Tom Oberheim was a pivotal figure in the development of the analogue synthesizer as a commercially viable musical instrument, and though current synthesis technology is largely digital, Tom believes analogue still has a part to play. His company, Marion Systems, has just unveiled its first synthesizer, which effectively combines vintage technology with a very forward-looking, modular design. Having ensured that the coffee was on, Tom filled me in on his recent past.
"There's no way to talk about Marion Systems without mentioning my old company Oberheim. For various reasons that I don't want to go into too deeply, in 1985, Oberheim Electronics became ECC Oberheim owned by my ex-lawyer. I stayed on for a couple of years, then left the company in 1987 to start Marion Systems. The first product I had planned was an industrial-strength sampler, but there followed a period when RAM prices went sky high, and since the product was based around a lot of memory, it would have been too expensive, so it was shelved. So, as a quickie product I did the Akai S900 upgrade board that converted the S900 to a 16-bit machine and that sold quite well. I also did some consulting, including work outside of the music business, then in December of 1988, Marion Systems started doing consulting for Roland. The first product they assigned to Marion Systems was what eventually became the DM80. I hired a couple of engineers, including Chris Meyer, who was the architect of the DM80, and for about 15 months we worked on the project until Roland took it back to Japan.
"I'd been thinking about doing another synth, and in early '91 I came up with the concept of the Marion MSR2, the modular synth we are showing at NAMM '94. One of the things I've been interested in for a long time is the layering of different sounds. The first time I heard that was in 1975 when Oberheim brought out the modular 4-voice synthesizer. I was talking to Harold Rhodes of Rhodes pianos and he suggested we try playing the synth from the keyboard of a Rhodes piano. I put a crude switch assembly under the keyboard, which meant you could hear the Rhodes sound layered with the 4-voice synth, and it sounded great. It was so synergistic; the result seemed to be so much more than what you'd expect. I wasn't able to pursue that concept with Oberheim for various reasons, but I wanted the MSR2 to be capable of sound layering.
"The MSR2 is a modular synthesizer comprising a 1U mainframe and two plug-in cards, or modules. The card modules utilise a simple connection system which provides MIDI, DC power and audio out. Effectively, each module card is a complete synthesizer, so there's no limitation as to what we can do in the future, and we plan to make the architecture open to third-party developers. Not only can we build synthesizer boards using different synthesis technologies, we could also build boards for effects like reverb, or we could put in a hard-disk recorder for specialised applications.
"The first module, which is the one we're showing here, is an 8-voice, multitimbral analogue synthesizer very much in the tradition of the stuff I did with my old company, except with some semiconductor technology improvements to make it more reliable and smaller. The oscillators use a hybrid analogue/digital approach so that they combine the sound of a VCO with digital stability. The circuit uses a high-resolution digital counter that produces a clock feeding into a custom analogue chip which, in turn, converts that clock to the desired analogue waveform. The pitch accuracy is much higher than typical digital machines and the resolution allows us to create very subtle oscillator detuning effects. Another feature of the HROs [High Resolution Oscillators] is that they can produce clipped triangle waves as well as all the other obvious analogue-type waveforms. The clipped triangle wave sounds very nice and you really can't get that sound any other way.
"Other than the HROs, the configuration is pretty traditional. Each voice has two oscillators, a 2- or 4-pole VCF (low-pass only), a VCA, and voltage-controlled pan, but there's also a lot of modulation stuff and some advanced MIDI facilities. The filter is based around a Curtis chip, so has a similar characteristic to the filter in the Matrix 12 and Matrix 6 or the OB8.
"What we have is a finished analogue module with plans to do a number of different modules over the next few years -- both on our own and in co-operation with other companies.
"That brings us onto the mainframe, which holds two modules. We wanted something that would give the user the ability to do complex things but without too much operational complexity. In addition to holding the two modules, the mainframe also includes a 6-channel programmable mixer, followed by a 2-channel, 7-band graphic equaliser, which is also programmable. The reason behind the 6-input mixer is that it can accommodate the stereo outputs from the two on-board modules, plus a further stereo output from an external synth if required. For example, you might want to add in the audio output from the keyboard synth being used to control the MSR2.
"Each module has its own traditional patch storage system, where each patch is a conglomeration of parameters. In this case there's 200 RAM and 200 ROM patches per module, plus 100 special-purpose patches for splits, layers and so on. But the mainframe also has patch storage at a higher level, and these we call Super Patches. A Super Patch remembers which patches you have set up on the individual modules and also the mixer settings, the graphic equaliser settings and even the patch you want on your controlling keyboard. This allows you to realise a fairly complex system using just an MSR2 and a controlling keyboard synth."
Is a Super Patch roughly equivalent to what some manufacturers call a Performance Patch?
"Yes, though you can use the modules in different ways, depending on what you want to do. The mainframe has two MIDI ports and you can either treat the modules as two independent synthesizers or you can configure the two modules to work as a single 16-voice unit. There's also a 25-pin connector which allows external audio to be brought in for processing by the VCFs and VCAs.
"We haven't made a firm decision yet as to what the next board will be, but it will probably be based on some kind of sample-playback system. Hopefully we can find an interesting approach to make this attractive to someone who already owns a sampler or sample-playback type of synthesizer."
Your past designs have usually been pretty user-friendly. Is this a tradition that you've continued in the MSR2?
"I think that's one of the really outstanding features of this machine. The user interface is very simple; you just work your way from the left of the panel to the right. At the far left is the power switch. In the first column you can select the Mainframe, Module 1 or Module 2. In the next column, you choose whether you want to work with the Presets, edit the presets or edit the system parameters. Then you go through the menus with one rotary encoder and change the parameter values with the other. Having two rotary encoders really makes working with menus a lot simpler. The menu list is quite extensive, but it's pretty easy to get around. A nice feature that we put in is markers; you can mark two pages and then jump back and forth between them without having to go through the entire menu, which can make editing a lot faster.
"Even so, there are a lot of parameters because each voice, in addition to the two HROs, has a DCF, modulation, plus three 6-parameter envelope generators, lots of trigger options and various oscillator sync options. The software engineer came up with a very nice idea, which was to use the voice number as one of the modulation sources, so you could control the voice panning or detuning depending on which voice is playing. Even though there's a lot to program through a front-panel window, I think it's a fairly nice system -- if I can understand it, I think most people will be able to handle it!"
Have you put any quick edit facilities in the software so you can get to the major parameters quickly?
"We haven't implemented a specific quick edit mode, though you could use the marker system to move directly between any two pages you want to edit. If I can think of a nice way to do it, I'd like to include some user-definable pages so you could create pages containing the parameters you're most likely to want to access."
Are you talking to any third-party software companies about editing packages for the MSR2?
"We're talking to a prominent US software company, but since we're just finishing off the software, we haven't been able to write up a very complete SysEx spec for the MSR2 yet; as soon as that's done, the plan is for them to do an editor."
I know that you already have some ideas for new modules for the system, but have you seen anything around the show that leads you to think the future of synthesis lies in any particular direction?
"There's a lot of smoke over physical modelling at the moment, but we'll have to wait and see if there's any real fire. If physical modelling turns out to be interesting, and anybody develops a chip set I can use, then it'll be a module. That's the kind of approach I take; if there's an interesting chip set available, we can build a module around it. We don't have to take months -- doing a module card is not like doing a complete product. We're not only going to do bread and butter stuff, we're probably going to do some oddball stuff too. For example, there's been a lot of interesting suggestions, including old machines that are gone that people would still like, that we can probably put on this card very quickly. Unlike a bigger company, we can do a module that addresses a very limited market and still make it viable. We are a small company and we're going to stay small so we can respond quickly."
At this point someone started to play a few of the patches that had already been programmed into the unit, and the only way I can describe the sound is 'classic Oberheim'. It was all there -- subtly evolving pads, searing filter sweeps and shimmering textures, as well as the inevitable over-the-top sounds designed to show off the gymnastic capabilities of the machine. I asked Tom when he expected to begin shipping production models of the MSR2 with the existing analogue modules.
"What we're showing here are production models, though there are some final refinements needed to the software and we still have to program all the patches for the presets. Two of my patch developers were in the earthquake zone and their patches are buried beneath rubble in their apartments at the moment! Other than that, we have quite a few production units ready built just waiting to go -- we should be shipping within the next month or so.
"I'm really quite excited by the concept of the MRS2 and I foresee most of my time during the next few years will be spent in developing this product line. I think I've put together a form factor that allows the product to grow in a number of directions, and because of the modular approach, the MRS2 isn't a product that's going to be made obsolete.
"My philosophy in building equipment over the past 25 years has been not to try to second guess the musician. Almost every every instrument I've ever built, I've put in the hands of musicians to see what happens. I just do what, from a technical point of view, seems reasonable and cost-effective, then let it go from there. Only time will tell if the MSR2 concept works out, and that's up to the people who use it. Certainly I think we've reached a point in this industry where we have access to all the technology that's needed -- now all we have to do is to put it into a form where people can use it conveniently. The idea of layering synthesized sounds certainly isn't new -- people do it in the studio all the time -- but typically you need to have a lot of MIDI cables, audio cables and an external mixer. Some people even need to have a technician to take care of all that for them, but if we can put all that in one box, then anyone can do it."