Tape loops are lengths of magnetic tape spliced end to end so that a recorded sound repeats for as long as the machine runs. The device is mechanically trivial and musically enormous: it supplied the raw method for phasing minimalism, for the first looping delay systems and for some of the most discussed ambient records. A description of the technique notes that a loop is made by cutting a length of tape and joining the ends, and that playing it on a reel-to-reel recorder makes the sound repeat endlessly. The history that follows runs from a scratched record in postwar Paris to decaying 1980s tapes in a New York loft, and it explains why repetition, drift and decay still define a large part of drone music.
Before tape: the locked groove
The loop idea predates tape. In a 2017 article in Organised Sound, Christophe Levaux describes how Pierre Schaeffer noticed in 1948 that dust and scratches on a record made the needle skip back and repeat the same sound. That accident, often cited as the start of repetitive audio history, led Schaeffer to experiment with closed grooves on shellac discs. Levaux adds a necessary correction: closed grooves had been patented in the United States in the 1870s, and looped tape was used in telephony roughly twenty years before Schaeffer. What distinguished Schaeffer was purpose, the deliberate repetition of a sound to reveal its character.
The same article states that Schaeffer set up the first magnetic tape loops in the early 1950s by splicing a recorded section end to end, and that looping was later built into machines such as the phonogène and the morphophone. These devices belong to the world of musique concrète, which the site separates from drone practice in its piece on the differences between concrete music and drone music.
Studio loops of the 1950s
Other composers adopted the technique quickly. According to the Wikipedia overview, Hugh Le Caine’s Dripsody (1955) was built from a single water drop using variable-speed playback, tape loops and 25 splices, and Karlheinz Stockhausen used loops extensively in Gesang der Jünglinge (1955–56) and Kontakte (1958–60). The studio loop was a compositional shortcut: a short event could be repeated, transposed by changing playback speed, and layered.
Terry Riley and the time lag accumulator
Terry Riley turned the loop into a performance system. In 1963, living in Paris, he recorded the Chet Baker Quartet for The Gift, an experimental theater piece directed by Ken Dewey, and processed the results. Sound American’s account says he connected two tape recorders so that the tape ran across both sets of heads, letting the second machine re-record and feed back the signal, with the delay time adjusted by hand. Riley described the early version of this device as the time lag accumulator, and used it on the quartet’s performance of Miles Davis’s “So What.” The result, Music for The Gift, is a layered and abstracted treatment of jazz. His earlier Mescalin Mix was made in San Francisco for a dance workshop run by Anna Halprin.
The lesson for later drone musicians is that a loop need not be a fixed object. Feedback between two machines produces accumulation, where each repetition adds material instead of repeating it exactly.
Steve Reich and phasing
Steve Reich built a compositional method from a mistake. According to NPR’s account, It’s Gonna Rain premiered in San Francisco in 1965 and was based on a 1964 recording of the Pentecostal preacher Brother Walter in Union Square Park. Two identical loops of the voice drifted out of sync on the machines, and Reich later said he slowed one by pressing a thumb against its reel. The shifting relationship, which he heard as a round, became the basis of phasing.
Come Out (1966) refined the idea. As Wikipedia summarizes it, civil rights activist Truman Nelson asked Reich for a tape collage to benefit the Harlem Six, and the piece is built from a four-second loop of Daniel Hamm’s statement. The phrase “come out to show them” is re-recorded on two channels that start in unison and drift apart, then split into four voices and eight until the words dissolve into rhythm and tone. The article also notes that Odyssey Records released the piece on LP in 1967 alongside works by Richard Maxfield and Pauline Oliveros, whose own tape work is covered in the guide to Pauline Oliveros and deep listening.
Phasing depends on small differences in playback speed between two copies of the same loop, which makes the copies slowly separate. The musician controls the process by choosing when to intervene.
Fripp and Eno: tape delay as a duo
In the early 1970s, Brian Eno and Robert Fripp used a tape-delay feedback arrangement to build (No Pussyfooting). According to Wikipedia’s entry, two reel-to-reel recorders were set side by side so that sound recorded on the first was played back from the second and fed into the first again, creating a long looping delay. The sessions took place on September 8, 1972, at Eno’s home studio and on August 4 and 5, 1973, at Command Studios in London, and the album came out on Island Records on November 9, 1973. It has two long tracks, “The Heavenly Music Corporation” at 20:55 and “Swastika Girls” at 18:43.
Here the loop is not a spliced circle but a feedback chain between two machines. The performer builds layers in real time, which is the approach later adopted by guitarists using looping pedals.
Music for Airports and incommensurable loops
Eno applied a different loop principle to Music for Airports. A technical breakdown on Reverb Machine explains that the track “2/1” uses seven loops of different lengths, each repeating a single sung note, performed by three women and Eno. In a 1996 talk Eno gave durations such as 23 1/2 seconds, 25 7/8 seconds and 29 15/16 seconds for some of them, and described the lengths as incommensurable, meaning that the loops are unlikely to realign. The tapes reportedly ran around tubular aluminum chairs in Conny Plank’s studio, so the timings drift, perhaps because of how the tape was wrapped around the chair legs.
The method produces a piece with no fixed form. Because the cycle lengths do not divide evenly, new combinations of notes keep appearing, which is exactly the generative logic that attracts drone and ambient composers.
William Basinski and the loop that dies
William Basinski reversed the relationship between loop and medium. The Quietus interview says he transferred tape loops made in New York in the early 1980s during a session in July 2001, found the oxide shedding as the tape passed through the machine, and let the melodies decay on playback. He decided to stop adding anything and let the loops run. On September 11, 2001, he and neighbors watched the attacks from his roof while the music played.
Basinski released the loops himself starting around 2002, and a deluxe box set followed in 2012, according to the same interview. The resulting series is covered in the glossary entry on The Disintegration Loops. The work turns a loop’s usual stability into a measurable physical process: each pass removes a little more of the recording, so the piece is also a record of the machine wearing away its own source. More background is in the glossary entry on William Basinski.
Practical notes and limits
Loop length is constrained by hardware. On a standard reel-to-reel machine a loop can only be a few seconds long, and using two machines side by side extends the length, as noted in the Wikipedia overview. The same article records a 76.62 meter loop made by Hainbach and Look Mum No Computer in 2023.
The overview notes that analog sequencers, digital sequencers and MIDI ended the tape loop era in mainstream production. The same ideas survive as delay with feedback, copies of a sample at slightly different rates and loops of unequal length. Readers who want to hear these ideas in context can start with the essential drone music albums, then return to the originals described above.
