The Roland RS202 and Multivox MS202 are two classic string synthesizers that seem to share more than a passing physical and sonic resemblance. We investigate the machines and the myths, and solve the mystery once and for all.
Some keyboard instruments are of such obvious importance, or boast such interesting stories that they demand huge Retrozone articles covering many pages of Sound On Sound. A few, as I've discovered when writing about the Yamaha GX1 and the ARP Quadra, deserve two months' coverage. But there are some about which it seems that there is almost nothing to be said. These are keyboards of such painful obscurity, or of such obvious simplicity, that it seems almost impossible to envisage how one could write a few hundred words them, let alone a few thousand. Take the subject of this month's Retrozone as an example. It's a string machine. It has just a handful of controls, and other than generating a pleasing wash of stringy, synthy noise, it does bugger all. So, what else is there to say, you ask? Well... there's lots. As you will see, we have only to scratch away a little of the surface to find that things are never quite as straightforward as they seem...
Before we discuss the Multivox MX202, let's travel back in time to an era before the invention of modern polyphonic keyboards. If you wanted to create a sustained sound for pop or rock music in the early part of 1966, you had three choices: you could play an organ, an organ, or an organ. Or a Mellotron. OK, that's four choices, but with a MkII Mellotron costing a mind-numbing £5000 (which was well over double the price of a nice three-bedroom semi in a middle-class suburb of Basingstoke), you were unlikely to get your hands on one of those for Saturday's gig down the pub, and even in the studio, the Mellotron was a rarity. If you were one of The Beatles, you had one. But if you were Gordon Reid (aged 7), there was no chance of you playing a keyboard that emulated the sustained tones of brass or massed strings. If you wanted to thicken the sound of your music by adding the sounds of an orchestra, you hired an orchestra.
These days, Ken Freeman is best known for writing the theme music for BBC1 dramas like Casualty and Holby City. But back in the mid-'60s, he was designing a radical new keyboard that he called a String Synthesizer. This used the 'divide-down' technology of cheap organs, but each note comprised three oscillators affected by up to three vibrato oscillators, and thus sounded 'chorused'. Unlike the cheap organs on which it was based, the String Synthesizer had a lush tone that was not dissimilar to that of massed strings.
At the start of the 1970s, after a failed attempt to launch the String Synthesizer in partnership with a company called Ling Dynamic Systems, Freeman was almost bankrupt and was unable to put his keyboard into production. So it was left to Lowrey to release a somewhat cut-down version called the String Symphoniser. This sacrificed one of the oscillators, so while it still sounded good, it never quite matched Freeman's prototypes. In the meantime, Freeman had not been secretive about his ideas, and neither had he patented them. Consequently, by the time that the String Symphoniser appeared, there were other manufacturers ready to enter the Strings arena..
The Dutch organ manufacturer, Eminent, was perhaps the first company to produce a commercially successful 'string' synth. Marketed variously as the Eminent Solina, the ARP Solina, and the ARP Model 2100 SEIV String Ensemble, this employed a far simpler sound-generating system than the one developed by Freeman. Rather than create a lush chorus using multiple oscillators and multiple modulations per note, the Solina comprised just a single set of divide-down oscillators and a three-stage chorus/delay unit. As we now know, the Solina was a huge success and became a mainstay of the mid-'70s pop and rock world. I remember The Enid even used two of them, modified so the player could control the amount of chorus applied to the sound. This allowed them to produce a much wider range of solo and ensemble sounds than would otherwise have been possible. The Solina eventually spawned true American string synths such as the ARP Omni and Omni II (which proved to be among the most successful instruments of the decade), in addition to the ARP Quadra.
But it was in Italy that this form of synthesis was to find its spiritual home, with the Logan String Melody in the vanguard of an army of Mediterranean string machines. These would eventually include the Logan String Melody II and String Orchestra, the Elka Rhapsody 490 and 610, the Farfisa Syntorchestra, the Godwin String Concert, the Vox String Thing, and too many Crumars to name.
As a result, it became easy to generalise the geography of the keyboard market: monosynths came from the USA (Moog and ARP); electromechanical pianos and organs came from the USA (Fender, Wurlitzer, RMI and Hammond); and quirky stuff such as Clavinets, string synths and cheesy organs hailed from Europe, particularly Italy and Germany. But what about the Japanese?
By 1975, Roland, Korg and Yamaha had already entered the modern keyboard market, offering unusual monosynths that avoided direct competition with ARP and Moog by the simple expedience of being cheaper and in some ways less powerful. It came as no surprise when, in the middle of that year, Roland launched the RS101 Strings.
A keyboard of the 'synthesizer in its own flight-case' school of design, the RS101 was very compact and sturdy. It was also much lighter than the ARP Solina, and this, together with its size, ensured that it was more portable and convenient. What's more, it offered many significant advantages over the Solina, so it should have been a huge success. However, the RS101 is one of those rare analogue instruments that appears to have sunk almost without trace, and I have no idea why.
The RS101 offered independent Strings I, Strings II and Brass presets either side of a keyboard split located two-fifths of the way up its five-octave F-to-F keyboard. Each side of the split offered a variable Tone control and independent Slow Attack, variable Sustain and Volume Soft options, allowing you to play bi-timbrally. There was also vibrato and the first appearance of Roland's gorgeous Ensemble effect. Furthermore, the RS101 offered individual VCAs and envelopes for each note. Like the Logan String Melody II, it could articulate each note correctly, and this made it far more playable than the Solina. However, despite the functionality, it bombed.
If the RS101 proved to be an unexpected lame duck, the same could not be said of its 1976 successor, the RS202. Although very similar to the RS101, the RS202 had two new features: a user-controlled delay for the vibrato, and the Ensemble Off/I/II switch that was to become Roland's trademark for the next decade.
With the Ensemble switched off, the RS202 showed its true colours. Strings II produced a thin sawtooth-esque waveform, and if you moderated this using the Tone controls, playing it in the upper register using Slow Attack and a hint of delayed vibrato, it produced a more than passable solo violin. Similarly, you could use Strings I, which generated a fatter timbre an octave below Strings II, to produce a very acceptable cello in the lower octaves. Both of these voices benefitted from the application of judicious external EQ and reverb, and were classy sounds.
If any of the three voices disappointed, it was the one called Brass, if only because it sounded little like brass. Fatter than either of the String settings, and at the same pitch as Strings I, it produced a timbre that was, in truth, rather electronic. If I had to put a traditional brassy name to it, I would say that with the Tone set carefully, it sounded a bit like muted horns — but only a bit. Nonetheless, many players found uses for it.
Experimenting with the Tone controls suggested that these were band-pass filters, with a significant amount of resonance applied to the Brass filter. These allowed you to mute or emphasise certain frequency areas to generate a narrow, but useful range of different timbres.
As you would expect, combining two or three of the voices thickened the sound, and polyphonic playing with all three selected added a slight, if unexpected movement to the sound. Perhaps this was due to the phase shifts introduced by the independent filters in each signal path. But it was the Ensemble that defined the RS202; the 'II' setting introduced a rich chorus, whereas setting the switch to 'I' produced the faster, deeper ensemble effect that became an instant classic, and has remained so to this day.
No fewer than three delay lines based on a 512-stage BBD chip called the MN3002 generated the Ensemble. By determining the amount of delay, the modulation rate and depth, and the relative phase between the three signal paths, the designers could have selected between a wide range of delay, reverb, tremolo, vibrato, phasing and chorus effects. However, like the developers of the ARP Solina before them, Roland's engineers got it just right — the RS202 in full flight sounded gorgeous.
Because of the complexity of the sound generated by the Ensemble, the RS202 could dominate a mix, overpowering other keyboards and instruments and filling every space in the spectrum. However, its greatest strength lay in its ability to produce thinner, more subtle timbres that had a character all of their own. Distinguishable from the Solina and the rash of Italian jobs that appeared around that time, these sounds had a transparent quality that sat beautifully in a mix, complementing other instruments, yet never proving uninteresting.
As a result, the RS202 was soon adopted by luminaries such as Tony Banks, Rod Argent, Tomita and Nick Magnus, as well as by the keyboard players in Jethro Tull and Magnum. But my favourite RS202 player was the recently deceased Pete Bardens. Listen to Camel's Rain Dances LP (1977) to hear the RS202 played through a phaser to create the dreamy sweep that characterised the band's output until Bardens' departure.
In 1998, while on a trip to the United States, I discovered the Multivox MX202 I now own in a tiny but fantastically well-stocked guitar shop in Nashville, Tennessee.
It was in a back room with a price sticker stating that it could be mine for the princely sum of $1375. When I enquired, the shop staff assured me that this was the correct price, because the instrument was new. I asked to inspect it and a short conversation confirmed that it had been in stock for the previous 21 years!
To cut a long and enjoyable haggle short, I offered the store $50 for it and we eventually agreed a figure of $75. To be honest, I think that they were amazed to get a cent for it, and I was very happy with my bargain. The only problem was then to get it back to the UK in the same condition as I had found it, which meant wrapping it in bubbles and carrying it as hand baggage on five flights before returning home.
Given the success of the Solina and RS202, it was perhaps inevitable that other companies should attempt to jump on the string-machine bandwagon. So our American counterparts were not too surprised when in 1977 a shadowy company named Multivox launched the MX202 Strings and Brass.
A keyboard of the 'synthesizer in its own flight-case' school of design, the MX202 was very compact and sturdy. It offered independent Strings 1, Strings 2, and Brass presets either side of a keyboard split located two-fifths of the way up its five-octave keyboard. Each side of the split also offered a variable Tone control, Slow Attack and variable Sustain and Volume Soft options, thus allowing you to play bi-timbrally. It also featured a user-controlled delay for the vibrato, and Ensemble Off/1/2 switches that... Hang on! I'm getting a horrible sense of déjà vu here.
With a couple of minor differences, the MX202 was a dead ringer for the RS202. The keyboard was C-to-C instead of F-to-F, losing seven semitones in the bass and extending the upper range by the same amount. Secondly, the Ensemble selectors were three buttons marked Off, 1 and 2, rather than a single three-position toggle switch marked Off, I and II. And the fader caps were a different shape and colour. Everything else is the same, from the identical layout of the controls, to the gate output on the back that allows you to play the Strings through the filters and amplifiers of a suitably equipped synthesizer.
Placing the two synths alongside each other is an interesting exercise, as is attempting to find the sonic differences between them. The truth is that, unless you're prepared to listen and compare them side-by-side, without external EQ or effects, you'll be hard pushed to identify one from the other. Although there are inconsistencies, they're fairly minor, and in many cases you wouldn't be aware of them in a mix. For example, the attack of the MX202 is softer and less 'immediate' than the RS202, and the unequalised bottom end of the RS202 is slightly thicker and rounder than the MX202. When playing, the lack of the bottom seven keys compromises the cello impersonation, suggesting that Roland's choice of an F-to-F keyboard is the better one. But none of the differences are huge and, given the very specific nature and sound of the famous Roland ensemble, this can't be coincidence, or even parallel engineering by engineers intent on achieving the same result. Let's face it, companies have since tried to emulate the Roland chorus, and have failed. Something fishy seems to be going on.
At this point, you'll be expecting me to tell you who Multivox were, and how the company got away with such an outrageous feat of plagiarism. For a long time, I had no idea. Years ago, I scoured the press and literature of the era, but there was little information available. The earliest concrete reference I could find in the UK lay in a 1978 advertisement placed by a London company calling itself PKP (Professional Keyboard Products). This company, which I seem to remember also imported the Octave Kitten, launched Multivox in the UK, promoting its MX20 and MX30 electric pianos alongside the MX2 Leslie simulator, the MX3000 multi-keyboard (see the box below) and the MX201 echo unit. The last of these looks a bit suspicious: the MX201 couldn't by any chance have been a copy of Roland's revered RE201 Space Echo, could it?
Things began to look even more suspicious when Multivox released the MX880 and MX2000, two monosynths that, at first sight, looked identical to the Roland SH2000, even to the point of using identical names for their weird, synthy voices. Although these offered two oscillators compared with the SH2000's one, both Multivox synths were almost direct copies of Roland's 1973 design.
For many years, I believed that Multivox had copied Roland's designs and that the company was a corporate parasite. Certainly, public opinion has long been that Roland was the innovator, and that Multivox was a rip-off merchant. After all, the MX202 appeared a year after the RS202, and the MX2000 appeared about five years after the SH2000. But if this were the case, why did Roland not sue Multivox and stop them from selling such blatant copies as the MX202, MX880 and MX2000?
In the late 1970s, the Multivox logo included the name Sorkin Music, and adverts gave the company's name as "Multivox/Sorkin Music of 370 Motor Parkway, Hauppauge, New York". So who or what was Sorkin?
One synth site on the web suggests that it was the manufacturer of some later Multivox products, but I believe this to be incorrect. I suspect that Sorkin was a music store that marketed Multivox products in the USA. If you know for certain, please let me know via Sound On Sound.
Hillwood, Multivox and Firstman were among the first generation of analogue synth names swept away by the digital revolution of the early 1980s. But it appears that Kazuo Morioko did not depart the world of synthesizers; he went to work for both Hohner and Akai, where, in 1983 or thereabouts, he helped to design parts of its aborted Akai Studio System, including the MG1212 multitrack recorder and the AX80 polysynth featured in last month's Retrozone.
The serial number plate on the back of the MX202 tells us that the full name of its manufacturer was the Multivox Corporation of America which was based in Hauppauge, New York, USA. Alongside this, the plate declares that the keyboard was made in Japan. So here we have a clone of a Japanese keyboard, produced by an American company, but manufactured in Japan. Could the MX202 be a Roland product in disguise?
Flipping the lid on the MX202 provides a bit of a shock. I expected to see either the Roland or Multivox name on the circuit boards — but I was disappointed. Instead, I found the name Hillwood, a little-known Japanese manufacturer who were active in the late 1970s and early 1980s. This was a company founded by a Japanese gentleman called Kazuo Morioka, whose name translates as First Man Wood Hill. Morioka thus gave his name to the Hillwood synths and keyboards released throughout the mid-to-late '70s, as well as the short-lived manufacturer Firstman. Furthermore, Morioka was working in partnership with other manufacturers, rebadging some Hillwood products as Pulsar, and developing yet more for another Italian company, Solton. But where did Roland fit into all of this?
Opening up my RS202 produced a second surprise. The boards in the Roland may be laid out differently from those in the MX202, but apart from a bit of point-to-point cabling in the Multivox where the Roland uses connectors, you could pull a board out of one of these synths, and stick it in the other. I'm not talking just about geometry here. Other than the keyboard mechanism and the details of which screw goes where, these boards are the same, right down to component level. In other words, the MX202 and the RS202 are the same instrument. But whereas the Multivox tells you who built it, the RS202 does no such thing. There isn't a written mark, a stencil, or even a sticker to identify it as a Roland (or otherwise). Sure, there's the Roland name on the external serial number plate and a QA sticker inside the storage cavity for the mains cable, but these could have been added after manufacture.
So the question may not be who built the Multivox MX202, but who designed and built the RS202? And what of the RS101? Let's try to piece it all together.
The MX3000 was Multivox's finest hour. With bass, poly, and lead sections, an ensemble effect, and a phaser, it falls into the same 'multi-keyboard' camp as the ARP Quadra, which I covered for Retrozone a few months ago (see SOS April and May 2002).
The bass section offered three basic sounds, with three contour types, each with variable release. The lead synth consisted of a traditional VCO/VCF/VCA signal path plus an LFO and, despite its apparent limitations, it produced some interesting sounds. Nevertheless, the polyphonic section held the most promise. This offered eight editable voices, of which three (piano, harpsichord and honky-tonk) were velocity-sensitive.
Given the similarities to the Quadra, it makes you wonder how much of the MX3000 was copied from ARP's ideas or designs. Keep an eye out for the MX3000. If you find one and don't want it, I'll buy it. Then I'll pop it open next to my Quadra and let you know.
Keyboards were not the only Multivox products that 'shared' design features with Roland products. Take the FR6M Rhythm Ace as an example. This was a simple, analogue beatbox with 27 preset rhythms, much like any number of others from the same era. But its knobs and switches were identical to those used by Roland at the time, and the design and case were very similar to those of the CR68 and CR78.
History has demonstrated that Hillwood and Multivox were capable of producing numerous synthesizers, electric pianos, bass pedals, multi-keyboards, effects units (in particular, rotary speaker effects and echo units) and even analogue sequencers. Several later products (with no Roland equivalents) developed by Hillwood but badged as Multivox demonstrate that there was a long-term relationship between the companies. That Hillwood should design and then license one or more products to a third party like Roland does not seem impossible. Unlikely, but not impossible.
On the other hand, Roland might have developed the SH2000, RS101 and RS202, and later licensed its designs to Hillwood/Multivox. But, given that the SH2000 was one of a family of three synthesizers — the SH3 (and its immediate successor, the SH3A), the SH1000 and the SH2000 — that marked Roland's birth, it seems unlikely that just one of these would have been designed or built elsewhere.
There's a third possibility. In the revised edition of the A-Z of Analogue Synthesizers, Peter Forrest has now suggested that the truth may actually lie in the middle ground of joint development. In his interpretation, Multivox would have co-designed the RS101, RS202 and SH2000, licensed them to Roland and sacrificed its own rights in the designs for an agreed period. Later, Multivox licensed the designs to Hillwood, which built the Multivox and Hillwood variants. Does this sound plausible? At first sight, not really — but there are precedents. It's not uncommon for co-developers to hand over the rights to the products they have helped to design, either in perpetuity or for an agreed period.
But the truth turns out to be much simpler than any such convoluted speculation. I contacted Mr Hiro Ueno, the Director of the Development Division at Roland Corporation in Japan, and he informed me that there was neither a licence agreement nor a joint development agreement between Hillwood and Roland. So my initial hunch was correct all along: Hillwood was a parasite, blatantly copying Roland's designs, right down to the component level.
I asked Mr Ueno why Roland didn't sue Hillwood over this. He told me, "There were many copied models, not only from Hillwood but also from several companies in other countries. At that time, we were not large enough to have a specialist for these issues. Of course, we thought that they should be punished, but we were fully occupied designing new models, so we had no time to deal with this."
I told Mr Ueno about my discoveries inside the MX202, and he continued, "I didn't know that Hillwood copied not only the design concept but also the layout and components of the printed circuit board, but this makes sense. It's the easiest way to copy the original design because they don't need to understand the details in the circuit. However, I'm sure that it would have been difficult to retain the quality of the products. We were confident users would choose our original instruments and that those other companies would never survive in the end."
Of course, Mr Ueno is right. Twenty-five years later, Roland is still with us and remains one of the most important innovators in our industry. Hillwood? Gone, and long forgotten.
I reckon there's a lesson here that should be learned by the handful of plagiarising companies in our industry that continue to operate on the fringes of legality and respectability. You can only copy for so long, and then, unless you show some design initiative and quality of your own, customers will begin to drift away to competitors that demonstrate flair and innovation.
At the start of this article, I promised that I could take a keyboard of painful obscurity and obvious simplicity, and find that there was far more to it than met the eye. Indeed, given the host of issues that it has raised, the MX202 is perhaps one of the most intriguing keyboards I have ever covered in Retrozone. As I've written many times before in SOS... I love this stuff!
In Part 2 of my retrospective about the ARP Quadra (see SOS May 2002) I stated that the RS202 was a paraphonic instrument. This was a mistake, and a silly one. Like the RS101, the RS202 offered individual VCAs and envelopes for each note. Furthermore, it had no dynamic filters, so the question of paraphonic versus polyphonic architecture in the filter section did not arise.