For the best part of 20 years, musicians have been asking certain Japanese manufacturers for recreations of their classic synths, and finally one of them has been brave enough to do it. How does Korg's reborn MS20 compare to its illustrious ancestor?
Look back at contributions to the various synthesizer forums in early January and you'll see how vehemently the leaked news of the MS20 Mini was derided. I remember one writer confidently stating that, "the chance of Korg bringing out an MS20 in 2013 is about the same as you and I being able to travel to the moon and back in 2013”. Well, I haven't been to the moon (or back) yet, although there are still eight months to go, so who knows what may happen. On the other hand, I'm already sitting next to an MS20 Mini, which Korg describe as "the classic MS20 recreated”. I wonder if it is.
Many people have speculated that the release of Korg's Monotrons and Monotribe had, at least in part, been experiments to see whether the unending hype surrounding monophonic analogue synthesizers could be translated into sales. If true, Korg clearly feel that the experiment had been a success because Fumio Mieda (the designer of the original MS20) and Hiroaki Nishijima (Korg's current chief engineer) recently went on record to say that "the next step should be a fully-fledged analogue synthesizer”. But should the company design a new instrument, or dredge up a design from the past? Ultimately, I suspect that — having come this far — the decision to remake the MS20 was unavoidable. It dominates the wish-lists of enthusiasts the world over, and Korg had already recreated it in software within the OASYS and the Kronos. The company had even built the MS20iC USB Controller for the plug-in version in the Legacy Collection. So the next step was to decide how closely to mimic the original.
As you can see, the original design was retained and, when possible, many of the original components were used internally. Where these were no longer available or had become illegal because of the ROHS (Regulations On Harmful Substances) legislation that came into force a few years ago, close equivalents were used. Consequently, the most noticeable difference between the Mini and the original MS20 is its size. The official line is that Korg felt that "it would be devaluing to make the new model exactly the same size as the original”. In other words, it's smaller out of respect to the original. I don't accept that for a nanosecond. I think that it was all about hitting the target price, and a full-size model might have made that impossible. But let's be clear, the MS20 Mini is not built into the same case as the MS20iC controller. The keyboard is wider and deeper, the chassis is larger, the spacing between the knobs and sockets is greater and, well, you get the picture.
Although the MS20 looks complex, it's based upon a fairly conventional dual-oscillator VCO/VCF/VCA monosynth. The first oscillator offers four footages ranging from 32' to 4', and four waveforms — triangle, sawtooth, variable-width pulse, and white noise — but there's no pulse-width modulation, which is rather disappointing. The second oscillator offers a different range (16' to 2') plus detune of a little over an octave, and different waveforms: in this case, sawtooth, square, a fixed pulse, and ring modulation. The last of these is misnamed: the waveform is generated by the equivalent of a mathematical operation performed upon the pulse wave generated by VCO1 and the square wave generated by VCO2, and this creates a range of effects similar to ring modulation.
I compared the oscillators of my original 'blackboard' MS20 to the Mini (see 'The MS Family' box) and noticed that the raw waveforms seemed brighter on the original. So I closed the low-pass filter slightly to match them. Now they sounded all-but identical for all waves and at all octaves. This was true even for the Ring setting, notwithstanding the difficulty of matching the controls to obtain the same sound. This was a very good start.
The outputs from the oscillators are mixed and then passed to the filters. There were two generations of MS20, the first of which used proprietary Korg35 filters, while later units used a more standard design. Arguments have raged for years about the pros and cons of these, but almost always in favour of the earlier device. Consequently, Korg decided to use the '35' in the Mini, which was wise. There would have been an outcry had they not.
The first filter in the signal path is a 6dB/oct high-pass filter with resonance (which Korg call 'Peak') and it will self-oscillate at high settings. For many years, commentators described this as a 12dB/oct device, and I suspect that I contributed to that error. But when, in 2000, Korg released the schematics of what lay inside its sealed circuitry, its true nature was revealed. The second is a 12dB/oct low-pass filter, again with resonance and the ability to self-oscillate. This was where the first germane difference between the two synthesizers revealed itself: the waveforms of the self-oscillation generated by the Mini were noticeably brighter than those of my MS20, which means that some patches may not translate perfectly from one synth to the other.
I tested the filters further by sweeping their cutoff frequencies downward while making the filters self-oscillate in the presence of harmonically rich signals from the oscillators. Whereas the filters on other synths pick out a descending series of harmonics as they go, the MS20's filters scream and protest, leaving holes and even creating artifacts that sweep in the 'wrong' direction. Setting this up on both synths produced almost identical results. Then I noticed that the top of the sweep on my MS20 was slightly higher in frequency than the Mini's, so I reduced its contour amount and tried again. The results were remarkable. Notwithstanding the slight additional brightness, any remaining discrepancies lay within the bounds of the differences between two ostensibly identical synthesizers.
At the end of the signal path lies the audio VCA, and this is the one audio area in which the Mini differs from its predecessor by intent. The VCA in the original MS20s produced a pervasive hiss behind the wanted sound, so Korg's engineers made the decision — and another wise one, in my view — to improve this. The new VCA has no adverse effect on the wanted signal, but reduces the unwanted noise, which I wholeheartedly welcome.
There are two contour generators in the MS20. The first, EG1, is a trapezoid with four stages — Delay, Attack, Sustain (which is always at maximum) and Release — and this is pre-patched simultaneously to the pitches of both oscillators. However, instead of adding the contour voltage to the existing pitch CV, the knob that controls the sweep determines from how far below the played pitch the sweep starts and ends. The second, EG2, generates an HADSR contour, where 'H' is a Hold before the Attack begins. This is pre-patched to the audio VCA, where it acts as you would expect, and to the cutoff frequencies of the filters, where the Sustain level takes the value determined by each of the cutoff frequency knobs. When you press a key, the contour sweeps the filter up from below the programmed cutoff frequency, reaches its zenith some way above it (unless S=10), and then decays back to the Sustain level. When you release the key, the cutoff frequency then drops back down to where it started (unless S=0). Depending upon the position of the Sustain knob and other elements within the patch, all manner of unusual results can be obtained.
I would like to say that the Mini's envelopes imitated the MS20's and, for EG1, this was true. But, after a while, I realised that their EG2s were behaving rather differently, especially when I tapped the inverted contour from the patch panel. This seemed to be beyond the scope of component differences, so I measured the output voltages from the EG2 Rev Out socket. I found that when my MS20 was set up to produce an 5V on/off contour, it produced voltages of 4.82V and 0V. By contrast, the Mini produced voltages of 6.48V and 1.23V, which could lead to all manner of differences in sounds. I suspect that this is simply a calibration error and easily fixed.
There's a single LFO in the MS20. This generates triangular and pulse waves, with a knob to sweep the former from sawtooth to triangle to ramp, and to adjust the duty cycle of the latter. The triangular output from this is pre-patched to the oscillators' pitches (for vibrato) and to the cutoff frequencies of both filters (for filter sweeps and 'wah'). However, you can only access the pulse waves from the patch panel.
Ah yes... the patch panel again. If, in 1977, the instructions from Korg's senior management had been to make the MS20 look as elaborate and confusing as possible, I hope that the designer received a promotion. The MS20 looked like something out of a sci-fi movie, but all was not as it seemed.
If you look at the six knobs that control the VCO and VCF modulation amounts, you'll see that they have names such as MG/T.Ext and EG2/Ext. The first part of the name refers to the pre-patched source controlled by each knob, and the second refers to the socket that, if fed a signal from elsewhere, overrides the internal patching. The sockets don't carry the same names as the knobs (for example, T.Ext is called Total on the patch panel) but that all adds to the mystery that is the MS20.
So far, so good, but if you now turn to the top line of modules and sockets on the patch panel, you'll see that there's no way to divert the audio signal from its predetermined VCO/VCF/VCA path. The company claimed that the MS20 offered an 'extremely flexible patching system', but this is a huge exaggeration. Unlike, say, an ARP2600, the patchbay accesses only the control and modulation paths, and even here there are significant limitations.
Nonetheless, in addition to providing access to the pitch CV (Hz/V) and Gate (S-Trig) inputs and outputs, the inverted contours from EG1 and EG2, the pulse output from the LFO and the pink noise from the noise generator, the panel provides access to five facilities that require physical patching before they will do anything. These are the S&H generator, the patchable VCA, the control wheel, the momentary switch and the External Signal Processor, the last of which comprises a pitch-to-CV converter, a band-pass filter, an S-Trig output, and an envelope follower. Korg claimed in 1978, and still claim today, that this allows players to use the MS20 as a guitar synthesizer, but the ESP was always of dubious performance, and it was difficult to get it to track anything without glitching. It was equally difficult to get it to output a scaled Hz/V control voltage over anything but the narrowest range of frequencies. It would have been relatively simple to improve this on the Mini but, again, its performance is virtually identical with that of the original. You may think, therefore, that it inconveniences trillions of electrons for no discernible benefit, but owners have found cunning uses for the ESP, such as shaping synthesized sounds using human voices, or deriving triggers from external audio.
You have to spend some time with the Mini to appreciate just how closely it's modelled on the original synth. Let me offer an example. If you patch the output from the wheel to the VCO2 CV input, it doesn't add to the CV generated by the keyboard, it replaces it. Since the wheel outputs a voltage in the range -5V to +5V, the pitch of the oscillator drops to zero when the wheel is in its central, detent, position. If you then move the wheel further toward you to produce negative voltages, VCO2 remains silent (which is correct) but VCO1 goes sharp. You might imagine that a redesign would eliminate what is, after all, a fault, but the Mini recreates this perfectly.
Here's another example. It's not often realised that, in contrast to its keyboard CV inputs, the MS20's Total and Freq inputs were designed to the V/Oct standard. However, they don't track accurately enough to play the synth in conventional fashion from a V/Oct source. It would have been possible for Korg's engineers to correct this, or even to convert the whole synth to V/Oct, but they didn't. The same dodgy tracking exists as before, and quite rightly so.
Here's a third. If you insert a cable fully into an MS20 control socket, the pre-patched CV is disconnected, and the voltage supplied by the cable replaces it. However, if you insert the cable halfway, the pre-patched signal path isn't broken, and the two CVs are mixed. This means that, for example, you can control a cutoff frequency using the wheel and EG2 simultaneously. It's not as easy to do this with the 3.5mm cables and sockets on the Mini, but it can be done.
I could go on, but everything that I would write would reinforce the same message. The Mini is an MS20 in almost every way. Indeed, by the time I wrote this, I was pretty much convinced that the two significant differences noted above — the brighter waveform of filter self-oscillation and the altered response to EG2 — had become more noticeable as the review had progressed, perhaps because a voltage trimmer had slipped, or some other component had drifted out of spec. The only other germane difference I found was that the VCAs controlling the modulation paths on my MS20 were slightly 'hotter' than those on the Mini. Whatever the reason for this, its only consequence was that, when I set a knob to 10 on the new synth, I often found myself setting it to nine or thereabouts on the old one. So, in the majority of cases, you can program the two synths to sound the same. Sure, there are times when even tiny differences can change the ways in which sounds react or develop, but if you love the sound of the original, you're going to love the sound of the Mini too.
Some people claim that no-one has any problems playing mini-keyboards, but that's bollocks. If I reach for an octave on one synth, I don't expect to hit a ninth or even a 10th on another. Imagine if the Minimoog Voyager had appeared with miniature keys. If it had, I doubt that Moog Music would exist today. Had they asked, I would have beseeched Korg to build a genuine, full-sized recreation of the MS20, even though this would have cost a little more. If I'm blunt, I think that the company has missed the chance to create a 21st century classic.
Partial mitigation lies around the back where, in essence, the Mini's MIDI inputs (both USB and five-pin DIN) convert it into a module that you can play from a full-sized controller keyboard. Only Note On/Off messages are recognised, and those only on Channel 1, but the excellent pitch stability and tracking of the original synth has been retained, so you can play the Mini over a much wider range than its own keyboard allows. Unfortunately, no modulation messages are acted upon, so you can't use performance controls unless you add a suitable MIDI/CV converter to the set-up. This isn't daft; adding things such as the A-Ds and D-As necessary for additional control might have compromised the authenticity of the sound. The Mini also acts as a basic MIDI to CV and Gate converter, producing the appropriate signals at its Kbd CV Out and Trig Out sockets when you play it via MIDI.
This then brings us to the question of build quality; how does the MS20 Mini feel? It's actually quite nice — well built in a lightweight sort of way. However, whereas the pots on my 35 year-old MS20 are rock-solid, those on the Mini exhibit a fair degree of play. Then there's the issue of the patch sockets... 3.5mm sockets can become unreliable when cables are frequently inserted and removed, let alone when they are abused. I had to jiggle the output cable a couple of times to ensure a solid connection during the course of this review and I wish that, if not elsewhere, Korg had retained a quarter-inch socket for this. So treat the sockets and the patch cables with respect, and make sure that you've got at least one 3.5mm to quarter-inch cable to hand, or the chances are that you won't be able to plug the Mini into anything!
One final point: despite room for an internal PSU and an IEC socket, the Mini uses an external power supply. I hate those bloody things. What's more, this one delivers a chunky 1.7A DC, so it's going to be tricky to replace it if you lose or damage the original on the day of a gig.
Aficionados have been clamouring for the re-release of the MS20 for a decade or more. Now they have it, with all of its original quirks, qualities and limitations. Korg can even claim that the new version is better than the original, because the sound is pretty much the same but with less hiss. (I agree with this.) On the other hand, some will feel that it's a mere shadow of the original because its keyboard is too small. (I agree with this too.) Either way, it isn't just for aficionados and it isn't just pandering to nostalgia; many people who weren't even born in 1978 are lusting after it, and it's priced so competitively that, if the keyboard doesn't prove to be a stumbling block, I could imagine it becoming the instrument of choice for newcomers to the wonders of vintage analogue synthesis. As for experienced users, I'm beginning to wonder whether any existing owners of original MS20s will sell them and buy two or three Minis with the proceeds.
Korg's second generation of monosynths included the MS20's little brother, the MS10, and the MS50, a single-oscillator expander that complemented rather than duplicated the facilities of the MS20 itself. These, together with the SQ10 analogue sequencer and VC10 vocoder, formed a family, and very nice it was too.
There were four accessory units to go with these. The MS01 was a footpedal that generated CVs, the MS02 interface converted between Hz/V and V/Oct signals, the MS03 was a pitch-to-voltage converter similar to the ESP on the MS20 itself, and the MS04 was a CV generator and modulation pedal.
There was one other MS-series synth. Nearly four feet wide and two feet high, the 'blackboard' MS20 was built for educational use. Only a handful were made, and just three are known to exist in the UK; the other two reside at Korg UK.
Launched in 2004, the Korg Legacy Collection included a software emulation of the MS20. This offered polyphony, fabulous performance features, powerful effects processors, and dozens of other bonuses, and it even came with a miniature USB controller, the MS20iC, which looks like the MS20 Mini but is a bit smaller again. The sound of the Legacy MS20 is not always identical to that of the original synth, but it's remarkably similar. If you're attracted by the possibilities of MIDI control over all of the functions of a polyphonic MS20, look no further.
In 1973, the Keio ORGan company's first synth offered just one oscillator, had no performance controls, and used such odd nomenclature that it was often hard to work out what it did. Nonetheless, the Korg 700 turned out to be a fine little instrument, and players such as Vangelis and Kitaro were soon to be seen in its company. The following year, the 700S appeared, and this added a second oscillator, a noise source, ring modulation and filter modulation. It was an altogether more powerful synth, and it was followed by the wonderful 800DV, the 900PS and M500 preset synths, and the quirky 770.
Then, in 1978, the Korg MS20 appeared. Based upon technology introduced the previous year in the PS3100 and PS3300 polysynths, it appeared to break new ground in price and performance. I played one of the first in the UK, and it was love at first twiddle. I thought that it must be incredibly powerful, immaturity ensured that I was blinded by its appearance rather than listening to its sound, and thus I spent my money... badly. Its patch panel proved to be far more limited than I had expected and it never sounded as I believed it should.
To understand this, you have to appreciate that the MS20 appeared in an era in which all synths were 'Moogs' but, no matter what you did, an MS20 never sounded like a Moog, so for many years it was viewed by the synth community as sounding thin, tepid, and uninspiring. Consequently, there was a time when you could barely give an MS20 away. Mine sat in its box for years before I decided to sell it. It was immaculate, and I obtained just £95 for it. But by the turn of the century, some unlikely synths were becoming popular. The Roland TB303 had led the charge, and the MS20 soon followed it from zero to hero. A mint example offered for £95 today would have every enthusiast in Europe jumping on a plane to collect it.
Interestingly, when Korg's engineers created the ROM for their 'T' series workstations, they included a bunch of waveforms sampled from the MS20. Given the popularity of the 'T's and their descendants, it's possible that the MS20 has produced, in disembodied fashion, some of the most successful sounds of the modern era.
Polyphony: One note, high priority, single triggering.
Number of oscillators: Two: triangle, sawtooth, variable pulse, noise and 'ring' waveforms.
Number of filters: Two: 6dB/oct HP and 12dB/oct LP, resonant to self oscillation.
Number of contour generators: Two: one DAR and one HADSR.
Number of LFOs: One: triangular and pulse waves available.
Additional modules: S&H, patchable VCA, ESP (external signal processor).
Keyboard: 37 mini keys, no velocity or aftertouch sensitivity.
MIDI: USB In and Out, five-pin DIN In. Note On/Off only. Channel 1 only.
Audio outputs: One line level, 3.5mm TS mono, one headphones, 3.5mm TRS mono.
Power supply: External 9V DC, 1.7A.