The all‑time best‑selling synthesizer, Korg's M1 laid the groundwork for synths that followed. We go behind the scenes to reveal the secrets of its success.
In a marketplace where a synth that sells a few tens of thousands of units is considered a success, one that reportedly sold 250,000 surely exceeds a manufacturer's wildest hopes. Such an instrument was the Korg M1, the widely‑beloved Sample + Synthesis workstation that can rightly be called the most popular synth of all time. Released in 1988 at a UK retail price of £1499, it was manufactured until 1995 — and seven years is a very long time in music technology. Although Korg won't verify the quarter of a million figure I've just mentioned, they do tell me that 100,000 were manufactured during the first two years of the M1's life, serial number 100,000 having rolled off the production line in November 1990.
Why such enormous success for this particular instrument? As you'd expect, there was more than one reason...
Sampling, that mainstay of modern music, was growing in popularity at the time of the M1's gestation, but DRAM (Dynamic Random Access Memory) chips were very expensive, which helped to make samplers quite expensive too. Although the M1 isn't a sampler, its ROM (Read Only Memory) contains four megabytes of musically useful and downright stunning 16‑bit PCM (pulse code modulation) tones. Included are superb drum and percussion hits — a first for a sample‑playback synthesizer — and exotic instruments that previously hadn't been heard by many in the mainstream. The M1 also has onboard effects, which are more diverse and of better quality than those found in the near‑contemporary (and very successful) Roland D50. In addition, the M1 has a built‑in eight‑track MIDI sequencer with battery‑backed memory. This sequencer might not be as user‑friendly as the slightly earlier Ensoniq ESQ1, but it's enhanced by the inclusions of pattern construction and drum machine‑style loop recording. Perhaps none of the M1's basic facilities was completely unique to it, but they were specified and combined in a way which obviously gave it an edge.
For its time, the M1 had a very good feature list, and it's not bad even now: a 61‑note keyboard that senses both key velocity and aftertouch, a joystick for pitch‑bend and modulation control, 16‑note polyphony, eight‑part multitimbral operation with dynamic voice allocation, and 86 16‑bit sampled waveforms within that 4Mb ROM memory I mentioned earlier.
User memory can be flexibly allocated, between program, combination, and sequencer storage: you can choose to store either 100 Programs, 100 Combis, and 4400 sequencer events, or 50 Programs, 50 Combis, and 7700 sequencer events. A Combi (Combination) consists of up to eight Programs, allowing you to assign different sounds in layers or split zones, or set up voices on specific MIDI channels for multitimbral sequencing applications.
Two card slots are provided as a means of quick and simple memory expansion. One slot accepts RAM cards for storing and directly accessing Programs, Combis, and sequence data. The other takes PCM cards containing alternative waveform data. Plenty of third‑party companies provided support of this kind — I've listed them in the 'M1 Variations & Add‑Ons' box.
The M1 shipped with one hell of a sound set — and it was unusual in that it had the same sound set for every country it shipped to. The situation had previously been that manufacturers commonly shipped their new synths and samplers with sounds pertinent to specific countries. As Jack Hotop, Korg's premier sound programmer, and a man who played a key part in the M1's development, told me: "In the old days, a new synthesizer would come out and programmers in Italy, Germany, the UK, and the United States would create sounds for it. Everybody would say, 'I know the market in my country, so we should put my sounds in this synthesizer.' Consequently, Korg would have to ship their synths and samplers loaded with sounds created specifically for England, Canada, and so on. We weren't really unified. Then Korg Inc. in Japan bought Unicord, the US distributor, and formed Korg USA. Now we were one big happy family."
Korg founder and chairman Tsutomu Katoh, and his son, Seiki, then decided that each instrument should ship with the same factory sounds, regardless of its destination. The company assembled an international team to develop the M1's sounds. As well as Jack, the group featured Peter Schwartz, a composer, keyboard player, programmer, and musical director who's worked with David Bowie, Cher, Enya, Madonna, the Pet Shop Boys, and Hanson. (Not only did Peter clean up the M1's strings and vibes sounds, he also helped to create other crucial timbres.) The team was completed by Athan Billias, Ben Dowling, Robby Kilgore, Jim Bescher, Michael Geisel, and Michele Paciulli.
On paper, the lists of 100 multisounds and 44 drum and percussion samples in the M1 seem quaint now, given the current state of digital synthesis and the gallons of gigabytes of sampled sounds in the known universe. Still, the sounds Jack and the gang squeezed into the M1's modest 4Mb of sample ROM played a key part in making the instrument a phenomenal success.
If we were to categorise these sonic offerings, we could come up with four groups. There are sampled attack transients, followed by single‑ and multi‑cycle loops or lengthier multisampled loops; single‑cycle sustained waveforms that lack attack transients; percussive attack samples; and some rhythmic loops. In plain English, you get timbres such as acoustic piano, luscious strings, realistic acoustic guitar, eerie woodwinds, exotic sitar and kalimba, a wind‑chime pattern, a vibrant hammered metal pole, and killer drums. "You don't need lots of memory for certain types of sounds," Jack explains. "Short loops will work for flute, trumpet, French horn, marimba, and Clavinet, leaving you more memory for choir and low piano notes, which need more memory and crossfading."
At times the sound collecting went smoothly, but the pace was frantic, due to production deadlines. "I bought a pan flute in Yokohama," Jack reveals, "and Athan, who had been living in Tokyo, found a huge sake bottle to sample. It made a really low tone that was deeper than most other blown‑bottle samples. Athan used to play flute, so he played the pan flute and blew across the top of the sake bottle for samples that are in the M1."
New York session musician and synth programmer Robby Kilgore, who's worked with Carly Simon, Riuichi Sakamoto, Michael Brecker, Steve Winwood, Steps Ahead, Laurie Anderson, and many others, delivered big time when the sound team was in search of a proper bass timbre. "We needed a slap bass," Jack recalls. "Robby Kilgore said, 'I play bass.' But the strings on the bass were dead, so they didn't slap well. We needed something zingy, but we couldn't get zingy. So Robby pulled out a pick and started playing the Gunsmoke theme with muted pick bass. It was magical. You still hear the M1's Pick Bass a lot today. In the end, we found a good slap bass in the Korg DSM1 library, from a split bass‑and‑brass program."
Individual M1 programs can use one or two oscillators. The former is required if you want to trigger 16 notes at a time, although Korg's effective implementation of dynamic voice allocation, which will quickly reassign silent oscillators as needed, can often compensate for taxing polyphony requirements. Single‑oscillator sounds generally aren't too thin‑sounding either, thanks to the M1's complex sampled waveforms and onboard chorus and reverb effects.
You can adjust the volume balance of the two oscillators in a double program, and detune the second oscillator using interval and fine‑tune parameters. In addition, the attack of the second oscillator can be delayed by up to five seconds, for further complexity.
Each oscillator is paired with its own digital low‑pass filter and filter envelope. The filters aren't resonant, and sweeping their cutoff won't sound as smooth as it would with analogue filters. However, those aren't major irritations, as the sound set is harmonically rich enough to offset the lack of resonance, and the effects can mask any swept‑filter graininess.
Separate LFOs are provided for vibrato and filter modulation, and in a double‑oscillator voice you can assign each LFO to modulate either or both voice channels. It's also possible to program each LFO's rate to be modulated independently by joystick movements.
One thing that makes the M1 particularly special is Korg's implementation of modulation sources. Jack Hotop explains: "I went to Japan eight times for the M1, starting in late '86. I got to work directly with Korg's great engineers, who aren't just guys running around in lab coats. They're musicians, too, and it shows in their work. When I first met with them about the M1, I asked them to provide more resolution for the LFOs at minimum and maximum values, and to add envelope intensity. Peter and I asked them to allow both positive and negative velocity values for crossfading, and velocity and keyboard tracking control for the envelope segments. We felt these were crucial enhancements for bringing short samples to life. The envelopes are the legs that our programs stand on, the arms that they reach out with. These changes would make our programs as expressive as possible."
Korg's engineers didn't settle for simple ADSR envelopes. Along with the usual attack, decay, sustain, and release rates and levels, the filter and amplitude EGs include parameters for attack level, break point, slope time, and release level. Added to the pitch EG's envelope parameters are start and release levels; there's no sustain‑level parameter because the key's basic pitch determines that level. All three EGs can be programmed to respond to key velocities by varying the overall envelope level and time. In addition, the filter and amplifier EGs can be set to track to key position for different types of response.
When you dig into the M1's effects section, you'll find a considerable amount of power — especially considering that the concept of putting multi‑effects processors inside a synthesizer was a very fresh one at the time of the synth's creation. Among the effects algorithms are reverbs, early reflections, stereo delays, chorus and flange, tremolo, two‑band shelving EQ, distortion, an exciter, a Leslie simulation, and numerous multi‑effects combinations. Both series and parallel signal‑routing configurations are available for the two processors.
Since the M1 has only the two effects processors, all of the Programs assigned to a Combi can't have access to the same effects they used in Program mode — a shortcoming which dogged early workstation synths from more than one manufacturer. However, routines are provided for copying effects parameters from one Program or Combi to another, and there are specialised Combi effects in which each processor runs two effects algorithms simultaneously, such as delay and reverb. If a dual‑algorithm Combi effect is assigned to each of the two processors, you could construct a four‑Program Combi in which each Program has access to its own effect. Usefully, any program in a Combi can be set to bypass the effects entirely.
Anyone who's taken the time to properly mic an acoustic drum kit can surely appreciate the convenience of a synthesizer loaded with great drum sounds — and the M1 was the first. You can store four independent drum kits, each made from up to 30 different sounds assigned to specific MIDI note numbers. Since the M1 has four outputs, drum or percussion sounds may be assigned to specific outputs for external processing.
One of the best drummers I've ever played with assembled a great kit comprising two Roland Octapads and an M1. Thanks to the great M1 sounds, Korg's flexible drum‑kit setup possibilities, and the way he programmed his kits, he could immediately switch from one set of timbres to another. Few can tell that the drums in our recordings weren't acoustic.
The M1's sequencer, admittedly, imposes some significant limitations: you have to define a song's time signature up front (and it can't be changed later), only two, three, four, five, or six beats per measure are allowed, and a song can't exceed 250 bars in length. However, on the up‑side, enough memory is provided for 10 songs and 100 patterns; automated punch‑in and ‑out are supported; data on up to eight incoming MIDI channels can be recorded simultaneously; you can sneak program changes into any of the eight tracks; there are editing utilities for copying, inserting, deleting, erasing, and quantising measures; and step‑editing allows you to change the pitch, duration, and velocity of notes. Quantising can be applied during or after note input, and resolutions range from a quarter‑note to 1/48 of a quarter‑note — essentially no quantisation, as that's the rate at which the M1's clock runs.
Yes, the M1's sequencer is limited in comparison with what we have today — but it's easily good enough to allow you to get valid work done.
Jack Hotop sums up the M1's importance to the music industry: "The M1 had a lot of innovative features for its day. It had sampled PCM sounds in ROM, a built‑in sequencer, and a good variety of digital multi‑effects. We also gave it immediate memory expansion via card slots. Insert a RAM card and you could access more banks of sounds. If you wanted more samples, you could plug in a PCM card that you could carry in your pocket."
Was the M1 the original workstation? "Workstation, schmirkstation. The Ensoniq ESQ1 gave you sounds and a sequencer; that was a workstation. How about the Kurzweil 250? That was sure a workstation. They were fine instruments, and they were part of the evolution. It isn't a contest. What is a workstation? It's the evolution of keyboard instruments. It's also the integration of man, music, the sounds of the world around us, and technology. We've seen everything from the harpsichord, pianoforte and pipe organ to the Prophet 5, Oberheims, Jupiter 8 and DX7. These are the stations where we work. They're all music workstations. But is it work? Why don't we just call them playstations? Because Sony already did, that's why!"
If you're in the market for a powerful and still‑popular vintage digital synth, consider the M1. It packs lots of goodies that make it a great musical tool. If you were lucky enough to own one — and smart enough to keep it — you'll know how special it is.
Soon after releasing the M1 keyboard, Korg introduced the M1R. It's a great 2U rackmount unit — it's what I own — that went out of production in 1992, three years before Korg stopped making the M1.
In 1990 Korg brought out the M1REX, an expanded version of the M1R that comes with 8Mb of 16‑bit PCM samples from Korg's T‑series, and a ROM card containing 100 Programs, 100 Combinations, and 44 drum sounds. The EXK‑M1 and EXK‑M1R were ROM memory expansion kits for upgrading the M1 and M1R with T‑series sounds.
If the M1's limited sequencer memory cramps your style, Cannon Research's 1989 Frontal Lobe would please you no end. Developed by Michael Cannon, the Frontal Lobe provides an extra 13,000 or 62,000 events, depending on the model. It also fortifies the M1 with a 3.5‑inch disk drive for sequence, program, and combi storage. In '91 Cannon came out with the Version III Frontal Lobe, a 16‑track sequencer that could address 32 MIDI channels via dual ports and was available with 9,000‑ and 49,000‑event capacities. Also in '91, Cannon released the PCM Channel, which allows the M1, M1R, M3R, and T‑series synths to play samples stored on Frontal Lobe disks.
One of the neatest products introduced at the March 1991 Musik Messe in Frankfurt, Germany, was the Zadok SAM1. Made in the Netherlands, the SAM1 could read samples stored on disk in Korg DSM1 and T1, Akai S950 and S1000, and Digidesign Sound Designer formats, allowing you to play the sound data on a Korg M1, M1R, M3R, T‑series, or Wavestation synth.
As the M1 flourished, third‑party companies developed tons of sounds for it. Among them were Eye & I, Greenhouse Sound, InVision Interactive, Kid Nepro, Livewire Audio, Patch/Works, Pro‑Rec, Soundsations, Sound Source, Synthware, Technosis, and Valhala. Most of them aren't around any more, but a few have survived, such as Eye & I (www.voicecrystal.com) and Kid Nepro (www.kidnepro.com). The Northern California company InVision specialised in enhancement ROM boards for synths including the Emu Proteus 1. (An InVision Protologic board lives inside my Proteus, and I wouldn't have it any other way.) InVision's Plus+1 ROM Wave upgrade for the M1 was so good that Korg offered an M1 Plus with the upgrade pre‑installed.
Korg built on the incredible success of the M1, and their growing reputation as the workstation manufacturer, with a series of later workstations: 1989's T‑series; the 01/W (1991); the X‑series of 1993/94; the ground‑breaking, deluxe Trinity, launched in 1995; 1996's N‑series; 1999's Triton; and, most recently, the Karma and Triton Le (2001).
We don't have space to do more than mention the incredible Korg non‑workstation synths that came after the M1: 1990's wavetable‑based Wavestation, 1995's physically modelled, analogue‑style Prophecy lead synth, 1997's Z1, and 2000's knob‑laden MS2000. It's an impressive track record.
Korg's synth‑making history goes back further than you may have imagined. The company started doing business as Keio Electronic Laboratories in 1962, and their first product was the Disc Rotary Electric Auto Rhythm machine, or Doncamatic DA20 (c. 1963). The Korg name — an amalgam of Keio (pronounced Kayo) and organ — first appeared on an experimental keyboard that had programmable voices in 1968. Fifty DA20s were produced. March 1973 saw the birth of the miniKorg monophonic synth, whose popularity convinced Korg founder Tsutomu Katoh that he should commit more resources to synthesizer development.
During the mid '70s, Korg brought numerous synths to market. The dual‑oscillator 700S monophonic came out in May '74, followed by the duophonic 800DV Maxi‑Korg in March '75, the 900PS preset synth and SB100 Synthe‑Bass in November '75, the dual‑oscillator 770 monosynth in September '76, and the M500 preset synth in September '77. None of these instruments set the world on fire, either in terms of the numbers sold or their synthesis capabilities. (They're extensively profiled in SOS's April '98 Retrozone feature.)
The big turnaround in Korg synthesizer success came in December 1977 with the introduction of the PS3100 (see Retrozone February 2001) and PS3300 modular polyphonic synths. These were serious, powerful, and expensive systems that challenged modular synths from ARP, Moog, Roland, and other manufacturers. Korg followed up in kind with the PS3200 modular polyphonic in December '78, but catered for a wider modular user base in May of the same year by unleashing the popular patchable MS10 and MS20 monosynths (see Retrozone November 1996), and the patchable MS50 monosynth module in January 1979. Other Korg synths that appeared in 1979 included the decent‑sounding KP30 Sigma preset monophonic performance synth and the pricey Trident synth/brass/strings machine (see the July '95 Retrozone).
November 1981 saw Korg release a pair of interesting synths, the Mono/Poly and Polysix. While the former could function as a monophonic or four‑voice polyphonic synth, the latter provided six‑voice polyphony and complete programmability for an unprecedented price of only £899. The Polysix was replaced in 1982 by the Poly‑61, which listed for about the same in the UK but included numerous enhancements.
Then came MIDI. Korg's first synth to support this earth‑shattering development was the eight‑voice Poly‑800, which appeared toward the end of 1983. Costing a mere £529, the Poly‑800 was a fully programmable synth at a breakthrough price. With only a four‑octave keyboard, it was tiny and lightweight. Korg even had the foresight to make it battery‑powered, and fitted side‑mounted buttons so that you could slap on a guitar strap, sling the keyboard around your neck, and carry it as if it were a guitar. The Poly‑800 was a very popular item, with a reported 100,000 being sold. Its companion, the EX800 synth module, came along in 1984.
Three years before the coming of the M1, Korg launched the DW8000 (see Retrozone December 1998), a hybrid eight‑voice synth that, like the previous DW6000 (Korg's first hybrid), combined digital oscillators with analogue VCAs and VCFs. The DW8000's keyboard sensed velocity and aftertouch, and had a built‑in digital delay line.
In 1986, Korg unveiled their first sampler (which was also a synth), the DSS1, profiled in November 1997's Retrozone. Korg's Jack Hotop remembers it well: "The story of the M1 really begins with the DSS1. Finally Korg had come out with a sampler. But it was in the mid‑'80s, during the heyday of samplers. The Akai S612 and the early S900 rackmount samplers were popular, as were the Sequential Prophet 2000, Emu Emulator II, Fairlight CMI, and Synclavier. All of a sudden there came a flood of samplers from Casio, Korg, Kurzweil, and Roland. Every sampler you could name had more memory than the DSS1. They had at least half a meg, when the DSS1 only had 256k! But the DSS1 also had some magical stuff: two built‑in DDLs that you could route signals through in series or parallel, programmable EQ, and variable bit resolution down to two bits. It also had crossfading on the oscillators, you could draw single‑cycle waveforms using a value slider, and it would let you create waveforms by plotting harmonics numerically — not unlike additive synthesis. But the DSS1's interface intimidated some people, and there weren't many resources available for learning about sampling techniques."