A history of sampling

The gear in a modern sample-based setup descends from decisions made across eight decades about one question: what do you do with a recorded sound? Pierre Schaeffer asked it in a Paris radio studio in the 1940s. Roger Linn asked it again in 1988. The Eurorack designers asking it now are working in the same tradition, and the answers have accumulated into the hardware and workflows we use today.

This is a survey of how sampling developed: the instruments, the techniques, the legal battles, and the cultural moments that shaped what the word “sampler” means now.

Tape and the birth of musique concrète

The first samplers were tape recorders. In the late 1940s, Pierre Schaeffer was working as a sound engineer at Radiodiffusion-Télévision Française in Paris when he started experimenting with recorded sound as raw material for composition. His Études de bruits (1948) — “Studies in noises” — assembled everyday sounds, slowed, reversed, and looped using the tools at hand: phonograph discs and, soon after, magnetic tape.

The technique he developed, which he called musique concrète, treated recording not as documentation but as a first step. The recorded sound was concrete: something you could cut, splice, and re-assemble. A train, a bell, a voice — material, not event.

This was a fundamental reframing. Before Schaeffer, recordings were reproductions. After him, they were instruments.

Collaborators including Pierre Henry pushed the ideas further into the 1950s and beyond. Henry’s Variations pour une porte et un soupir (1963), a piece built entirely from the creaks of a single door, is still striking for what it asks the listener to hear.

The tape splicer and the studio became the sampler. Everything that followed is software and electronics working out the same underlying idea.

The Mellotron and the keyboard sampler

The Mellotron (1963, Birmingham, England) was the first instrument to put samples under a keyboard. Each key triggered a strip of magnetic tape containing a pre-recorded sound — strings, flute, choir — that played back for as long as the key was held, up to about eight seconds, before the tape mechanism ran out.

Its American predecessor, Harry Chamberlin’s instrument from the 1940s and 50s, used the same principle. The Mellotron was a refined, commercially available version developed in the UK from Chamberlin’s design.

The Beatles used it on “Strawberry Fields Forever” (1967), the flute sound at the opening. King Crimson’s Robert Fripp used it for orchestra-sized texture from a single player. Most of progressive rock reached for it when they needed strings without a string section.

The Beatles, "Strawberry Fields Forever" (1967) — the flute sound at 0:01 is a Mellotron Mark II.

The Mellotron’s limitations were part of its character: eight seconds per note, mechanical unreliability, temperature sensitivity, the slight pitch drift and breathy quality that came from analog tape running through imperfect transport. These weren’t problems that later got fixed; they were the sound, and musicians used them deliberately.

The insight that outlasted the hardware: a keyboard controller plus recorded sounds equals a playable instrument. Everything from the Ensoniq Mirage to Kontakt to the Make Noise Morphagene runs on the same premise.

Digital arrives: the Fairlight and Synclavier

The first digital samplers were studio tools built for studios with studio budgets.

The Fairlight CMI (Computer Musical Instrument), developed in Sydney by Peter Vogel and Kim Ryrie, arrived around 1979. It was among the first commercially available instruments to use the word “sampling” for what it did. Users could record any sound and play it back at different pitches across a keyboard. Early buyers included Peter Gabriel, Kate Bush, Stevie Wonder, and Herbie Hancock — musicians with both the budget (upward of $25,000 at launch) and the curiosity to push unfamiliar tools.

The Synclavier, from New England Digital, arrived in the same period. A fully configured Synclavier could cost several hundred thousand dollars by the early 1980s. These were not instruments for individual musicians; they were capital investments for professional facilities.

Both machines stored samples on floppy disks and, later, hard drives. The file management problem (what format, what naming convention, where to put things) appeared at the very beginning of digital sampling and has never fully gone away.

The affordable sampler and what it unlocked

The E-mu Emulator (1981), at roughly $8,000, brought sampling within reach of working professional musicians. Still expensive, but purchasable as a musician rather than as a studio executive. It recorded at 8-bit quality, which colored the sound noticeably — thin, grainy — but the concept was sound.

The Ensoniq Mirage (1985) dropped the price to around $1,700. For the first time, amateur and semi-professional musicians could own a sampler. The Mirage was 8-bit, limited in memory, awkward to program — and enormously influential, because the barrier had dropped low enough that a much wider population of musicians could get their hands on one.

Akai’s S612 and S900 (mid-1980s) set the template for the next decade: 12-bit then 16-bit quality, floppy disk storage, MIDI integration, a straightforward architecture that let you record a sound, tune it, and play it from a keyboard. The S900 was a studio workhorse for years.

Bit depth, sample rate, and the sound they made

Digital samplers run on two numbers: bit depth and sample rate. Both constrain the recording, and for decades, both were limiting enough to define the characteristic sound of an era.

Bit depth determines how precisely a sound wave’s amplitude is measured at each sample point. Each additional bit roughly doubles the resolution and adds about 6 dB of dynamic range. Eight-bit audio yields 256 possible amplitude levels and a theoretical dynamic range of around 48 dB, which is less than vinyl (around 70 dB) and far below the range of human hearing (around 120 dB). The result is an audible noise floor between sounds and a slight crunch on loud transients: the signature of early digital.

Sample rate determines the highest frequency a recording can represent. The Nyquist theorem requires at least twice the target frequency: a 26 kHz sample rate yields a frequency ceiling around 13 kHz, while 44.1 kHz raises it to about 22 kHz, past the threshold of most adult hearing.

The first generation of commercial samplers worked at 8-bit across the board, with sample rates ranging from roughly 16 kHz to 33 kHz. The noise floor was part of the sound. Sampled material took on a shimmer and compression that, in context, felt alive in a way that later “clean” digital did not.

The E-mu SP-1200 (1987) recorded at 12-bit and a deliberately low 26 kHz: a combination that compressed the high frequencies and gave sampled drums a dense, crunchy quality. Hip-hop producers including Pete Rock, DJ Premier, and Large Professor used this not as a limitation to work around but as a production tool. The worse technical spec produced the better sound, for what they were making. This is why old gear commands premiums: sometimes the floor is the feature.

The Akai S1000 (1988) changed the reference. At 16-bit and 44.1 kHz, it matched the CD specification that had launched five years earlier. Sixteen bits gives roughly 96 dB of dynamic range, which covers the full musical range with clean headroom. The S1000 sounded “transparent” to producers used to 12-bit texture. Most found it a relief. For the following 15 years, 16-bit/44.1 kHz was the professional floor.

Professional work shifted toward 24-bit through the 1990s and 2000s, mostly in recording and mastering contexts where the extra headroom reduced noise accumulated through long mix chains. Most hardware samplers stayed at 16-bit for playback, since the improvement above that threshold is inaudible on most speakers. The Make Noise Morphagene uses 32-bit floating-point audio: a format where there is no fixed amplitude ceiling, so signals can’t clip inside the module regardless of what’s happening in the patch. The cost is file size: a 32-bit float file is twice the size of a 16-bit file of the same length. At Morphagene’s maximum reel length of 174 seconds of stereo audio, that’s still under 80 MB, a number that would have seemed impossible for a sampler in 1985.

The sample length problem had a different arc. Early digital samplers were constrained not by format but by RAM: a few megabytes at most, at a time when memory cost hundreds of dollars per megabyte. The Ensoniq Mirage offered about a second of audio per key in most modes. Every second of audio required physical memory you had to buy, which is why early samplers have a fundamentally different vocabulary from modern hardware: sounds had to be building blocks, not recordings. By the S1000 era, several seconds per slot were routine. The Octatrack’s Static machines, which stream from CompactFlash rather than loading into RAM, circumvent the constraint entirely. The Morphagene’s 174-second reel limit is a design choice, not a hardware necessity.

Hip-hop and the cultural explosion

While electronic musicians were building patches and scoring films, hip-hop producers in New York were doing something different with the same tools.

The Amen break, a six-second drum fill from “Amen, Brother” by The Winstons (1969), became one of the most sampled pieces of audio in history. The same short loop, chopped and rearranged, runs through decades of hip-hop, jungle, drum and bass, and countless other genres. It’s direct evidence that a short recorded fragment, placed in the right hands, becomes raw material for an entire musical tradition.

The E-mu SP-1200 (1987), with its 12-bit sampling and distinctively gritty character, became the production tool of choice for Golden Age hip-hop. Its low sample rate wasn’t a limitation to work around; it was a sound to chase. Producers including Pete Rock, DJ Premier, and Large Professor used it specifically for the texture the bit depth gave a sampled break.

The Akai MPC60 (1988), designed by Roger Linn in collaboration with Akai, combined a sampler with a 4x4 grid of velocity-sensitive pads and a step sequencer in a single unit. The workflow it established — load a sample, chop it into slices, sequence slices across pads — is still the basis of most sample-based production today. The MPC line ran through multiple generations over 35 years, and its influence is visible in every grid-based hardware controller built since.

The legal shift

Sampling in the 1980s was largely informal. Producers sampled freely, clearances were rarely sought, and the legal framework hadn’t caught up to the practice.

The Grand Upright Music v. Warner Bros. Records decision in 1991 changed that. The case involved Biz Markie’s use of Gilbert O’Sullivan’s “Alone Again (Naturally)” without a license. Judge Kevin Duffy’s ruling was brief and pointed: sampling without permission was copyright infringement. The era of informal sampling was over.

The downstream effects reshaped the sound of hip-hop. Clearance costs made dense sample collages prohibitively expensive for commercial release. Producers shifted toward original instrumentation, live musicians, and interpolation (re-recording a melody rather than sampling the original recording). The Amen break and other cheaply licensed material kept circulating; everything else got expensive.

A later ruling, Bridgeport Music v. Dimension Films (Sixth Circuit, 2004), went further: any unlicensed use of a recorded sound, even a single note, constituted infringement in that circuit. “Get a license or do not sample” was the formulation that came out of it.

The legal framework didn’t kill sampling. It redirected it. Producers found workarounds, built sample-free workflows, and the underground kept sampling with less commercial exposure. But the era of freely borrowed recorded material in commercial music largely ended in the early 1990s.

Software and the sample library

The 1990s brought DAWs, and DAWs brought sampling into software. The sample pack emerged: collections of loops and one-shots organized by genre, BPM, and key, distributed first on CD-ROM then online.

Native Instruments Kontakt (2002) defined what a software sampler could be: a scripting environment for building instruments from sampled audio, capable of multi-gigabyte libraries with velocity layers, round-robin alternation, and articulation switching. Orchestral libraries, grand pianos, vintage synthesizers — previously requiring dedicated hardware or live players — became software products.

The file management problem scaled with the library size. A 2 TB drive in 2010 could hold more samples than an 1980s studio had ever touched. Finding the right sound in a library of 500,000 files requires different tools than finding it in a library of 500.

The hardware comeback

Something shifted in the mid-2010s. After two decades of pulling production toward laptops and DAWs, a significant portion of the electronic music world started pulling back toward hardware.

The reasons were partly practical: hardware forces decisions, breaks the computer’s infinite undo, and exists in space in a way that changes how you interact with it. Partly aesthetic: the warmth of 12-bit sampling, the mechanical constraints of tape, the physicality of knobs over pixels.

The Elektron Octatrack (2011) and Digitakt (2017) combined the MPC’s sampling workflow with deep sequencing and a hardware-first design philosophy. The Roland SP series, the MPC Live and MPC X, the Teenage Engineering OP-1: hardware samplers proliferated and diversified.

In Eurorack, the approach went further. The Make Noise Morphagene translated Schaeffer’s tape-splicing ideas into a voltage-controlled module: load a reel of audio, navigate it by splice, scatter it into granular texture. The connection to the tape studio isn’t metaphorical; it’s structural. The Intellijel Multigrain took the granular engine in a different direction, toward texture synthesis from banks of short samples. The Qu-Bit Nebulae, the Arbhar, the Lúbadh: the Eurorack ecosystem has produced more variety in sampling approaches than any single hardware era before it. The Morphagene prep guide covers what preparing audio for that tape-derived workflow actually looks like in practice.

The old problem remained. Every generation of hardware invented its own storage format and folder structure: FAT32 microSD for the Morphagene, FAT32 USB for the Squid Salmple, CompactFlash for the Octatrack, internal +Drive for the Digitakt. The job of getting the right file into the right format in the right place on the right card followed sampling from Schaeffer’s studio into the present. Tools like SampleStack exist specifically because of that persistence: converting format and writing folder structure across all of this hardware is the same job it has always been, just with more devices to target.