Sampling — as a technique bearing this name and loosely following procedures we recognise today — has been around for little over 10 years. Surely no-one who witnessed the appearance of the first dedicated sampling instrument, Emu's Emulator, ever imagined it would have such staying power. Even Dave Rossum, founding father of Emu and inventor of the Emulator, laughed when I asked him if he regarded sampling as the future of the industry back in the early 1980s.

Yet has Emu, or anyone else of significance, produced a keyboard instrument that wasn't in some way based upon samples in the past five years? And if, in your wildest dreams, you'd had to choose between the E-II and something like the eponymous 360 Systems' sample replay keyboard as the significant, seminal product in those early days, which would it have been?

The user-sampling E-II may have gathered the immediate sales (even at slightly over £10,000!), but it's the instruments whose sounds are simply based upon samples — be they sample + synthesis designs like the D50 and M1, or relatively 'straight' sample replay units like the Roland U-series and Emu's Proteus families that have generated the really big bucks.

Sampling, in one form or another as a hidden technology on all 'keyboard' instruments, from synthesizers to pianos to drum modules, is now firmly established as the leading source of sound production. Sampling as a (theoretically) user-undertaken activity, although it remains an important part of the keyboard player's life, singularly failed to kill off the synthesizer as many feared it would once the concept caught on in 1985. The reason for this is that operating a sampler is not, nor most likely will ever be, as easy a operating a synth.

Although sampling first came to the industry's attention as a function on the Fairlight CMI, and subsequently as the stand-alone Emulator, the big breakthrough for musicians came in 1985 with the simultaneous appearance of Ensoniq and their Mirage, the first sampling keyboard to break the magic £1,000 barrier, and so bring sampling within reach of the masses. Although the Mirage proved spectacularly successful, the low cost sampling market which at the time it seemed to create proved, in fact, somewhat of a mirage itself.

The interest generated by the Mirage simply whetted the appetite. Manufacturers quickly found that the limitations of 2.8" Quick Disks, of memory, screen size, and to an extent basic sound quality in this price range, combined to make the whole business of sampling only viable for the musician at more sophisticated, and so more expensive, levels. Accordingly entry level instruments from other companies met with limited success. Thereafter sampling has always been viewed, if not quite as the rich man's domain, certainly as the sport of those serious enough to spend twice as much to indulge this particular interest as they would otherwise do on a synth.

This early confusion of format, systems, market and applications pulled sharply into focus in 1986 for the British market when Akai launched the S900. The S900 was the right instrument at the right time, offering the right facilities at the right price. At a stroke, the 3.5" disk, large-screen format, rack mount sampler was de rigeur.

Unfortunately for Casio, Roland, and Korg, and disastrously for Yamaha, Akai had not only got there first in terms of hardware, but were also quick to encourage the development of a large sound library. The net result was that everyone else had slim pickings.

Akai did not enjoy, and have not enjoyed, quite the same level of success in America though, and this is for two reasons: Akai's presence as an MI company in America was at nowhere near the level that it is in Europe. More importantly, in America the two domestic companies of Emu and Ensoniq were able to fend off opposition to any of their products on their home turf. America, after all, is not noted for reticence when it comes to looking after her own. Many a British musician has been taken aback in the States when they bowl into the studio with their Akai disks to be greeted not by the customary S1000 but by an E-III or an Emax Rack.

It's fair to say that Akai has continued its dominance of the European sampler market through the S1000 series (though the keyboard version has been something of a turkey), and S950. But the sailing is not quite as plain as it used to be.

Roland, who made a creditable stab at things with the S50/550 came out with all guns blazing on the top of the range S770/750. Emu's stature in Europe continues to grow, and if the Emax series has gone off the boil somewhat the forthcoming E-III rack looks powerful enough to take on all challengers. Meanwhile Ensoniq, who for years were disastrously mishandled outside America, now look secure enough (as do the products) for people in the market for a keyboard sampler — Ensoniq's strength — to buy the products with peace of mind.

Although the safe route might still appear to be with Akai, today, with the opening up of the sounds market thanks not only to MIDI sample dump standard but also the growth of dedicated sound companies, and with the simple passage of time whereby everyone (practically) seems to have got copies of everything in every format, the choice is by no means straightforward.

Some interesting developments are on the horizon too. With sampling newcomers Peavey's concept of the SX sampling expander, a unit that simply deals with the process of turning a sound into a sample (storage, editing, and playback are undertaken elsewhere), and Kurzweil's K2000 (listed as a synth in last month' table) and its ability to load in samples as sound sources, it is quite possible that the whole process of sampling will soon be turned on its head.

Sampling has always had two main applications: as a source of new sounds to be played instrumentally, and as a digital recording medium for loops, post production or, as is currently happening more and more, for complete recording as part of a direct-to-disk system. To date there has been considerable cross-over at the top end of the market. But it seems likely that the future for sampling as an instrumental technology may lie in add-on units for, or features within, synthesizers or workstations, leaving the dedicated sampler free to concentrate on its various recording applications.

The reason for this is not only that the marriage of instrument and recorder has always been a somewhat unhappy one, but that since very few people spend much time actively sampling, preferring to acquire ready-made samples from third parties, what on earth is the point of offering complex user sampling facilities on each unit? However, the perverse nature of the music world being what it is, who wants to be seen with a playback only instrument? No-one, of course. So offering sample input via separate add-on modules, or as features on a larger instrument, could prove to be a most happy, and face-saving solution. This may be the future. What are the chief concerns of the present?

THANKS FOR THE MEMORY

Internal memory is one of the most important aspects of sampling life since it dictates how many sounds you can use at any one time, how long they are (or how much you have to trade quality off against sample length). In some cases memory size also affects polyphony.

Few samplers now come with less than 2MB (megabytes) of internal memory, which is generally quite sufficient for a decent piano sample, and some strings or percussion. But for any major feats of looping, live vocal triggering, or of post-production arrangement re-hashing, 8MB would seem a more sensible amount. (Sample memory size is also sometimes listed in mega-words, or occasionally megabits. Here's the difference: a byte is eight bits, so one megabyte equals eight megabits; a 'word' is 8 bits long in an 8-bit sampler, and 16 bits long in a 16-bit sampler. So in a 16-bit sampler, one megabyte equals half a megaword.) Memory is seldom cheap from instrument manufacturers themselves, but fortunately third party memory expansions are becoming more widely available, and of course for instruments that can accept low cost SIMMs chips the financial outlay can be kept at very respectable levels.

STORAGE

Internal memory is one thing. External memory is another hot ticket. The question is not so much "should I bother with a hard disk or some such storage media", but "which one or what type?"

For a while instruments with built-in hard disks seemed very attractive. The problem is, once you've filled up your hard disk you're back to square one: floppies. Although removable media units from the likes of Syquest look more viable, because you can store data in 45MB chunks on cartridge, the price and availability of removable optical disk drives has reached a stage where they might provide a better solution for some in terms of price per MB of storage capacity. There are still compatibility problems with some samplers, however, so make sure that any external drive you are looking at will actually work with your sampler.

It is important to keep in mind that any such device will not compensate for lack of internal memory. Yes, it is often possible to initiate sample loading from hard disk with MIDI program changes, but this can be a dangerous system to rely upon in a live situation.

SCSI (Small Computer System Interface) is the connection used to link a sampler to hard disk (or to another sampler for sample dumps). On most high end samplers SCSI is standard; invariably it is an option on others.

"Surely no-one who witnessed the appearance of the first dedicated sampling instrument, Emu's Emulator, ever imagined that sampling would have such staying power."

CD-ROM, a medium that puts ROM sample data on CD, was popularised by Emu way back in the days of the E-II. Never the most fashionable of systems purely for sound storage (CD ROMs are also used for dictionaries, encyclopaedias etc.), it is nonetheless extremely fast and effective when it comes to auditioning sounds; handy if all you want is a large library of third party sounds to choose from Although the drives can be universal, the CDs themselves are of course dedicated to specific products or ranges of products.

SOUNDS

The public's seemingly insatiable appetite for buying new sounds — for synths and samplers — is testament to the level of our collective interest in and ability at programming or sampling for ourselves. The point is fast approaching where exclusivity is a thing of the past. Certain samplers reproduce certain types of sound more faithfully, or more sympathetically than others, but you'd be hard pressed not to find Fairlight Strings, Marcato Strings (originally an Emu warhorse), orchestra hits, Peter Gabriel's shakuhachi, and all the other sampling cliches freely available no matter which instrument you choose.

To an extent, samplers have just become long-winded preset synths; the fact that their sounds happen to be samples is of little importance. However absurd the situation, for the time being the market for new samples remains buoyant. As to which medium is best for acquiring the samples, the current hot ticket is still CD — regular audio CD — whereby you can load in (ie. sample) actual sounds or indeed loops off CD, mapping and/or tailoring them to suit your own application.

Undeniably, the sampling CD solves one of sampling's past problems: that of being able to find suitable raw material (how often are you going to be in a position to record a choir, or a string orchestra?). The other advantage is that the sampling CD is not tied to any one instrument because what you are dealing in is audio, rather than data.

The downside of the method is that you are still left with the business of organising your samples — looping, mapping, tuning etc., which can be laborious and skillful procedures. The ready-to-go disk or CD-ROM thus still has some advantages.

EDITING

Sample editing facilities — the means by which you manipulate the sound you've just sampled come under two headings: digital and analogue. Digital editing facilities enable you to play with the sample data by looping and truncating etc., whereas analogue editing lets you process the sound with filters, envelope generators and LFOs to shape your final product.

In spite of relatively large screen formats being the norm, few would dispute that sample editing programs such as Digidesign's Sound Designer II for the Mac or Steinberg's Avalon for the Atari will not greatly increase your ability to tailor samples quickly, and exactly as you want. In this respect Roland are at somewhat of an advantage since their samplers offer a CRT connection as standard — you can just plug in a computer monitor and go, using a Roland mouse to find your way around.

'Analogue' processing can be vital. Without a filter it will be almost impossible to change the basic feel and timbre of your samples. The better the filtering, the more control you will have in this respect. Powerful envelope generators also help enormously, not only with obvious problems like softening the attack of a sound, but also when it comes to having to make vocals sound less synthetic, adding realism to string samples, or even turning brass samples into strings. Generally, analogue-style editing features help out in the creative department.

YOU MUST REMEMBER THIS...

If you're in the market for a sampler, consider the following points in relation to any machine you look at:

Internal memory: what is standard, what is maximum, how much are upgrades?

SCSI: Is it standard; if not, how much does it cost to add a SCSI interface?

Polyphony: vital to check. 16-voice can still be quite limiting.

Extraneous features: do you need that sequencer/SMPTE generator/those digital effects/synth wave generating?

The public's seemingly insatiable appetite for buying new sounds — for synths and samplers — is testament to the level of our collective interest in and ability at programming or sampling for ourselves.

(Click image for higher resolution version)