Sampling has a surprisingly long history with the first recognisable machine appearing as early as 1938 in the form of a 1936 invention by Frederick Sammis: the “Singing Keyboard”. A precursor of modern samplers, the instrument played electro-optical recordings of audio waves stored on strips of 35mm film which were triggered and pitched when the player pressed a key. Little is known of this machine and its uses but its underlying approach clearly prefigures the first practical sampling technology as employed in the American Chamberlin and its later British equivalent, the Mellotron.

These instruments were essentially tape players of extraordinary mechanical configuration. A capstan roller stretched the length of the machine and short strips of tape, one for each key, could be played back individually. Neither machine had the ability to record its own samples but a range of sounds was made available on prerecorded tapes which were factory-loaded into a supporting frame.  Most tapes held three separate audio tracks, allowing any one machine to create up to three different sounds. Typically, these would be strings, voices and flutes but a usefully wide range of instrumental and effects recordings was made available and the split keyboard of the later Mellotrons allowed the simultaneous use of two voices although over a more restricted range.

The notable characteristic of both the Chamberlin and the Mellotron was that they were capable of playing a sustained (ie not looped) sound although, in practice, the duration of any one note was limited to around 8 seconds after which time it would cut out automatically as the tape reached its end, thereby requiring a specific playing technique. The general view at the time was that the Mellotron (later renamed the Novatron for copyright reasons) was unwieldy and unreliable, being especially prone to tuning problems. However, the number of such machines still in use suggests that this may have been more a function of poor maintenance rather than shortcomings in design. The sounds produced by these machines, although derived from “real” instruments corresponded poorly to their original sources: as imitative instruments they were not therefore particularly successful but, as with many such instances, they acquired a charm and following of their own and have remained in use, albeit intermittently and at the whims of musical fashion ever since.

The machine that virtually created the practice of sampling as it is now known did so from necessity rather than out of any intention to invent a new approach to musical production was the Fairlight CMI (Computer Musical Instrument). Deriving its underlying concept from computer technology rather than musical instrument design and its name apparently from a ferry crossing Sydney harbour, the Fairlight became what it did since it was initially an unsuccessful synthesiser.

The designers, Peter Vogel and Kim Ryrie set out in the mid 1970s to create a wholly digital synthesiser. To do this they used two parallel microprocessors - Motorola 6800s (as used singly in the Atari computer) to create a pilot machine, the Quasar. This preproduction machine was intended more as a proof-of-concept than anything else and quickly demonstrated the limitations of digital synthesis as it was then: waveforms of more or less any description were easily created but harmonic partials were simply beyond its capabilities and such sounds as were created were unsatisfactory for instrumental purposes. This problem was not unique to Fairlight: their great rival, the American Synclavier suffered from similar limitations (vide the early works by Frank Zappa).

The solution adopted by Vogel and Ryrie was remarkably audacious for its time: they proposed to use “real” sounds in digitised form as their primary source. Conventional instruments were to be digitally recorded into the machine and then manipulated to produce the desired output. This fell far short of their original intention and was a solution that neither welcomed or approved of. Kim Ryrie (quoted in AudioMedia magazine) puts it thus:

“We had set out to develop realtime pure synthesis technology that allowed us to control every parameter of the sound. We wanted to digitally create sounds that were very similar to acoustic musical instruments and that had the same amount of control as a player of an acoustic instrument has over his or her instrument. Sampling gave us the complexity of sound that we had failed to create digitally but not the control we were looking for. We could only control things like the attack, sustain, vibrato and decay of a sample and this was a very, very severe limitation of the original goal that we had set ourselves. We regarded using recorded real-life sounds as a compromise - as cheating - and we didn’t feel particularly proud of it.”

The sound quality of the early Fairlight was poor by modern standards but, given the extent of its instrumental library (supplied on 8.5 inch floppy disks), it was immediately perceived (as the Mellotron had been some 10 years earlier) as a major threat by the Musicians Union which made strenuous attempts to ban its use. Following a lengthy promotional tour by Vogel, singer Peter Gabriel became interested by the Fairlight and set up a British company to import, promote and market it, thereby giving it immediate cachet amongst the rock cognoscenti. The great innovation that Gabriel immediately made was to begin to employ the ability of the Fairlight to record external sounds. This allowed him to use non-instrumental sounds as if they were instruments, with, for example, their pitch transposed via the keyboard. Thus it was that one of his subsequent albums carried such instrumental credits as “scraped exhaust pipe” and “tapped paving slab” and it was a tribute to the musicality of the Fairlight that their sounds integrated seamlessly with those of “conventional” instruments.

Not only could the Fairlight sample but it could place the samples in a sequence shown visually on the display terminal. Although crude by modern standards, the now-legendary “Page R” sequencer allowed composers without musical performance skills to work in a way not dissimilar to that employed by those who were conventionally trained. Not only this, but they were presented with a potentially infinite instrumental palette.  This capability was quickly seized upon by English producer Trevor Horn and his collaborators, later collectively known as “The Art of Noise” in an early effort to transform the pop recording process by the extensive use of sampling and sequencing. Notes or phrases, played, sung or “found” would be sampled, looped and sequenced to create entire songs often with minimal participation by the named performers

By now, successive models of the Fairlight had dramatically improved upon the originally poor sonic performance and created a viable sample-based instrument. One of its major limitations, shared like so much else with the NED Synclavier was that it was horrendously expensive and difficult to use. These factors necessarily limited its widespread acceptance. There began a perceived need for cheaper and simpler alternatives. One of these derived from American rap techniques and the related art of “scratching”. Here, the sound source would be a record or records, manipulated manually to allow controlled repetition of rhythms or phrases, sometimes to the accompaniment of a drum machine. Although limited in its transformative capability, this approach had the merits of being simple, direct and, above all, highly affordable.

A somewhat more sophisticated approach was adopted in many British studios where digital delay lines were used as basic samplers, sometimes triggered by incoming audio. Here established technology wass used in a manner never anticipated by its designers. Instead of recording and briefly storing a sound prior to its release in order to create a repeat or echo, the sound would be held in memory for a potentially indefinite time. This sound would then be played out only upon receipt of a suitable trigger signal, derived perhaps from an audio source previously recorded to multitrack tape. This allowed accurate replacement of, say, a kick drum part with a preferred sound without compromising the qualities of the original performance. Some such units featured short scale keyboards and had a usable, albeit limited instrumental capability. Fully featured samplers remained exotically expensive until the arrival in 1981 of the Emulator. Costing only $8000 as opposed to the $25000 dollars demanded for a Mark 1 Fairlight, the Emulator was a simple, user-friendly sampling keyboard that quickly gained enormous success, coming first to the attention of the British public with the release of Paul Hardcastle’s “Nineteen”.

Using musical and other samples from a TV documentary about the Vietnam war, together with more conventional techniques, Hardcastle created a style reminiscent of scratching but with far greater controllability.  In so doing, he presented the public with one of the first sampling stereotypes: the fast repeating stutter of “N,N,N,N Nineteen….”

Notwithstanding the relatively low audio quality of most samplers, they rapidly gained acceptance for their ability to render “real” instruments with passable accuracy. Not only this but also the perceived “sterility” of the sounds generated by the new generation of digitally based synthesisers such as the DX-7 resulted in a demand for a warmer, more natural sound. This resulted in a legendary partnership: the DX-7 synthesiser linked via MIDI to the Ensoniq Mirage sampler with one keyboard controlling both machines. The DX-7 contributed sonic purity and definition while the Mirage added “realism”.

With the advent of MIDI, it became possible to control multiple devices from one keyboard or other controller (such as drum pads, computer etc) and hence some units ceased to be fitted with their own controllers, becoming sound generating boxes that required external MIDI control in order to function. For reasons that are far from clear, keyboards generally contained synthesis systems whereas, with a few exceptions, samplers lost their keyboards and became rack-mounted boxes. By far and away the most commercially successful such units have been the Akai range whose usability has resulted in their becoming the de facto industry standard over a period of many years.

The widespread adoption of sampling, initially as a means of instrumental replication and subsequently as a high-tech version of the compositional techniques relying on “beats” and “breaks” as originally developed in the West Indian tradition of “toasting” and adopted by American hiphop artists has been substantially influenced by developments in the computer industry. The increasing cheapness of memory chips and other storage media has made possible the creation and archiving of larger sample file sizes with the result that audio quality can be improved without unacceptable storage problems. Current systems are typically capable of quality that compares favourably with that of the compact disc: the medium upon which most of their output finally comes to rest.