Introduced in 1994, the Korg X5 synthesizer was basically an update on the tone production engine of the world-beatingly successful Korg M1. From the M1's launch in 1988, and essentially through to the end of the 20th century culminating in the late '90s N Series, Korg used their AI² Sample & Synthesis system in synths and modules.
Unlike the M1, however, the X5 featured no onboard sequencing – it was purely a synth, and not a ‘workstation’. Whilst the X5 was hardly a major advancement in technology from Korg (the module version of this keyboard had already been on the market for a while in the shape of the 05R/W), it did benefit musicians in some important respects. Firstly, the sound-programming system was widely known, so a large number of digital synth players would not need to learn a new concept. And secondly, the R&D cost for this synthesis system had been more than recovered, so purchasers were not going to pay premiums on the retail price. Relatively speaking, the X5 was an inexpensive synth, and in its day, it did offer a lot for the money.
The X5’s raw waveforms comprised hundreds of selectable multisamples – everything from the predictable range of pianos and organs to exotic instruments and drum kit sounds. Up to two of these samples could be combined in a basic Program Mode preset, with a variety of synthesis parameters and effects allowing the player to warp the sounds to taste. The user could then combine up to eight of these basic presets into a much more powerful ‘Combi Mode’ preset. In Combi Mode, the X5 could build some phenomenally big sounds with a heck of a lot going on. Combi Mode also incorporated digital effects, but these overrode the effects programmed into the basic patches in Program Mode. Combi Mode additionally allowed the user to allocate different MIDI channels to each of the composite sounds. That, in a nutshell, was AI².
The huge drawback with all of this was that the original X5 was only 32-note polyphonic. If you were stacking two samples per Program, and eight Programs per Combi, then with each sample taking up a voice (2 x 8 = 16), your Combi would only be two-note polyphonic (32 / 16 = 2). The fewer samples you managed to use per preset, the greater the amount of polyphony you retained. Of course, it was quite possible to create good patches using just a single sample, and that would allow the full 32-note polyphony.