most early-twentieth-century recordings and you will hear a performance in the traditional sense. That is, you are hearing a single and complete take, in which the beginning, middle, and end of the piece were recorded in that order on the same day, in the same place, and by the same performer or group. This was hardly out of a desire for authenticity; it was a product of necessity. However, since the introduction of magnetic tape (in the late 1940s) and digital recording (in the late 1970s), it has been possible to offer the illusion of a traditional performance as well as to create “performances” that could never have existed. With the ability to manipulate sound through such technology, musicians have been able to transcend time, space, and human limitations, and in the process have created wholly new sounds, works, genres, and performance traditions.

One of the most basic manipulations is splicing, in which passages recorded at different times are joined together. The Beatles’s “Strawberry Fields Forever” (1967) provides a famous example. The Beatles did over two dozen takes of the song, none of which completely satisfied John Lennon. But he did like the first half of Take 7 and the second half of Take 26. So he asked George Martin, their producer, to put the two together. Unfortunately, they were in different keys and tempos. The two takes, however, were related in such a way that when one was sped up and the other slowed down so that the tempos matched, the pitches also matched. Thus the two takes could be joined, the splice occurring at about 0:59 on the word going in “Let me take you down ’cause I’m going to Strawberry Fields.” Although the splice is nearly undetectable, the slightly altered speed of Lennon’s voice helps give the song its distinctively dreamlike quality.

You can build recordings entirely from samples, synthesized and sequenced sounds and jamming. It’s common in pop, dance music and hip-hop to go into the studio with no prepared material at all, and come out with finished songs. Songwriting, improvising and recording have collapsed into a single act.

The job of assembling an electronically-oriented recording is a new form of musicianship unto itself.

It is important to realize that sound is manipulated in the studio not (or not typically) by performers, but by a variety of sound engineers and producers, sometimes referred to collectively as recordists. Recordists fall outside (or perhaps in between) the traditional triad of composer, performer, and listener. They might be thought of as sound shapers, artists in their own right who collaborate with performers and composers.

[T]he very possibility of manipulating sound after its creation—from splicing to digital pitch correction—forces us to reformulate our ideas about composition, performance, and the relationship between the two.

[R]ecording does more than influence the activities of composers, performers, and listeners. It affects the relationship among these actors and in fact challenges the stability, even the validity, of the triad. It is no longer necessary for listeners and performers, or for performers and composers, to work together in order to create music. Yet at the same time, listeners and composers have discovered a more intimate relationship, one that can bypass the mediation of performers, while performers can work in solitude, without the need to stand before listeners. Performances and works are no longer clearly distinct, for recordings can take on the function and meaning of both. Just as recordings can be heard as spontaneous interpretive acts, their repetition can transform them into compositions, works that can be analyzed, historicized, canonized, politicized, and problematized. Nor are production and reproduction so easily separated when preexisting sounds can be manipulated in real time.

A lot to think about. But not everything in the book is so serious. From the introduction:

Several years ago a friend asked me to explain the subject of this book, then in its early stages of development. Opting for a dramatic approach, I pulled a CD at random from a nearby shelf and brandished it in front of me. “This,” I declared, “has changed the way we listen to, perform, and compose music.” My friend squinted at the CD, gave me a quizzical look, and asked, “That did?” “Yes!” I answered with gusto. Seeming unconvinced, he clarified his question. “Van Halen changed the way we listen to, perform, and compose music?”

Yes, yes they did.

The brain is a pattern recognition machine. We like repetition and symmetry. But we only like it up to a point. Once we’ve recognized and memorized the pattern, we get bored and stop paying attention. If the pattern changes or breaks, it grabs our attention again. If the pattern-breaking happens repetitively, itself forming a new pattern, we find it super gratifying.

Count to four, over and over again, at a rate of one beat every three quarters of a second or so. On beat one, play a low-pitched percussion sound. This could be a kick drum, a foot stomp, a grunt or ‘bff’ sound, or the bottom of a paint bucket, whatever. On beat three, play a higher-pitched sound: a snare drum, a clap or snap, a plosive syllable like gh or kh, the rim of a paint bucket. This pattern, alternating high and low, will get monotous quickly. To make it funky, you need to introduce some predictable unpredictability. The best way to do that is to change the placement of every alternating bass drum hit. Anticipate it by a beat, or half a beat. Delay it a beat or half a beat. Or do both: play a kick both on beat four and a little before the next beat two. Or just skip every other kick. Maybe also add an extra snare hit in an unexpected place: the very last beat of every other phrase, for instance.

For another layer of texture, add an even higher-pitched sound on every beat, or every half beat. It could be a hi-hat, a shaker, a tambourine, a plosive like ‘ts’, or the metal handle on the paint bucket. Maybe even try doubling up the tempo for a beat at the beginning or end of the phrase. And that should be it for complexity. Cluttering up a beat too much diminishes its power. The silences between the drum hits are as important as the drum hits themselves. Leaving out every other kick drum hit is an effective strategy because the ear fills in the silence by itself, and that imaginative participation is a big pleasure.

There’s a natural tendency to want to immediately start putting in variations, embellishments and fills. We get nervous that the beat isn’t going to be “interesting” enough. It’s true that patterns get monotonous after a while, but too much variation releases the tension that a good pattern can build. If the pattern gets on your nerves quickly, find a better pattern. Rather than let the air out of the balloon too early, you want to pick your symmetry-breaking spots carefully: in the last measure of a sixteen-measure phrase, or even better, a thirty-two bar phrase. Building up lots of tension gives the listener greater satisfaction when it’s finally released. And when it comes time to throw in a variation, using unexpected silences is more attention-grabbing than adding extra sounds. Eliminate the kicks, or the snares, or everything but the hi-hats, or drop the beat out entirely but keep the pulse going, and watch the crowd react with pleasure when the full beat kicks back in.

My musician friends complain that drum machines don’t sound as good as human drummers because they’re so predictable and robotic. I myself have not found this to be true, unless the human in question is a deep Jedi master like Clyde Stubblefield or Questlove. The best drummers are the ones who play the most tightly and predictably. Questlove in particular got to be as good as he is in part by listening to a lot of drum machine music, and striving to sound like one. My friend Doug, one of the best drummers I know, used to leave his metronome on for hours at a time while he did stuff around the house. The beauty of drum machines and the software that emulates them is that they discourage you from needless complexity. They make it effortless to program a two-measure pattern, and annoying to vary or embellish it.

American popular music in all of its varieties mostly agrees about the power of four:

Every beat is four sixteenth notes long.
Every measure is four beats long.
Every phrase is four or eight measures long.
Every section is four or eight phrases long.

Sometimes you see sections that are three phrases long, as in twelve-bar blues. Very rarely a more eccentric phrase length shows up. But most of the music that most of us like here in the western world is based on powers of two. Instead of breaking symmetry totally, you can create extra tension by gently bending the form. You can lay behind the beat or lean out in front of it. You can stretch out every other beat a little, a practice known to musicians as swing. Quality drum programming software lets you bend and stretch time in musical-sounding ways without fracturing the pattern and breaking the spell.