Tom Oberheim has returned to the analogue synth fold with a revised and updated version of his classic '70s monosynth, the celebrated Synthesizer Expander Module.
The Oberheim Synthesizer Expander Module was one of the world's first self‑contained synthesizer modules, developed so that owners could play sequenced lines without tying up valuable synths such as Minimoogs or ARP2600s. But, right from the start, there was something a bit special about the SEM, so it wasn't a total surprise when Tom Oberheim used it as the basis of all manner of instruments including dual‑ and four‑module synths (the '2‑Voice' and '4‑Voice'), a 4‑Voice expander, and a huge eight‑SEM dual‑manual polysynth (the '8‑Voice'). Even today, vintage Oberheims that are not based on SEMs are judged by how well they emulate the SEM sound, hence the reverence the OB1 and OBX command, and the lesser esteem bestowed upon the OBXa and OB8.
Consequently, there's a thriving market for second‑hand SEMs and SEM‑based synths, but Oberheim himself wasn't interested in resurrecting the design until pressure from Roger Linn and a bunch of end‑users at an event in Barcelona caused him to re‑think. Today, thanks to them, the SEM is back in production, and Oberheim claims that, "the new SEM is pretty much the same as the original... in fact, I made the circuitry as close to the original as possible.” There was an easy way to check this; I took one of the SEMs out of my 4‑Voice and removed the review unit from its case. Superficially, they appeared somewhat different, but after a few seconds' inspection it became apparent that, despite the use of surface‑mount components and hugely improved board quality, much was the same. I wondered if they also sounded the same...
When compared with the originals, the oscillators in the new SEM are remarkably accurate; both waveforms (sawtooth and pulse) sound great, and the 'sync' parties like it's 1979. To be fair, there are small differences, but the differences between the four SEMs in my 4‑Voice are almost as great as those between them and the reissue. Perhaps the only germane difference lies in the method of control. Whereas the original SEMs used somewhat temperamental dual‑concentric knobs for coarse‑ and fine‑tuning, the new one uses a conventional knob for coarse tuning and a smaller one for fine tuning. In truth, the new system is more accurate and easier to control, so you'll hear no complaints from me.
Whereas much of the SEM was quite basic, its 12dB/octave filter was remarkable, and in many ways defined the instrument. Unlike, say, a 24dB/octave Minimoog, ARP2600 or Odyssey filter, it didn't self‑oscillate, which made it impossible to obtain some important 'synth' sounds. However, it offered four modes (low‑pass, high‑pass, notch and band‑pass), which made it possible to obtain a diverse range of sounds not available on the competition. You selected these using a single knob that offered the band‑pass response via a switched position at its maximum anti‑clockwise position, or you could sweep through the other three as you turned the knob clockwise. This meant that you could obtain filter profiles that were notch‑like, but biased toward low‑pass or high‑pass by turning the knob to somewhere lying between either of its extremes and its 12 o'clock position.
The new SEM retains this architecture but, instead of having a band‑pass position on the selector knob, it has a dedicated switch. However, while static sounds can be programmed to be all but indistinguishable on the original and reissued SEMs, the minimum and maximum cut‑off frequencies of the new unit, and its response to ENV2 (the 'filter envelope') are somewhat different. And, in case you're wondering, the filters in the four SEMs in my 4‑Voice are remarkably consistent from one to the next; on this occasion, it's the modern unit that's the odd one out.
These differences make themselves further felt when inspecting the dual ADSD envelope generators and the LFO. For example, the slowest attack and decay/release rates are slower on the originals, and the maximum modulation depths can be greater or lesser (depending upon the destination). Meanwhile, the maximum LFO rate (over 50Hz) and depth are both greater on the reissue. This means that it can be hard — sometimes impossible — to program complex sounds that are identical on the original and the reissue. Don't misunderstand me... the new module is a genuine SEM in both sound and character, not just a visually similar but functionally inferior knock‑off such as you might pick up in a Bangkok street market. But some patches are going to respond differently, while others might never have been available on the original in the first place.
When assembled within an Oberheim synth, an original SEM may be augmented by extra goodies such as a keyboard controller, a programmer, a portamento section, tuning and filter cutoff frequency controls, octave transpose, and a stereo output module with level and pan for each SEM. Inevitably, the additional facilities on the reissue are somewhat different, and the exact nature of these will depend on which version you buy.
The review unit came with the MIDI/CV panel installed, and this retains facilities such as the portamento knob and octave transposer, but adds a bunch more that are appropriate to the modern world. Firstly, there's the MIDI/Channel panel, with channel selection, the choice of the one MIDI CC# that the unit will recognise at any given time, and the pitch‑bend range. Below this, a pair of panels allow you to select low‑, high‑ or last note priority, the last of which offers retriggering, if desired. Alongside the retrigger on/off button lies something new that makes a significant difference to the sound: filter tracking, offered as a single '100 percent' option.
Unfortunately, the Source Destination panel is not as powerful as it looks, because it only allows you to direct one from a choice of velocity, aftertouch or the selected MIDI CC# to the four possible destinations (VCO1 pitch, VCO2 pitch, VCF cutoff frequency and VCA gain). This means that you can't direct (say) velocity to the filter cutoff while using aftertouch to play pitch‑bends and vibrato. Likewise, the VCA will only respond to a MIDI‑derived CV when its switch is set to the 'On' position, thus disconnecting it from ENV1. This means that you can make the loudness of a sound velocity sensitive, or you can shape it using the envelope, but you can't do both simultaneously. The manual actually addresses this point, claiming that making the VCA respond to both MIDI and the envelope simultaneously would have entailed unacceptable changes to the design, which is a shame.
Even when compared with other synths of its era, the SEM was a simple instrument, and the reissue retains its immediacy. Sit it on a desk, plug in MIDI, plug in an audio output cable (and a couple of audio inputs if you want to use it as a sound mangler) and plug in the cable from the external power supply. There are no configuration pages or software downloads to frustrate and confuse you; just let it warm up for a few moments and it's ready to go.
But the simplicity of the original SEM didn't mean that it was sonically limited and, despite the differences noted above, I found the reissue to have the same large, warm, and sometimes aggressive American character as the original. And where it differed, it often did so in a good way. For example, the greater envelope depth applied to VCO2 when programming 'sync' sounds means that the reissue can create much more potent sweeps than the original, and I can't see that anybody is going to complain about that!
You can also use this version of the SEM as a stand‑alone MIDI/CV converter: the pitch CV is output from CV Out 1, the auxiliary CV is output from CV Out 2, and the derived gate is output from... you guessed it, Gate Out. I tested this by hooking up the SEM's CV1 and Gate outputs to the CV In and Gate In of a Roland SH101, and it worked perfectly, first time. The only significant limitation is that the SEM's secondary CV is output only through CV Out 2. Most of the time, this is fine, but when (say) aftertouch is routed to oscillator pitch, the SEM itself responds, but the SH101 has no knowledge of the second CV, so the two instruments go out of tune with one another. It's a shame that the summed CVs are not available from Out 1 when appropriate. Be aware that this might limit what you can do in this regard.
Extending the patch, I took the output from the SH101 and fed it into one of the SEM's audio inputs. I was then able to shape the Roland's output further using the SEM's filter and amplifier. Mixing the Roland's output with the Oberheim's own oscillators, and then filtering and shaping them all together, I discovered that I had an even more powerful and flexible instrument at my fingertips.
So are there any criticisms? Of course there are. The first may seem trivial, but it caused me some head‑scratching before I realised what was going on. In short, CV2 is called CV2 because CV1 (which isn't called anything) is the pitch CV. Oberheim's documentation calls CV2 (and this is an exact quote), the "second auxiliary control voltage”. Even adding a comma would have helped (the "second, auxiliary control voltage”) but naming it the Aux CV would have been even more sensible. To be honest, the manual has more than its fair share of typos, so a thorough proof‑read wouldn't go amiss.
Secondly, it's worth restating that the MIDI version of the SEM responds only to note on/off and pitch‑bend, plus a choice between velocity, aftertouch or a single user‑defined MIDI CC number. I can't see that limiting it this much was necessary to preserve the architecture or the sound of the original and, like me, you might find it frustrating.
Thirdly, while you can store a single setup of the MIDI/CV panel, there are no patch memories. Some people have suggested that this is a benefit, because the addition of memories reduces the resolution of the analogue controls on the top panel, thus leading to audible stepping. This is arrant nonsense. Sure, control voltages are sampled with finite resolution when a memory is stored, but with modern technology this isn't an issue. I suspect that those people are confusing MIDI control of an analogue parameter (which usually results in just 127 discrete values) with digital storage, which, at 32‑bit resolution, offers billions of possible values. Consequently, I suspect that the decision to preclude memories was actually based upon the increased complexity and additional build cost that would have resulted.
Finally, I always question the use of an external power supply with professional kit, and the SEM's case is large enough to house an internal PSU with a standard IEC socket. But if you incorporate your own PSU, you have to submit the whole unit for expensive and rigorous electrical safety testing, whereas if you use an external low‑voltage supply with an existing CE or FCC stamp, you do not. Nonetheless, Oberheim's choice of a 24V DC PSU is a little worrying; there may not be many replacements on the market, should you ever need one.
There's a lot of crap written about analogue synths, and statements such as "[the SEM] sounds great because there are no digital circuits or thin‑sounding 'voice chips' in there” are guaranteed to annoy, especially when the same writers may eulogise about Memorymoogs and other vintage synths that contain huge piles of those nasty‑sounding voice chips and digital circuits. Likewise, claims that not just the original circuitry but the original components were used in the reissue are disingenuous, and Oberheim himself claims much more modestly that "its circuitry is almost identical to the original”. So let's ignore all of this and get to the root of the issue. Does the new SEM sound good, does it sound like the original, and should you want one?
I imagine that there must have been many temptations to tweak the design and bring the SEM into the 21st century, so Oberheim has to be complimented on resisting them. Consequently, the reissue not only sounds very good, it sounds very similar to the original, which is great news for its many fans. And, given that vintage units change hands for around £750, the reissue is good value for money. Sure, there are cheaper alternatives (see box) but, for some people, the SEM is an instrument of choice regardless of price. I'm rather disappointed that you can't hook up multiple units to create a polysynth but, as a monosynth, the reissue is everything that the original was, and more. So, to answer the final question, you'll probably want one. But be prepared to wait — there's already a queue.
The most obvious equivalent to the SEM is the Analogue Solutions SEMblance, which is clearly based upon the Oberheim. However, unlike the SEM, it offers extras including Sample & Hold, a noise generator, and greater modulation capabilities. So, while the SEMblance has a similar sound and specification to the SEM, its manufacturer says that it "is not intended to be and is not a clone of the SEM” and goes on to say, with refreshing honesty, that, "If you want a true SEM sound, buy a SEM!” Nonetheless, at just £399$749, the SEMblance has to be a contender for your cash.
Perhaps the most interesting alternative to either of these is the DSI MoPho. Its detailed on‑board programming system and software editor are a million miles from the philosophy of the SEM but, in recreating the sounds of their classic forebears, they share much common ground. Unfortunately for Oberheim, the large panel and hardware of the SEM means that it's significantly more expensive than the MoPho. It's also more expensive than the DSI Tetra, which is in essence four MoPhos in a box, and which I have previously compared favourably to the 4‑Voice. But, for some people, only the immediacy of something like the SEM will do.
The SEM reviewed here is not the only version available, and the most significant alternative replaces the MIDI/CV converter with a patch panel offering 33 analogue inputs and outputs on 3.5mm jacks. This loses the portamento and audio input pre‑amplifier of the MIDI version but in many ways is more powerful because it provides a more complete means for controlling the SEM than the limited capabilities of the MIDI interface. It's an ideal companion for almost all modular or semi‑modular synths, as well as for analogue sequencers and, despite its lack of CV mixers or multiples, you can even treat it as a mini‑modular in its own right, because there's nothing stopping you from patching its outputs into its own inputs, to create sounds that are unobtainable from the MIDI version. Well, unobtainable, except that all of the patch points brought to the front panel are also available on the circuit boards of the SEM itself, so I have no doubt that some MIDI versions will be modified by their owners. Indeed, if you fancy a bit of gratuitous modding, you can even purchase the naked SEM (without the case, without the MIDI or patch modules, and without a power supply; £$TBA) for integration into other systems.
There are two audio inputs on the back panel of the MIDI/CV SEM. EXT1 features a preamp, so that you can drag low‑level signals such as guitars and microphones up to line level. Unfortunately, you can only mix one of these inputs into the signal path at a time. By way of recompense, perhaps, the EXT2 channel can also produce a useful, internally generated A440 reference tone.
One of my first questions when hearing about the MIDI/CV SEM was whether I would be able to stack four of them to create a modern 4‑Voice expander. Sure, it wouldn't be cheap, but for some people, the temptation of four MIDI SEMs sitting alongside one another (or, better still, mounted in a single case) might just be enough to provoke a call to the local office of the Honourable Association of White Slave Traders.
Unfortunately (or fortunately, for the loved ones who were just about to be carted off to sunnier climes), polyphonic use would require a MIDI Overflow function that passes unplayed notes from one module to the next, and the SEM's MIDI/CV panel doesn't offer this. It's a shame but, given that Oberheim has blessed the SEM with a MIDI Thru, you have to wonder if this is something that he might consider for the future.
- Number of voices: one.
- Oscillators: two, each offering sawtooth wave and variable-width pulse wave with PWM, plus sync.
- LFO: one, generating sine wave only.
- Filter: one x 12dB/oct multi-mode low‑pass, high‑pass, band‑pass and notch.
- Contour generators: two, ADSD.
- External audio inputs: two.
- Channels: 1-16.
- Responds to: note on/off, velocity, aftertouch, pitch‑bend, one MIDI CC# in the range 1-99.
- Key assignments: low, high, latest (with optional retrigger).
- Transpose: plus or minus two octaves.
- Other: portamento, filter tracking, LFO retrigger, audio preamp for input one.
- Format: Desktop module (not rackmount).
- Size: Approximately 29 x 21 x 13cm.
- Weight: 2.7kg.
- Power supply: external 24VDC 'wall‑wart'.