Noise for Airports

Pianistic Translations

In doing research on the player piano, a certain temptation has come up many times. Given the popularity of the phonograph as an object of academic inquiry (and the persistence of its basic working principles), it is basically mandatory that I compare the pianistic reproduction I’m looking at to phonographic reproduction.

So first, there is a question: What kinds of things am I comparing? I just called my topic “pianistic” reproduction, which is basically a working term meaning “with discrete notes and attacks, like a piano.” “Phonographic” reproduction, on the other hand, would mean “like a phonograph.” Basing my terms specifically on the technologies is not ideal: like a piano in what way? like a phonograph in what way? Jonathan Sterne did the hard work for the phonograph and ended up with “tympanic reproduction”—sound reproduction that is modeled on the eardrum (like all microphones and speakers). That seems to collect together iPods, phonographs, 8-track tapes, etc. in a meaningful way—based on a foundational shared principle. At the moment, my only parallel move would be to alter “pianistic” so that instead of referring to “piano” the technology, it refers to “pianism” the mode of engagement with keyboard instruments. This is a little obtuse, still working out terms, but I hope you get the idea.

Click through for a bunch of examples, after the jump.

Read More

Machines to Piss Them Off

This weekend, I was at the New England Conservatory to see a performance of George Antheil’s Ballet Mécanique. The piece, originally written in 1924, called for a siren, several electric bells, two airplane propellers, a set of xylophones, a pair of live pianists, some bass drums, and sixteen player pianos. Syncing the pianos proved impossible at the time, so eventually the piece was performed with a single player piano. In the 1990s, Paul Lehrman and Schirmer revisited the piece with MIDI, used the magic of computers to sync up the pianos, and the rest is history (or a series of concerts).

There is a certain pleasure in being able to finally play a piece as the artist “intended”—Antheil never heard the piece as originally scored. But, of course, “intent” is a tricky thing. What do you do if an artist intends something that is not possible? The modern, MIDI-enabled Yamaha Disklaviers are not the same instruments as the player pianos of the 1920s. They can be synced together, and they are certainly pianos, but they do not function in at all the same way as the pneumatic contraptions Antheil would have known. One effect of this, I realized part way through the performance, was that Antheil had no way of knowing what 16 player pianos playing in sync would sound like.

It turns out, sort of muddy. The performance I saw had just 8 Disklaviers (the player piano part is written for four voices, with four pianos each, but we just had two pianos per voice), but even so, when all those sustain pedals are down and the pianos are hammering away, it sounds like mush. The rapid-fire melodies and clangorous chords just form a sort of “mrrrrrrr”-sounding backdrop to the much more lively playing of expert xylophonists and noisy sirens and bells. Antheil wouldn’t have known this—by intending something not possible, he had no way of evaluating its sound.

Not that I minded the muddiness. What it made plain to me, though, was how Antheil’s concern was not so much sonic as spectacular. The idea of sixteen player pianos bashing away at once is more important than the sound of them. As a contrast, take Conlon Nancarrow’s pieces: they play tons of notes, with the sustain down (and in a few cases call for two player pianos at once). However, Nancarrow modified his pianos, putting metal strips on the hammers of one, lacquered leather and tacks on the hammers of the other. These modifications resulted from actual experience with the sounds of his pieces, and a concern for them as sounds. The hardened hammers made for more clearly defined attacks that stand out above the muddy resonance of sustained notes.

As Paul Lehrman told it in a talk before the concert, Antheil wanted to cause a riot. He was considered an enfant terrible of the music world, known for putting a gun on top of his piano before a performance, so people wouldn’t leave. (And his eventual autobiography would be titled Bad Boy of Music.) The piece was to be premiered at the Théâtre des Champs-Élysées, where Stravinsky’s Rite of Spring had caused riots a decade earlier. Listening to the piece, the riot-causing intent is quite legible—the siren sounds totally outrageous, the electric bells unreasonably loud. He succeeded, solidifying his reputation as a guy who liked to piss people off. Unfortunately, the US premiere in New York was advertised as a guaranteed riot, and the folks who showed up, apparently un-enraged by Antheil’s musical tantrum, didn’t riot. This solidified Antheil’s reputation as a guy who liked to piss people off, but not always how he intended.

My advisor, who was at the concert, asked me whether Antheil had some over-arching theory of mechanization in music. Player pianos, sirens, electric bells, and airplane propellors would seem to indicate that he had at least some predilection for the mechanical. (Not to mention the xylophone parts, which, according to Paul Lehrman, though played by very dextrous humans, were originally intended for xylophone-playing machines.) It seems that, for Antheil, using machines was just a means to provoke (although this is almost surely an over-generalization). The pianos bash away, and at one point, the human performers are silent while the pianos play solo, with excruciatingly long silences. The conductor stands somewhat helpless, while the audience waits for the inscrutable machines to get on with it.

Turning over all kinds of musical agency to these machines, Antheil seems to consider this dynamic of machine control as something to exploit to provocative ends. (Ironically enough, as he was eventually unable to control his player pianos enough to get them in sync, and at the New York premiere, an inexperienced siren player missed his cue, sending wailing noise through the rapidly emptying house.)

Ballet Mécanique was originally conceived as the score to a film of the same name by Fernand Léger. The two ended up going in separate directions and producing independent works (the film is about 15 minutes, the score about 30), but various attempts have been made to match up some of the music with the film. Embedded below (dashboard people, you’ll have to go to the actual site), you can watch the film in two parts and hear the piece, as recorded with a single player piano.

Mechanical Fidelity

This post is the first of many to come that will be little snippets of thesis-related thought. I’m focusing on a series of objects, people, and practices from the long and weird history of player pianos, and these blurbs are what come out. They’re in no particular order, and they might not make sense individually (although I hope they do). Comments on any aspect are welcomed!

Mechanical Fidelity

[photo ©Mark Manring]

Zenph Studios has a peculiar business model. From their website:

Zenph® Studios takes audio recordings and turns them back into live performances, precisely replicating what was originally recorded. Our software-based process extracts every musical nuance of a recorded performance, and stores the data in a high-resolution digital file. These re-performance files contain the details of how every note in the composition was played, including pedal actions, volume, and articulations – all with millisecond timings.

These digital files are played back on modern player pianos and recorded for sale. Zenph focuses on recreating the original operation of the piano—improvements to recording technology ensure that the new recording will be an improvement over the old. However, the idea of “re-performance” and the relationship between player pianos and acoustic recording have a long history.

Read More

Data Frenzy at Sanders Theatre

Last night, I had the pleasure of being at the North American premiere of Japanese noise/data artist Ryoji Ikeda’s datamatics [ver. 2.0]. The “performance” was really more of a film: the audience watched a large, high-resolution projection of data, while listening to Ikeda’s signature electronically-produced sounds. (To complicate things a bit, the web page for datamatics says that “Ikeda employs real–time programme computations and data scanning” to produce at least some of the visuals,” so it’s not entirely like a film, I guess.) Ikeda was nowhere to be seen, unless you were one of the few people who realized that he was actually the quiet Japanese man by the projectors.

Ikeda’s work thrills at the edges of the sensorium, massing individual sounds until they blur into a mass, playing frequencies at the upper limit of human hearing, and running data across the screen at an ever-quickening, rapidly illegible pace. Warnings were posted at the door about strobe effects and loud volumes (woe to the accidental epileptic who encounters Ikeda’s work unaware), although neither were to be found in unusual amounts for an audience familiar with rock concerts. (My one complaint, and a small one, is that there was not enough volume; Ikeda’s work in bass registers is phenomenal, and the speaker system just wasn’t enough to rattle the audience. I heard that this was perhaps a concession to the fragility of the stained glass windows in the theater—understandable, but a pity.)

The work went through several “movements,” for lack of a better word, revolving (sometimes literally) around the stars, the genome, and then the show itself. This shifting in scale, from the microscopic to the macroscopic and referential to self-reflective, was a central concern, as was the palpably physical effect of such a data deluge. Having gone to the show with a synthetic biologist, I can report that at least one scientist enjoyed the implicit commentary about the incredible mass of modern scientific data, and the effect that these data have on perception. Ikeda may race through the genome in a matter of minutes, but even with years it stretches the limits of comprehension.

Of particular interest was the venue’s role in the history of modern data, and the invention of the science of acoustics. As Emily Thompson describes in her book, The Soundscape of Modernity, Sanders Theatre was considered “acoustically superb” and the president of Harvard wanted to identify what made it that way. Physicist Wallace Sabine was tasked with quantifying its acoustic properties, and he undertook extensive experimentation to measure reverberation.

Sabine amassed a great deal of data, and obsessed with the precision and accuracy of his measurements, “once threw out over three thousand measurements, representing several months’ work, after determining that the clothing worn by the observer (himself) had a small but measurable effect upon the outcome of his experiments” (Thompson 36). Seemingly overwhelmed by his data, Sabine had a difficult time determining how the properties of a room affected reverberation, until one day, he figured it out:

I was floundering in a confusion of observations and results which last night resolved themselves in the clearest manner. You may be interested to know that the curve, in which the duration of the residual sound is plotted against the absorbing material, is a rectangular hyperbola with displaced origin. […] This opens up a wide field. (Thompson 39)

Sabine’s organ pipe sounded data out of Sanders Theatre at the turn of the 20th century, bringing the former art of acoustics closer to science—the “confusion of observations” to the “clearest manner” of equations. Ikeda’s datamatics, at the start of the 21st century, condenses enormous masses of data back into confusion, toying with the clarity and discreteness of data and bringing it closer to art.

(You can watch a lecture by Emily Thompson on her book here.)

Machines to Interpret for Them

I’m reading Inventing Entertainment, an enthusiast’s history of the player piano, and while it is generally light reading (most interesting for anecdotes), I’m occasionally finding really provocative stuff in it. One such passage, an excerpt from Stanislaw Lem’s His Master’s Voice:

“By chance,” explains the main character, Peter Hogarth, an eminent mathematician, “a program tape for a digital computer might also fit into a player piano, and although the program has nothing, absolutely nothing, to do with music—it might refer to some fifth-order equation—nevertheless, when it is put in the machine, it produces notes. And it might also happen that not all the notes thus produced will be in total chaos, but that here and there one will hear a musical phrase.”

The book is about scientists trying to decipher messages from outer space—to convert them from one form of encoding to another that we can understand. This passage is a great literal example of Stuart Hall’s encoding/decoding model of communication.

In Hall’s model, messages are always encoded by the sender and decoded by the recipient. This encoding and decoding can be sociocultural (like in-group language, or based on common cultural references), and that is Hall’s main focus, but the model also works for technological encoding and decoding. What Hogarth describes in the passage above is the co-dependence of data and decoder for meaning making. The punch card on its own does not mean anything; only in conjunction with the computer does it mean mathematically. In conjunction with the player piano, these holes could signify music.

This idea is actually carried out in Yasunao Tone’s Musica Iconologos. The album is based on two Chinese poems. Digital images of the poem’s ideograms (or perhaps of photos representing the ideograms, I’m still unclear on that) are the equivalent of Hogarth’s computer tape. These binary files, instead of being read to create images, are read as audio data in a different computer program. Through what Hall might call an “oppositional reading,” data encoded for one purpose is decoded for another, creating a screechy, noisy, Fluxus-y composition.

(That’s an embedded Flash player with an excerpt, for you RSSers and Tumblr dashboarders.)

In spite of Tone’s avowed interest in indeterminacy (he was loath to make recordings, since they are so repeatable—this data composition was a way to remove intentionality), the liner notes for Musica Iconologos reveal the human hand that intervened in this otherwise pure replacement of decoding processes:

The sound files produced were very short in duration, averaging out around 20 milliseconds in length. […] My task as the digital editor and sound designer was to uncover and shape the larger sounds that lay within each short 20 ms burst.  I accomplished this by expanding the sound to a length that best fit the meaning of the word or picture the sound represented.  Then, if necessary, I digitally mixed or merged several projections together to achieve a desired grouping of data, following the structure (word groupings) of the poem and the implicit meaning of the particular word or picture in question.  Also, where appropriate, digital pitch shifting was applied to certain sounds in order to reflect the phonetic implications of the spoken Chinese word.  It should be noted that there was never any exclusion or repeated inclusion of sounds based on their final result.  To his credit, Tone always remained true to the poem’s structure regardless of his personal impressions of the music, and in a sense the sounds were a type of “chance operation” in form, as their final organization was established long before the project began production.

That quotation is from the sound designer for the album, Craig Kendall. Because the sounds produced by a clean swap of decoding software were too short, he had to lengthen them. Beyond that, and quite surprising to me, is the discussion of how pitch and structure was altered for artistic effect. It is rather telling, I think, how apologetic he becomes at the end of this passage, trying to recover the idea that something with so much human judgment can remain a “chance operation.” (Of course, even things that seem more obviously to be chance operations are never really devoid of human influence.)

I’m always looking for more examples of this kind of data music, so if you know of anything cool, drop me a line!

Machines to Listen for You

A recent story on NPR’s Morning Edition describes a new software system to analyze songs for hit potential. Quoth Music Intelligence Solutions CEO David Meredith:

“[It’s] a series of algorithms that we use to look at what’s the potential of a song to be sticky with a listener,” Meredith says. “To have those patterns in the music that would correspond with what human brain waves would find pleasing.”

Meredith’s software uses algorithms to model how a brain finds “pleasure.” Even putting aside the strange attribution of taste to brain waves (as opposed to just the brain, or—I don’t know—the person?), this model has a lot in it to talk about. First off, the idea that pleasure is caused by “patterns in the music.”

What kind of patterns is he talking about?

Meredith says his software found that hits have certain common patterns of rhythm, harmony, chord progression, length and lyrics.

The choosing of parameters, even seemingly innocuous ones like “rhythm” and “harmony,” is an interpretive move. These parameters define the way the software can “hear,” grouping or separating various songs in contingent ways. Of course the story is vague about how exactly the software parses patterns in “lyrics” or “length”—you wouldn’t want to give up any trade secrets. But let’s take length as an easy one.

How would you incorporate length into an evaluation of a song’s “hit score?” It’s presumably a single number. Is shorter better? Maybe the ideal length is determined by a relationship to the various other scores? How do you extract a hit rating from it? By embedding these questions in software, Music Intelligence Solutions obscures them. The software must have a methodology, just as the music critic would, but since it outputs a score—7.6 out of ten is apparently “good for a platinum rating”—this methodology goes unexamined.

David Bell, of the hip-hop duo the Block Scholars, paid $90 to use it.
“To me, it’s an unbiased validation of your music,” Bell says. “It’s not your family turning around and saying, ‘Oh, you got a great song.’”

The computer told Bell he had a 7.1 — good, but not great. So he went back to the studio and remixed. He got his score up to 7.6 — good for a platinum rating. He could hold his head up.

The software configures musical production in a particular and contingent way. Bell produces a song, the machine evaluates it, he remixes, and the machine reevaluates. In the context of evaluative software, the labor of remixing is a negotiation with the machine.

The assumption is that the machine that turns music into numbers, processes them, and gives back a number, must be “unbiased.” By displacing evaluation from people to “a series of algorithms,” Music Intelligence Solutions banks on the obscuring power of technology: “your family” has bias, but a machine does not. Of course, the workings of the machine are entirely sculpted by bias. Each decision about what matters in a song is tweaked and informed by the programmers. The algorithms do not make themselves, nor do they decide which musical traits are significant.

Objectivity is produced through the use of numbers and software interfaces. Technologically mediated bias becomes objective evaluation. Software like Music Intelligence’s UPlaya propagates a certain view of what “hits” are, and even exceptions to the rule are integrated into a particular musical viewpoint:

It doesn’t surprise New Yorker music critic Sasha Frere-Jones that a computer can predict hits, but he says it can’t predict all the hits. Sometimes, songs come along that don’t fit the mold.

Songs that are popular in spite of their evaluation results (Frere-Jones uses the example of “Da Da Da” by Trio) become songs “that don’t fit the mold.” Now, not fitting the mold is another story about how songs become popular. On one side, we have software to account for certain kinds of success (like “I Gotta Feeling” by The Black Eyed Peas—8.9 out of 10), and on the other, a popular cultural myth about the appeal of mold-breaking. Evaluative software recasts songs that don’t meet its model as part of another cultural narrative.

Is this something to worry about? Probably not. For the integration of taste into business decisions, numbers can be very useful—they turn the the qualitative into the quantitative, with its sheen of objectivity. For listeners, though, I find it hard to imagine that you’ll be seeing ads for “The new Britney Spears: 9.1 out of 10 on the hit scale! You’ll Love It!” Actually, I take that back. You might really see these things, like impact factors on science journals, or prominent magazine reviews on CDs. If they actually use these in promotion, and not just in house, that would be both totally wild and fodder for like a million more blog posts.

Machines to Play Machines for Them

The reason posts have been a little light lately (not in number, but in content) is that the thesis season has begun in earnest. A little over a week ago my classmates and I gave our preliminary thesis presentations, making a “public commitment,” to use the words of my thesis advisor. This post is an adaptation/abridgement of that presentation into blog format, as another form of “public commitment” and as a place to direct people when they ask me “So what is your thesis about, anyway?”

The Musical Gorilla: Musicality and Automaticity in Mechanical Musical Instruments

In H.G. Wells’s Tono-Bungay, a woman sees a pianola in a man’s living room. She asks him, “Does this thing play?” He replies,

Like a musical gorilla, with fingers all of one length and a sort of soul.

The research I’m doing revolves around statements like this—evaluations of machines that make music. I’m interested in how the various parties involved with these instruments (makers, marketers, users, listeners, and so on) think of them. Are they machines? musicians? Should their music be evaluated like human-produced music? If not, then how should one evaluate machine-produced music?

Only when you can forget your fingers can your brain be perfectly free. It surely stands to reason, then, that the ready made technique of the player piano sets the musician’s brain free to attend to the purely artistic side of the performance.

-Ernest Newman

The gramophone concerns itself with “sound” in a general sense, reproducing it as it is heard, rather than as it is made (this is why your speakers don’t look like little mouths, violins, pianos, etc., but rather like abstracted ears). The various iterations of player pianos, on the other hand, are concerned with how a particular kind of sound is produced (this is why player pianos generally look like—yes—pianos).

There are a lot of good books written about the early history of sound recording, from all sorts of musicological, historical, and theoretical perspectives. When I wrote my undergrad thesis, I had to grapple with many of them. The player piano, on the other hand, is less thoroughly theorized (probably for good reason—How long has it been since you’ve seen a player piano in person? What about an iPod?).

For me, though, the absence of big, authoritative works is a blessing. The various ideas about sound reproduction from the numerous authors who’ve theorized it can come into play without crushing me under the weight of a thousand master’s theses about the history of recorded sound. Because, on a fundamental level, the work of machines like the pianola (or Pianola) is different. These are not machines that try to replicate and model natural phenomena: these are machines that play other machines.

“Where, in truth, is the non-mechanical musical instrument? […] The anti-piano-player pianist is, in fact, a million removes from mere nature; he would be helpless without the huge box of mechanical tricks in front of him.”

-Ernest Newman

A fundamental idea in the kind of technology studies-informed work I do is that machines are not cold, objective things that go about their existence independent of humans. In their use, manufacture, and physical composition, they reflect the influences of countless actors. This is as true of the piano as it is of the automatic piano-player. The piano represents a certain kind of musical interaction—the pressing of keys to call out notes in specific intervals, the ability to change dynamic level, the sustain pedal. The player piano is a physical reading of the piano’s interface, which in its design reflects decisions about what is “musical,” “essential,” “artistic,” or “automatic.”

So what?

“Everybody is an electronic musician. Some of us push only one button in a performance, and some of us push many buttons.”

-Vladimir Ussachevsky

In the modern musical environment, automated music is basically omnipresent. In addition to mp3 players and car radios, we have music games like Guitar Hero or Rock Band that reinsert physical machinery into the path of sound reproduction. Every song now on the Billboard charts (and elsewhere, for that matter) arises from the manipulation and interaction of machines to produce sounds, and with effects like Auto-Tune, this machinery comes to the foreground in new ways. The player piano plays a formative role in our modern relationship to musical machines (I think); by looking at this example at a historical remove, hopefully I can shed some light on our current music listening and making situation.

While exciting, this is all quite broad now (and the stuff in this blog post is maybe half of what I presented last week), so it will surely contract and expand in various ways and areas many times before I’m through.

Projects drift; that’s why they’re called research projects.

-Bruno Latour

Stay tuned for more exciting player-piano based updates with the thrilling “thesis” tag. (They’ll hopefully come in more manageably-sized chunks after this one.)