An integration loop

Westernmost Colos­sus of the Tem­ple of Re

Westernmost Colossus of the Temple of Re, 1850, Maxime Du Camp

I encourage you to press the play but­ton below, then con­tinue read­ing. Later, you’ll find another ver­sion of this piece, trans­formed by con­tri­bu­tions from dozens of other people.

Meet a melody:

(Alas! Your browser didn't run the JavaScript that was supposed to generate musical notation here.)

Its ori­gin is a mystery. Composer, performer, record­ing: all unknown. The year of its cre­ation is unknown. All we know is how it sounds: stately and nostalgic.

And we know where it played, at least once.

Monitoring a radio sta­tion in New York City, the com­poser William Basin­ski hears the melody, records it. He intends to use a frag­ment as a loop in an avant-garde music project. The tape goes into a box. It is the 1980s.

It is decades later: the sum­mer of 2001. Dig­i­tiz­ing a room full of for­got­ten material, Basinski finds this loop again. But the tape is old; as it moves through the player, it starts to come apart, the mag­netic medium peel­ing off its plas­tic backing, more and more with each repetition. Enthralled, Basin­ski keeps record­ing as the melody dis­in­te­grates before his eyes, his ears.

Standing on his rooftop in Brooklyn, Basin­ski watches the World Trade Cen­ter collapse. It is Sep­tem­ber 11, 2001. Suddenly, the sum­mer’s record­ings have a meaning, a purpose. He titles them The Dis­in­te­gra­tion Loops and offers them as an elegy for the dead. They are heart­break­ing and, before long, beloved.

Seated inside the Metropol­i­tan Museum of Art’s iconic Tem­ple of Dendur, the Word­less Music Orches­tra performs Maxim Moston’s arrangement of the first reel of The Dis­in­te­gra­tion Loops. After the last note sounds, the audi­ence sits for two full minutes in silence, then bursts into applause. It is Sep­tem­ber 11, 2011.

Listening to NPR, I hear this performance. I save the MP3 (forty minutes long!) and return to it often. It is my pri­mary expe­ri­ence of The Dis­in­te­gra­tion Loops. I truly love it.

Years later, work­ing in my media lab, I feed NPR’s record­ing into a neural net­work — a rough AI. The com­puter struggles; it doesn’t know any­thing about notes or scales, horns or drums, tape loops or memo­r­ial concerts — only samples, the stand­alone grains of dig­i­tal sound. But, in its struggle, I hear some­thing evocative: an uncertainty, a brokenness, that seems to match the mythos of Basin­ski’s piece. It is the sum­mer of 2019.

As the network churns, I ask it to generate short pieces, hun­dreds of them. For several months, this is my morn­ing ritual: wake up, make coffee, and lis­ten to the pieces the network has gen­er­ated overnight, reject­ing most, sav­ing one or two. From that curated col­lection, I select about a dozen and sequence them using Able­ton Live. I use sev­eral plu­g­ins to process and mas­ter the sequenced piece, mind­ful not to smooth over the crunch and hiss that is the neural net­work’s signature, and then I publish it.

So. This melody, which was played at some point on a real horn — well, maybe it was real; who knows? — has now passed through radio waves and mag­netic tape and dig­i­tal memory, into the mind of an arranger and back out into the phys­i­cal world — an echo in the Tem­ple of Dendur — then through the inter­net and now into a neural net­work. Yet, through all those degra­da­tions and digitizations, resus­ci­ta­tions and transformations — a heck of a flip-flop—it has never suc­cumbed to noise; not all the way.

A stu­dent at the Uni­ver­sity of Vir­ginia transcribed the melody that Basin­ski recorded, and I played it once myself on a mod­u­lar synthesizer. That’s what you’ll hear (or have heard) at the end of this piece: the ini­ti­a­tion of a new loop.

An integration loop.

After I published the first ver­sion of this post in April 2020, I invited any­one read­ing to join me by play­ing or singing the melody once through — 

(Alas! Your browser didn't run the JavaScript that was supposed to generate musical notation here.)

—using what­ever instru­ment (including their voice) and record­ing device (including their phone) was clos­est to hand. I received many con­tri­bu­tions, and together, they became this:

This piece com­bines con­tri­bu­tions from Aaron Lammer, Adam Roberts, Albert Kong, Alexandra Grabarchuk, Angela Winter, Anna­beth Carroll, Alex Hern, Alek­sandr Maletin, Anthony Weintraub, Chris Schlusser, Dan Bouk, David Renshaw, Deli­cious Democ­racy (Brianna McGowan and Sam Bonar), Evan Goldfine, Greg Baker, Gueorgui Tcherednitchenko, Hay­den Higgins, James Finnis, Jame­son Brown, Jason Black, Jennifer Stout, Jeremy Brooks, Jeremy Keith, Jessica Spengler, Joe Iovino, Jonah, Jonathan Troyer, Kevin Evans, Lily Sloan, Mark Sullivan, Marta Kvande, Matt Penniman, Matt Silas, Max Lambertini, MĂ©duse, Michael Ashbridge, Michael Donaldson, Nik Von Frankenberg, Niko­lai Polikurov, Niko­lay Shebanov, Sam Hollis, Shaun Williams, Stephan Terre, Thom Wong, Tilma of Malack (Thane of the Nether), Tomorrow’s Man, Will Fraker, William Bouk, and William Cohen, along with sev­eral con­trib­u­tors uncred­ited at their request.

Listen, and you’ll hear their voices (and syn­the­siz­ers and pianos and guitars — mandolins too) bat­tle the noise, over­match it. In my imagination, each con­tri­bu­tion is a rung in a lad­der out of the pit of con­fu­sion and loss, all of us both (a) car­ry­ing the melody for­ward and (b) being car­ried by it, up towards something new, some­thing whole.

Dah-dahhh … da-da-da, da-da-da!

Amenophis III

Colosse monolithe d'Amenophis III, 1849–50, Maxime Du Camp

March 2020, Oakland