Synclavier Regen

Can the Synclavier Regen live up to the near legendary status of its ancestors?

Back in the early 1980s there were two names that were almost guaranteed to make a keyboard player’s heart go all a‑flutter. The first was Fairlight. The second, less well known but with even greater mystique, was Synclavier. Part of the reason for this was that they were so far out of the reach of most musicians that legends were created around them — legends that sometimes far exceeded reality. So when the chance arose to buy an abandoned Synclavier II for next to nothing, I didn’t hesitate. Having handed over the cash, I then loaded my car with three large cases, a video monitor and keyboard from the dawn of computing, plus all manner of pedals, floppy disk drives and manuals, and drove them to a gentleman named Steve Hills who ran Synclavier European Services. He spent the next few hours swapping hardware and loading various software revisions until... voila! It leapt into life and functioned perfectly.

The following day, I proceeded to learn how to use it. Or rather, I didn’t. Sure, it looked gorgeous, but it was a bloody hassle to get anything beyond relatively simple tweaks of the factory sounds out of it. I eventually mastered it, but it hadn’t been my finest purchase. Huge, heavy, and always scaring me that it would take a trip to synthesizer heaven, it fell into disuse even though I still love the ridiculous old beast. But wouldn’t it be nice (I mused for many years) if Moore’s Law eventually made it possible to recreate 100 percent of the Synclavier for one percent of the size, weight and cost. I waited for three decades, but here it is. Or at least, here it might be. I wonder if it’s the real deal.

Understanding The Regen

The Regen isn’t a conventional synthesizer, so I’ll start by attempting to boil its extended Synclavier sound engine down to the essentials.

The bottom layer of a sound is called a Partial, and this is built from two waveforms configured as a 2‑op FM voice. Following in the footsteps of later Synclaviers, each carrier can be generated by either additive or subtractive synthesis, or it can be up to 128 samples placed side‑by‑side across the keyboard, or it can be the result of resynthesizing a sample. The modulator is always an additive waveform generated by up to 24 harmonics that can have any amplitudes and phases with respect to one another. A contour generator shapes the amount of modulation, thus controlling the harmonic content of the sound, while a second shapes the level of the Partial. There are two LFOs — one for vibrato and one for tremolo — and (for all but subtractive synthesis) a chorus effect created by cloning and detuning the results. If you don’t want to invoke FM, any carrier can be used as the underlying sound of a Partial.

Hang on a moment... what’s this resynthesis thingummybob? Invented when RAM was hyper‑expensive, it’s a method of slicing an audio sample into short snippets and recreating (as closely as possible) the sound in each using additive synthesis. In the Regen, you can choose how many slices you would like to use and determine whether you want them to start at the beginning of the sample or some specified time later. If you want to edit the slices, you can create different sounds in each and then play them back as a wave sequence, either crossfading or stepping from one to the next. And when the resynthesized sound is used as a carrier, all manner of unusual results can be obtained. Resynthesis doesn’t always work, and the results can be unpredictable if you present it with enharmonic sounds. Even when it works well the results can be a bit lo‑fi, although this can be interesting in itself, and there are so many things that you can do with the slices — transposing, cloning, looping, and modulating them — that you’re going to love it anyway.

Up to 12 Partials of any type can be combined in a single Timbre. You can determine the levels of each, and there are additional controls over the FM depth and tuning, as well as parameters that allow you to spread the Partials across the soundstage, add Timbre Detune — which is what we would now call Analogue Feel — and to stretch the pitch across the keyboard. (Strictly speaking, some of these act at the Partial level, but we won’t go into that.) A Timbre also includes a multi‑mode ‘per‑note’ filter shaped by a contour that offers control over the start, peak, sustain and end levels as well as the times of each of the stages and the curves of the decay and release. You can use this to create many contours that don’t conform to traditional ADSR shapes. In addition, there are controls for the keyboard mode and portamento, an arpeggiator, and a small range of additional effects: decimation, a ‘per‑Partial’ multi‑mode resonant filter, and a ‘per‑Partial’ reverb.

You can have up to 12 Timbres, each of which exists within a Track, so there are 12 of these within the top level, which is called a Session. Track parameters allow you to do things such as determine the volume, transposition, key mapping and MIDI channels of the Timbres so that you can create splits, layers, and multitimbral performances. There’s also a master reverb that affects all of the Tracks and is stored as part of the Session.

The original Synclavier included a sequencer, but the Regen doesn’t recreate this. That seems sensible; there are much better ways to generate sequences in the 21st Century. What it offers instead is the ability to embed a .MID file in a Session. If you load a Session containing one, you’re presented with play, stop and continue buttons, but now’t else. Since the Regen has 12 Tracks, only the first 12 channels of the .MID file are recognised so, if you’re going to import your own compositions, you’ll have to ensure that nothing important is lost. It’s also worth noting that tempo changes are not recognised.

If the replacement of an obsolete sequencer by a MIDI player is no great loss, the omission of the original’s ability to record and manipulate a sample is a thornier issue. I understand the argument that it’s easier to record and edit samples on a computer and then transfer them, but I still think that it would be nice to be able to sample on the Regen itself. The other addition that I would welcome would be a simpler method for importing Synclavier II sounds. You can do so now using a combination of the company’s Synclavier3 and Synclavier Go! products as intermediaries, but the method is long‑winded and requires two additional products. A direct import option would be much more sensible.

Obviously the Regen is considerably smaller than its predecessors, but at 310 x 260 x 42mm it’s compact by modern standards too.

Programming The Regen

To create or modify a sound, you use the column of silicone buttons to the right of the panel to select the Partials or Timbres that you want to edit, then select high‑level functions using large rubbery buttons, then select specific parameters using other large rubbery buttons, and finally use the Swiper and its associated touch‑sensitive buttons to edit the values. Unfortunately, the method of selecting Partials, Timbres, Tracks and Sessions caught me out time and again. When the left/right arrow button above these buttons is blue, they represent Tracks, whereupon a blue surround means that a Timbre has been inserted into a given Track, a cyan surround means that that Track is active except that, when the Solo button is lit, chartreuse means that that Track is soloed. But when the left/right button is red, you’re dealing with Partials, and the equivalent colours are red, magenta and green. It takes time to get to grips with this, especially since some programming choices will jump you from one level to another. Further confusion reigns if you forget where you are in the hierarchy when loading sounds. Don’t mix up your Sessions and your Timbres or, like me, you’ll find your ladies infected with xylophones (or some other mishap). The other concern I have about these buttons is a more prosaic one. All components can fail and, while you can be confident of being able to find a potentiometer, a fader or even an encoder that can be made to work in the event of a failure, the Regen’s touch‑sensitive buttons and Swiper could become unobtainium in a few years. Let’s hope that the company has bought a huge stock!

There are two further design decisions here that seem odd to me. Firstly, it takes three swipes to move parameters from their minimum to maximum values. Since there’s a ‘fine’ mode, I have no idea why the Swiper isn’t programmed to go from bottom to top (or vice versa) in a single motion. Secondly, its two screens are so recessed that, when the Regen is placed in front of you on a horizontal surface, you can lose sight of the bottom line of data on each. The obvious workaround is to angle the Regen toward you, but a better solution would be to use the mounts on the underside to bolt it to a 100mm VESA‑compliant monitor stand, swinging it into position when wanted, and swinging it away again when not. That’s rather neat.

I have always found the underlying Synclavier engine to be quite intuitive but, like its inspiration, the Regen rewards study. Unfortunately, the manual at the time of review lacked some things that I thought would make it quicker and easier to master. In particular I would have liked to have seen a block diagram to illustrate the synth engine, plus a parameter‑by‑parameter reference section. I discussed this with Craig at Synclavier and, within a few days of our conversation, I received the first draft of a signal flow diagram designed for inclusion in the manual. Although there’s much more detail than can be covered in a single graphic, I think that this will make a huge difference to the speed at which new users learn the system. That was an excellent response.

Once I was ready to start programming, I started with a single Partial and confined myself to basic waveforms for both the carrier and modulator. In no time at all, I had created sounds that were instantly recognisable as emanating from a Synclavier. Replacing simple waves with increasingly complex ones then led to sounds ranging from gorgeous cymbals that morphed into female voices, to strings, brass, pads, percussion, and the inevitable screams of aliased cacophony. Interestingly, it was also easy to create virtual analogue sounds that I could never have obtained from my Synclavier II. When I reminded myself that I could have up to 12 of these Partials in a single Timbre and up to 12 Timbres under every note, the power of the system became really apparent.

So let’s now ask the question that I’m sure you’re waiting for: does the Regen sound the same as my Synclavier II? You might think that, since we’re comparing digits with digits, it should be obvious that it does. But it’s not that simple. In the original, the pitch of each voice was determined by a variable clock, and mixing 16 voices at different clock rates was beyond the technology of the time. Consequently, each voice board in the Synclavier II has a dedicated D‑A converter (which has a very different architecture from today’s equivalents) and the outputs from these are sent to an analogue mixer before passing to the synth’s outputs. So let’s ask a more sensible question: can the Regen sound almost the same as my Synclavier II? Yes, it can and, unless you’re going to carry out a side‑by‑side comparison (and who but a sad old SOS reviewer would be idiotic enough to attempt that?) or are trying to recreate the tiniest nuances of an existing Synclavier composition, I doubt that any differences are going to matter.

If you want to affect dozens of parameters simultaneously to create complex and musically interesting timbral changes, the Regen allows you to do so in ways that would require a whole football team abusing dozens of knobs simultaneously.

Playing The Regen

The Synclavier II was a performance instrument, and so is the Regen. You might wonder how I can say that given its lack of knobs and faders, so let me kick a particularly annoying elephant out of the room. For someone like me, the way to create music on a synthesizer is by programming a sound beforehand, including the connection of any physical controllers to the parameters that I might want to affect in real time. In other words, I don’t use the programming controls as performance controls. The Regen conforms to this model so, if you want to grab a couple of knobs to make a sound go ‘wheeee’, you’ll have to look elsewhere. But if you want to affect dozens of parameters simultaneously to create complex and musically interesting timbral changes, the Regen allows you to do so in ways that would require a whole football team abusing dozens of knobs simultaneously. The elephant, therefore, is not the lack of knobs but the perceived need for them.

In addition to standard MIDI CCs and performance messages, the Regen recognises polyphonic aftertouch and MPE. For much of this review, I played it using a Roli Seaboard Rise 2 and this made it possible to do things such as adding brightness, vibrato and reverb to one note and not others, or bending just one note while leaving others unaffected, or even bending two notes in a chord in opposite directions. Nevertheless, there’s an oversight here: MIDI sync hasn’t been implemented. I raised this with Craig and he told me, “While it would be useful for some things like the arpeggiator, it’s not really essential given Regen doesn’t do anything in the sequencing domain.” I’m really surprised by this — many players will want to synchronise their arpeggios and LFOs to the track tempo. Fortunately Craig then added, “so this nice‑to‑have feature may be added at a later date if we get lots of requests”. OK chaps, I’m requesting.

Despite the power and flexibility of the system, some will inevitably ask whether the Regen would have been a better product if it had been a keyboard that echoes the look and feel of the Synclavier II. It would certainly have been more lust‑after‑able, but it would also have been much larger, much heavier, and much more expensive, and there are many small studios into which a Synclavier clone simply won’t fit. Others will ask whether a large, touch‑sensitive screen might have been a better choice than a complex panel, but I can see that this would feel too similar to a soft synth and wouldn’t offer the same experience as the Regen. All in all, I think that Synclavier have got it about right, although I wouldn’t object to a MkII version with a monitor output!

Buying The Regen?

The Regen isn’t designed for novices and, if you dive into it without thought and attack it with a blunt stick, you’ll probably end up with nothing useful. Nor is it designed for people who want to twiddle a bunch of knobs and call themselves music producers. Sure, there are some happy accidents to be had, but it’s only when you study the system and start to plan sounds in advance that the depth and power of the Synclavier engine reveals itself. As Craig told me, “The learning curve is undeniable, even for someone with prior experience of a Synclavier II. It’s a system that you don’t just buy and use every now and again, it’s something you have to commit to.” So perhaps this whole review boils down to a simple question: do you want a hardware synthesizer from a company called Synclavier that emulates and extends a vintage synth called a Synclavier, is as deep and as arcane as a Synclavier, sounds like a Synclavier and will take as long to master as a Synclavier... or don’t you? If you do, the size, weight and price just dropped by a couple of orders of magnitude.