The History Of Korg: Part 3

Cutting-edge innovation is by no means a guarantee of commercial success, but Korg have had a flair for both — although not always simultaneously. We look at how this continued to be true throughout the '90s to the present, when the company pushed the boundaries of physical modelling while refining their world-beating workstations.

In Part 2, we saw Korg trading at the peak of their fortunes until, in 1994, their former majority shareholder appeared to steal the limelight. Yamaha's VL1 and VL7 keyboard synths embraced physical modelling, and generated orchestral sounds of remarkable realism and expression. Korg had developed their own physical-modelling instrument but, unlike Yamaha, who had addressed mainstream synthesis, Korg had launched a commercially unsuccessful percussion synth based on a shadowy pair of developments called SynthKit and OASYS. So let's continue the story from there...

1995

Although rumours of OASYS had been circulating since 1992, none but a select few had seen it, and some doubted its existence. Nonetheless, it was real, although it looked unlike any previous keyboard, with the centre of its control panel dominated by a touch-sensitive screen that provided access to, and control over, its complex sound-generating system — or rather systems. Long before the advent of software synths, Korg R&D had designed OASYS to be capable of loading and running whatever algorithms the company developed for it. This meant that, unlike the fixed models in Yamaha's VL1 and VL7, the OASYS could
— in principle — support an unlimited range of sound-generation systems. The only limit would be the power required by the models... there was an upper limit to the complexity that OASYS could support.

The original OASYS keyboard, as previewed in SOS March 1995. Sadly, it was never fully released commercially in this form, but nevertheless, it did exist. Some of its technology subsequently appeared in the Prophecy, Trinity, Z1, and Triton.Korg R&D had used the software SynthKit to develop a wide range of models, but these were to be whittled down to a selection that included imitations of orchestral instruments, tonewheel organs, FM synthesis, stereo PCM synthesis, wave sequencing, additive synthesis, and various forms of analogue synthesis. The company even claimed that OASYS would support several models simultaneously, allowing players to mix orchestral models with analogue-synthesis models with drum samples, and so on. But OASYS suffered from two severe limitations: it was overpriced and underpowered. Korg's market research revealed that the projected end-user price of £10,000 was far too high for a 16-voice polysynth.

Nonetheless, in July 1995, a non-functioning prototype played a supporting role at a press and dealer event in the hallowed surroundings of the Science Museum in Kensington, London. The occasion was the launch of Korg's two most important products of the '90s, both of which leaned heavily upon OASYS. They were the Prophecy and the Trinity.

Costing just under a thousand pounds in the UK, the Prophecy lacked the expensive OASYS user interface, offered fewer algorithms, lost the ability to load new models, and was monophonic. However, thanks to velocity- and aftertouch-sensitivity and its unique 'log' (a combination of modulation wheel and a pressure-sensitive ribbon controller) the Prophecy's orchestral models sounded realistic and expressive without the use of a breath controller. What's more, its wealth of analogue models placed it in a class of its own. With multiple multi-mode filters, multiple multi-stage envelopes, powerful LFOs, waveshapers, modulation matrices, user-defined key priorities, an arpeggiator, effects, and programmable performance controls, it was fully capable of emulating the sounds of '70s analogue synths, yet was equally suited to modern musical styles.

In 1995, the depth and range of the Prophecy's modelling was a revelation, and it has never fully been plumbed. Dance and techno were by far the largest markets for the little synth, and most people in those genres were interested in nothing but the analogue models. What's more, the Prophecy was not destined to be a worldwide success. It was popular in the USA and in the UK, where, according Korg UK's Managing Director, Rob Castle, "customers seemed to understand the Prophecy", but elsewhere it failed to take off.

Launched the same day, the Trinity (see box, above) was even more significant, because it reaffirmed and reinforced Korg's position as the leading manufacturer of keyboard workstations. Its new ACCESS sound generator was unparalleled: it contained 48MB of PCM waveforms, and there were so many multisamples that the Program memory was insufficient to take advantage of all of them. ACCESS was also the first of Korg's workstation engines to incorporate resonant filters, and was therefore a huge leap forward.

The Trinity retained the touch-sensitive screen first seen on the OASYS. For the first time, a player could touch a parameter or icon and edit it by moving a finger across the screen itself. And then there were the keyboards... expensive Yamaha mechanisms that made the various models of Trinity far more responsive than Fatar-equipped models such as Korg's earlier X and N series. Nevertheless, for many people, the most important innovation was the Trinity's multitimbral effects section. This allowed you to allocate a set of so-called 'Insert' effects to one sound, and different effects to a second, or a third, or a fourth... up to the limits of the power available. The Trinity was, therefore, the first truly multitimbral synth.

The remainder of Korg's launches in 1995 rather paled in comparison. This was a shame, because the 64-voice X5D polysynth and its rackmount sibling, the X5DR, were the culmination of seven years of AI development. With a new 8MB ROM including 90 additional multisamples and 50 new percussion samples, they incorporated the largest and most powerful AI² engine yet heard.

Then there were three new interactive keyboards, plus the ih, a vocal harmoniser developed by IVL Laboratories, the company responsible for the Digitech range of Vocalist harmony processors. Designed for use alongside the 'i'-series keyboards, the ih was hampered by its single XLR input, which made it unuseable with unbalanced microphones, and unable to process instruments. This was a mistake which limited sales considerably, and it was an avoidable one (although it's possible that it came about because IVL's contract with Digitech precluded them from using a full set of facilities in any other manufacturer's products).

1996

AI² had proved to be Korg's cash-cow for half a decade, and the company didn't intend this to stop, unveiling the N364 and N264 workstations in 1996. Visually, these were almost identical to the X3 and X2. Internally, they used the 8MB, 64-voice polyphonic AI² engine from the X5D and X5DR. But though a re-hash of existing technology, they also offered some significant improvements over their predecessors, including static RAM that allowed the 'N's to retain sequences when switched off, and an arpeggiator. However, the big innovation was the inclusion of 'RPPR', or Real-time Pattern Play and Record. Unlike conventional sequences or interactive accompaniments, Patterns were riffs and phrases (your own, factory-written, or ripped from MIDI sequences) that you could assign to keys and replay in real time or into the sequencer to build tracks.

As had happened in previous years, 1996 witnessed many significant developments outside Korg's traditional core markets... Take the Pandora multi-effects processor as an example. Aimed at guitarists, this offered 60 effects, allowed you to use up to four simultaneously, was small, simple to use, light, cheap, and would even run on batteries. Sure, it had limitations, but it was no surprise to find the things appearing in the gig bags of guitarists everywhere. Keyboard players adopted them, too, using them to create guitar-type patches for lead synths.

The 1212 PCI audio I/O card, when marketed as a stand-alone product, was a great cross-platform success.

If Pandora demonstrated that Korg could compete at the low-cost end of the spectrum, the launch of the Soundlink DRS (Digital Recording System) demonstrated that it also wished to compete at the top, although only a few features linked the DRS to the original 1992 Soundlink system besides the name they shared in common. There were six units in the DRS range: a reference amplifier and speakers (which we can ignore), a pair of eight-channel A/D and D/A converters and, at the core of the system, the 168RC digital mixer and 1212 PCI audio I/O card. At £2500, the 168RC was not cheap, but with 16 simultaneous digital (ADAT) inputs and outputs, plus eight analogue inputs, its spec exceeded that of any other digital mixer in its class. It also retained the automation of its 1992 progenitor, and replaced the limited effects processors of the original Soundlink with new, dedicated EQs, plus two Trinity effects processors. Why the 168RC sold in such limited numbers is a mystery.

If the 168RC was less than a raging success (which it was) the same could not be said of the 1212 card. With 12 inputs and outputs, high-quality converters (for the time, anyway) and word clock, this was more than a match for the competition. What's more, it was not particularly expensive, and was compatible with both PC and Mac, which meant that players could use it with most MIDI + Audio sequencers. Given its price and broad application, the 1212 could be sold through almost any music shop, so it reached a far wider market and sold in far greater numbers than any other Soundlink product.

Nonetheless, the DRS system was not a success, perhaps because — even with all six units — you still needed a PC or Mac to provide recording and sequencing tools. Furthermore, the hi-tech recording market was changing rapidly, with the prices for features hitherto considered solely the preserve of high-end professional studios, such as automation, moving faders and digital recording, tumbling fast. But Korg were undeterred, as future releases were to demonstrate.

1997

The first new Korg to appear in 1997 was another AI² module, the NS5R, which forsook the company's conservative, black look, substituting a two-tone grey designed to sit alongside a PC. However, there was another cosmetic improvement that was really worthwhile; its large orange display facilitated dramatic improvements to the user interface. The NS5R had limited outputs, but with the 64-voice polyphony of recent AI² products, and 32-part multitimbrality, it was no toy.

Another interactive keyboard appeared in 1997, and this one deserves mention because, rather than sport a domestic or 'night-down-the-social' appearance, it was built into the N364 case and had a striking, professional look about it. Indeed, the iX300 was in some ways the most powerful AI² synth to date, although the number of Patch memories was very limited, and there were no Combis at all. But this was not the first time that Korg had (ab)used its AI² engine in this way. The iX300 was virtually identical in all but looks (and speakers) to the i5S... updated and redesigned to appeal to the more style-conscious musician, but otherwise the same instrument.

Probably the most powerful physical modelling keyboard instrument ever commercially released, the Z1, was, astoundingly, a relative commercial failure. It remains one of a kind to this day.

Two more of 1997's major launches were updates of existing products. The SGproX was Korg's latest 88-note stage piano, and offered a 15MB multisampled grand piano among its 64 patches. At the other end of the spectrum, a revised Pandora — the PX2 — extended the original design with 32 preset drum programs that allowed you to practice or jam along with nothing but a guitar, a PX2, and a set of headphones. However, the year's big news was the Z1.

When it appeared, the Z1 caused a stir, and rightly so — even now, it remains the pre-eminent instrument of its class. In essence, it was a 12-voice polyphonic, six-part multitimbral version of the Prophecy, with additional SynthKit models and greater programming flexibility than ever before. Housed in a standard Trinity case, it lost the touch-sensitive screen, and reverted to a conventional LCD, but gained an X-Y touch-sensitive pad that acted in many ways like a joystick, ribbon controller, and the Prophecy's 'log' all rolled into one (the pad, of course, was to prove significant as a product in its own right a couple of years later).

The additional SynthKit models included Hammond organ and electric piano, and for some players, the pianos alone justified the Z1's existence. Unfortunately, just as they had failed to do with the Wavestation, Korg never released a 76-note Z1 Pro or 88-note ProX. What's more, there was no Z1 module. At least there was a six-voice expansion board, so users could expand the polyphony to 18 voices. This was a boon for piano sounds, as well as for using the Z1 multitimbrally.

With its fantastically powerful MOSS engine, a fistful of control knobs, the X-Y pad, a polyphonic arpeggiator, a bundled software editor, and a return to the Trinity's high-quality keyboard, the Z1 should have become everybody's synth of choice. But it didn't. In part, this might have been the result of its limited polyphony, or its use of AI² type effects rather than the multitimbral ACCESS effects structure. But there was a more arcane reason...

Soon after releasing the Z1, Korg's support staff apparently began to receive complaints along these lines: "I love the Z1's analogue models, but why did you stick trumpets and violins in it? I don't want them." Amazingly, Korg had fallen foul of the '90s craze for all things analogue, and unreasoning hatred of all things digital. By adding brass, strings and FM models, the company had deterred buyers who would have been happier had it been a far less powerful synthesizer.

Consequently, the Z1 was not the worldwide success that it should have been, and its patchy sales were not sufficient to justify further development. Soon after launching the Z1, Korg discontinued development of its physical modelling of acoustic instruments, and although analogue modelling was to continue, the Z1 was the last Korg of its kind.

1998

In 1998, Korg seemed to draw breath, and the next 12 months were devoted to products that exploited existing technologies.

At the high end, the Trinity V3 was an expanded workstation that incorporated six Z1 voices. This raised the maximum polyphony to 38 in total, and made it possible to pass the Z1's modelled sounds through the Trinity's superior effects processors. The TR-Rack was another step forward. At first sight, this was a Trinity in a 1U case, but it too offered several improvements. In particular, the patch memory was doubled, the Trinity's 24MB ROM was expanded to 32MB (including Korg's new stereo sampled piano), and it was supplied with an OEM version of Emagic's Sound Diver editor.

The i30 was a serious home keyboard, featuring an 18MB AI2 sample-based sound engine, a touchscreen and an optional hard drive.

As for exploiting AI² still further, there were the N1 and N5 keyboard synths, plus the N1R rackmount. With a 12MB ROM in the N5, and 18MB in the N1 and N1R, these offered 64-voice polyphony, 32-part multitimbrality, an arpeggiator, and GM compatibility, but saved costs by having just two outputs and no expansion capabilities. The SG-Rack was another rehash of AI² technology, but with much of its ROM dedicated to a single concert grand piano multisample. Likewise, the 18MB AI² engine found its way into three more interactive keyboards, the i30, iS40 and iS50. Of these, the i30 was the most interesting because, as the flagship of the range, it combined the facilities of the 'i'-series with the touch-sensitive screen of the Trinity, plus an optional 1GB hard disk drive for storing sounds, sequences and accompaniments. The 'home' keyboard had truly come of age.

For Korg enthusiasts looking for something new, there were the DL8000R and AM8000R ToneWorks rackmount effects processors. New... except that both harkened back to earlier products. The DL (delay line) recalled the days when effects units performed one task well, providing exactly what it claimed: a box full of delays, early reflections and echoes. And, just like a vintage delay line, it offered an internal LFO that also allowed you to create phasey, flangey and chorused effects. The AM (ambient multi-effects) processor was more flexible, and not a bit vintage, offering dual signal paths and a wider range of effects, many of which appeared to be drawn from the Trinity. With their bold designs and huge LCD characters, the 8000s were at home on stage and in the studio and, although they were not huge sellers in their time, they may well become minor classics. But, as in so many previous years, Korg still had a certified ace up their corporate sleeve during 1998. This time, though, it wasn't a new method of synthesis... it was the diminutive D8 hard disk recorder.

If Korg had misfired with their earlier digital recorders, its research and development was rewarded in 1998. Unlike Soundlink, the D8 was small, light, and affordable. Unlike Soundlink DRS, it was self-contained, and needed no external equipment to record and mix tracks (although wisely, it could still sync — and sync to — a MIDI sequencer). The D8 was not the first affordable digital recorder — Roland and Fostex had already released their versions over two years previously — but, given its size and layout, its clean user interface, and its well-thought-out range of facilities, the D8 still did much to redefine the concept of the portable hardware multitracker. Korg sold bucketfuls of them, and the public perception of the low-cost multitracker was never the same again.

1999

By 1999, Korg were launching so many significant products every year that it's impossible to give them more than passing mention here. There were more AI² keyboards and modules (the N5EX and NX5R) with 96-voice polyphony and 48-part multitimbrality (!), another AI² stage piano (the SP100), three more 'i'-series keyboards and modules, four more domestic 'Concert' pianos, and an improved multitrack recorder (the D16). There was also a replacement for the Trinity...

At the heart of the Triton lay Korg's new HI ('hyper integrated') synthesis engine. This name, while full of hype (no pun intended) was not mischosen. Just as Korg had integrated the M1's three AI chips into the single AI² processor, the Trinity's ACCESS chips were integrated into a single HI processor.

It was easy to see why so many players viewed the Triton as the sequel to the Trinity. It retained the sound structure of its predecessors, and offered a similar multitimbral effects structure. On the other hand, 62-voice polyphony, expansion slots, powerful arpeggiators, and a multitimbral MOSS option (ie. the synthesis type found in the Z1) showed that the Triton was no mere cosmetic upgrade. For example, each 12-note polyphonic, 48-step arpeggio was almost an form of auto-accompaniment in its own right, and you could use the arpeggiator to create wave-sequencing effects reminiscent of the Wavestation. In addition, the sequencer incorporated preset and user patterns reminiscent of the N- and i-series' RPPR functions.

Despite the Triton's superiority in some departments, many players continued to prefer the Trinity, and with good reason... there were some things that the older model still did better. Furthermore, the Triton lost the Trinities' digital I/O and hard disk recorder options. But if there was one area in which the Triton clearly exceeded any Korg instrument from the previous dozen years, it was sampling. Sure, an expanded Trinity had allowed you to import and use samples, but the Triton was Korg's first sampler since the demise of the DSS1 and DSM1. Although the sampling implementation was not quite to the standard of a dedicated rackmount, the Triton treated its samples as Programs, which made it extremely powerful.

1999 was also the year that Korg finally released OASYS, although not in the form originally envisaged. The OASYS keyboard had been dumped in 1996, so the technology returned to its roots when Korg reinvented it as a PCI card. This offered the I/O capabilities of the 1212 (although with 24-bit converters), an automated 12-channel mixer, and top-of-the-range synthesis and effects, all powered by five DSPs that you could program using a software editor derived from SynthKit.

Whereas the Prophecy had offered nine physical models, and the Z1 incorporated 13, the OASYS PCI offered 28, with new plucked string models, slapped bass models, a percussion synth model, additional brass and wind instruments, plus — for the first time — human voice. Like the Z1, OASYS could combine two models in a Program, and multiple Programs in a Combi. Unlike the Z1, OASYS supported multitimbral effects, and with far greater flexibility than either the Trinity or the Triton, making it capable of incredibly complex timbres, although with limited polyphony.

A few months after launching OASYS PCI, Korg released a Windows version of the Editor, and then the SynthKit development platform itself, thus allowing end users to develop new physical models. The potential for generating and warping sounds was fantastic, but Korg discovered that too many musicians wanted all the hard work done for them. This meant that OASYS never lost its (undeserved) reputation for being too complex, and its price plummeted. Just two years after release, the last units were sold off cheaply in the USA. Korg had offered the world almost everything it had asked for, and the world had said 'no thanks'. Madness!

So what did the world want? The answer lay at the opposite end of the spectrum. For years, Korg had been offering powerful and flexible products to the most discerning hi-tech musicians in the world. But the serious money now lay in 'groove' music that required a specific palette of sounds produced by instruments centred on controls that users could grab and tweak in real time. So Korg, ever aware of the fickle tides of fashion, developed a range of products to suit. They were the Electribes.

Small enough to slip in a gig bag, the EA1 was a two-voice physically modelled 'analogue' synth, while its near-identical partner the ER1 used similar technology to provide drum and percussion sounds. Each incorporated a handful of Korg's trademark effects, each offered an audio input that allowed users to add effects to external signals, and each incorporated a simple, pattern-based sequencer. Aimed at users producing dance and ambient music, they offered the right combination of facilities at the right price. The Electribes were instant successes.

The same could also be said of the diminutive KP1 Kaoss Pad. An hybrid of an effects unit and a primitive sampler, it was also a controller built around the X-Y pad first seen on the Z1. However, its intended target was quite different from that of its sophisticated forebear. With a microphone input and a phono input for accepting signals direct from a turntable, it was targetted at DJs and remixers. You could also use the Kaoss Pad as a MIDI controller, sending the continuous controllers of your choice by running your finger over the pad, and/or by pressing its buttons. The KP1 was far from perfect, but it was affordable, well built, and fun.

2000

The following year saw another handful of updates and reworkings of existing technology. Of these, the most notable was the Triton Rack, which lost the Triton's touch-sensitive screen and two voices from its 62-voice polyphony, but added digital I/O, an increase from two to eight expansion slots, increased memory, more arpeggios, more sample RAM, and the new mLAN interface option. There were also upgrades to the sampling software, including crossfade looping, time-stretching and sample-slicing, all of which subsequently became available on the Triton keyboards when Korg released its v2 operating system.

Elsewhere, HI synthesis also turned up in a most unlikely place... the inside of the PA80 Professional Arranger. A hybrid of a Triton and an 'i'-series interactive keyboard, this offered expansion options such as flash ROM, a 2GB hard drive, a harmoniser, a guitar input with dedicated effects, and even a video interface that allowed you to use it as a karaoke machine. Nothing in the PA80 was radical — each technology having been used in the Triton, the i3, the ih, and in Toneworks pedals — but it had never been combined in this way before.

More continuity was offered by new AX-series multi-effects pedals, the Pandora PX3 and the PX3B, all of which were released under the Toneworks name. Then there was the ES1 Electribe 'S', a desktop sampler to join the EA1 and ER1. Quite different from a 'pro' sampler, the ES1 was targeted at live performers and DJs for whom features such as audio slicing and motion sequencing, pre-loaded factory samples and rhythm patterns, and coarse, aggressive effects were ideal.

Then there was the new, DSP-powered CX3 organ. This was (and perhaps remains) the most authentic imitation of the Hammond A100/B3/C3 and Leslie 122/147 combination yet developed. It offered a special keyboard that mimicked the Hammond response, and its sound was almost indistinguishable from that of the originals. The CX3 also offered an 'EX' mode that allowed you to add additional drawbars and percussion to the standard Hammond configuration. Admittedly, it was an expensive and specialised product but, like the Wavedrum, the CX3 was not designed to sell in tens of thousands. It was another example of Korg doing the job right for the sake of doing the job right.

The MS2000 provided more proof that physical modelling could be commercially successful, particularly if focused on electronica/dance-friendly analogue synth modelling.

Mention of Korg's 2000 product lineup cannot be made without the company's most significant products of the year; the MS2000 virtual analogue synth/vocoder and its rackmount sibling, the MS2000R, both of which recalled Korg's classic product line-up of the MS20, MS50, VC10 Vocoder and SQ10 Sequencer. But the MS2000 offered significant synthesis enhancements such as DWGS and PCM waveforms, additional filter modes, velocity sensitivity, a simple modulation matrix, EQ, dual effects sections, an arpeggiator, and a powerful MIDI specification. Likewise, its vocoder added formant shifting and envelope shaping to the traditional facilities. However, these seemed to add to, rather than detract from, the flavour of the originals, and the MS2000s carried much of the MS-series' ethos into the 21st century.

Inevitably, the MS2000's factory sounds leaned heavily on techno and dance, but the instrument was capable of much more. In particular, it produced classic 24dB-per-octave analogue sounds not available from the majority of vintage Korgs, and its digital waves generated sophisticated sounds reminiscent of FM and additive synthesis. In truth, the MS2000s were sonic chameleons that could sound as vintage or as modern as you wished. Simple, affordable, and flexible, they made a very favourable impression, and, by the time you read this, the MS2000 synthesis engine will be teetering on the brink of release in more Korg products such as the new Microsynth (which, all being well, should be reviewed in next month's SOS).

2001 & 2002

So now we reach the 21st century, with Korg ensconced at the top of its industry. Nonetheless, products continue to pour forth... far too many to do justice to here. Effects units, concert pianos, low-cost digital recorders, Electribes, Pandoras, Kaoss pads and mixers, stage pianos, a dual-manual organ, new versions of the Triton, another Professional Arranger, tuners, and an encyclopaedia of expansion cards and options for every range... To be honest, it would take a whole issue of Sound On Sound just to describe the company's output in the past 12 months. In fact, it would take more, because, at the time of writing, more products are scheduled for release before the end of 2002.

Nonetheless, one recent synth deserves special mention. This is the Korg KARMA, the host for Kay's Algorithmic Real-time Music Architecture.

If it were not for KARMA's algorithms, there would be little to discuss about it, as it's fundamentally a downgraded member of the Korg Triton family. But the algorithms are what make KARMA special... Far more than an arpeggiator, and much less rigid than an auto-accompaniment system, it allows players to generate performances styled upon (for example) a guitarist's natural pick and fingering, or a harpist's strum. This may sound like a sophisticated arpeggiator, but in many ways KARMA is exactly what an arpeggiator is not, because you don't have precise control over the output, merely over the nature of the output.

With 1190 editable algorithms, the KARMA is capable of remarkable complexity and realism. Inevitably, it comes full of grooves, but there are patches that imitate orchestras, or blues, or jazz, or progressive rock, and even some that create movement within a sound, much like wave sequencing. If the KARMA has a shortcoming, it's the loss of the Triton's touch-sensitive screen. Korg got their keyboard user interface right in 1995, and it's been going backwards ever since. The Trinity's proud, angled screen and GUI was a masterpiece. The Triton's screen was less pleasant, because it sat flush with the top panel, making it harder to use unless you had clear space above the synth. The KARMA merely offers an LCD and a "Jump To..." menu system. Nonetheless, the KARMA remains a unique synth, and one that has yet to achieve its full potential.

And Now...?

Over the past three months, we've traveled from a primitive electro-mechanical rhythm machine to the most sophisticated music-generating products yet seen. But, to be honest, we've done little more than scratch the surface of Korg's achievements.

And yet, if you look back over the years, you'll appreciate that the company could have expanded even further. Many ideas were discussed and discarded. Others, such as a six-VCO monosynth, a stroboscopic tuner, and a synthesizer played by touch-sensitive switches, never made it out of the laboratory. And imagine how many products there might have been if Korg had not ignored sampling from 1987 to 1999, or if they had continued to develop studio effects units, or their desktop sound modules, or their soundcards, or if they had persevered with their pro-audio recording and editing systems.

It's also interesting to note how Korg integrated other manufacturers' technology within their products. The mid-'80s FM synths, the IVL harmonisers and KARMA are just three examples of other manufacturers' technology packaged in innovative ways. And what of the longevity and flexibility of Korg's own developments? The synthesis engine developed for the Korg organ at the turn of the '70s spawned a family of synths that lasted half a decade. The sample-based AI and AI² have lasted even longer... 14 years so far, and still going strong. Even ACCESS and its HI derivative have already been around for seven years.

Tsutomu Katoh — still at the helm of Korg after 40 years. [Photo courtesy of Korg Japan]Clearly, Korg have been led by an astute man who understands his market, listens to his colleagues, and gives his developers the latitude to attempt something new, and take risks. Even the occasional heroic failures have maintained Korg's image as a bold company, and each spawned mainstream products that, in all likelihood, repaid the investments many times over.

So where does that leave Korg in 2002? The financial difficulties of the early years are behind the company, and — following Yamaha's investment from 1987 to 1993 — it has been back in Tsutomu Katoh's hands for nearly a decade. Katoh himself is still at the helm and, although in his late '70s, he looks closer to 60, still owns a nightclub in Tokyo, and still controls Korg as he has for almost half a century. Indeed, within the company, Katoh and his elder son, Seiki Katoh (formerly President of Korg USA and Korg UK) are referred to as the 'top management', without whose approval nothing happens. Then there's Fumio Mieda, still designing superb instruments, fully 35 years after the first Korg organ prototype. Even Katoh's younger son, Makoto, is involved; he's currently working for Korg Europe.

So it's only right that we should offer the last words to Tsutomu Katoh. We asked him about the future of Korg, and whether he could tell us the technologies he envisaged. He replied: "It is difficult to mention future products, because everything is changing so quickly. However, I have my desires and dreams for products. As a manufacturer supplying tools to create music, I would always like to make products that make musicians really happy.

"Here's one case I experienced. When Korg made the first hand-held tuner in 1975, I received many letters from music teachers. They wrote that, before they got that tuner, they had spent most of their lessons tuning instruments. After they got it, they could teach ensembles. When I read those letters, I was very glad.

"I always dream of providing useful instruments. Even if the scale is small, a company that can produce unique products is still my ideal. It would be the most desirable situation if we could enable musicians to create works that did not exist before, or if we could invent a new musical style by making unique instruments. We know that other companies copy our products, but I always say to our employees, 'learning from other companies' products is OK, but do not make copies... think and produce your own ideas.'

"I do not think any of the conventional hardware products or software synths which are now in fashion have reached their ultimate form. So inside Korg we are trying to seek the form that is easiest to use without regard to the boundaries of hardware and software."

The future sounds interesting. Here's to the next 40 years!

Sound On Sound would like to thank Junko Fukai of Korg Japan and Paul Bundock of Korg UK for their kind help in sourcing pictures for this article.

The History Of Korg: Part 1

The History Of Korg: Part 2

The History Of Korg: Part 3 of 3