Completing the quartet of vintage synth emulations they began with Moog Modular V, Arturia's latest plug-in aims to reproduce the sound of the greatest semi-modular of them all, ARP's 2600. We see how it fares up against the original...
I'm an ARP man; I always was. Sure, the Minimoog was a great synth... limited, but great. But apart from this and the original Taurus pedals, Moog's record was never better than patchy. Sure, you could compare the Minimoog and the Odyssey, but where were the equivalents to the wonderful Pro Soloist, the Axxe, and the Omni? The Satellite, the Micromoog and the Opus 3...? Give me a break. Then there was the ARP 2600. With no Moog equivalent, this combined pre-patched synthesis with the flexibility of a modular synth, all housed in a neat suitcase that didn't need a pair of roadies and a transit van to move it from one gig to the next.
Nevertheless, the bias toward Moog survives to this day, as demonstrated by the host of digital imitations of the Minimoog. But perhaps this is about to change, with the almost simultaneous release of two ARP 2600 software synths. This month, we'll look at the first of these. It's the fourth software-based emulated synth to emerge from Arturia, and as with their previous emulations, it aims to produce a sound as close as possible to that of the hardware instrument, whilst also sympathetically extending the original's feature-set with more modern facilities such as polyphony and MIDI capabilities. It's called 2600V, and as I've done with my previous reviews of Arturia's software emulations, I put it up against the original hardware synth to see just how close the emulation was.
I began by evaluating the primary modules in the signal path: the VCOs, VCF, VCA, the output section, and the envelope generators that control them. The first two graphs below show the sawtooth waveform generated by VCO2 on my grey-face ARP 2600, and on 2600V. The equivalence is clear, and, as one would hope, the sound from these is all but indistinguishable.
The authenticity started to wobble when I inspected the sine waves (see the last two graphs on this page). When directed straight to its output, 2600V 's sine is brighter than that of my ARP 2600 — and how can one sine wave be brighter than another? Patching the output to an oscilloscope revealed all — the sine wave on 2600V is not a sine wave at all. It contains significant overtones, although strangely, directing this wave through the pre-patched signal path eliminates the overtones, leaving the correct, pure tone (see the first two graphs on the next page). That's weird, but gratifying.
Then I came to the triangle waves. As you can see from the last two graphs on the next page, 2600V generates a 'shark's tooth' waveform, while my ARP produces something thoroughly triangle-like. In listening tests, 2600V 's sound is a little brighter and less 'bottomy' than the ARP's, although the difference is subtle. However, I've seen the waveform before; it's very similar to the triangle wave produced by Arturia's Minimoog V (printed in my March 2005 review of that product), and even though it's not identical to 2600V 's, I couldn't help but wonder if Arturia were recycling some of their technology, even though 2600V supposedly models a different synth.
Happily, 2600V 's pulse waves sound reasonably accurate whether static or modulated although, as on other software synths I've used, its square waves are too 'pure'; they're recreations of a near-perfect square wave rather than the imperfect, not quite 50-percent waves of a real analogue. The final audio source is a noise generator that approximates that of the ARP 2600, but with a wider range of colorations.
The VCOs on 2600V offer numerous enhancements over those on the original, all of which are welcome (for more on the various types of hardware 2600 that were released in the 1970s, see the box later in this article). Firstly, each offers keyboard tracking from zero to 100 percent. Secondly, each has a five-position switch offering 4', 8' 16' and 32' settings as well as low-frequency oscillation. Next, the Initial Frequency control is quantised in semitones, which is a huge improvement over the original hardware. In addition, VCO3 has a PWM CV input and sine- and triangle-wave outputs, all of which the ARP lacked.
There are a further two additions which I've separated from the others, because the manual describes them incorrectly. Firstly, it tells you that you can modulate the 'widths' of the sawtooth, triangle and pulse waves of VCOs 2 and 3, but in fact the sawtooth waves are unaffected. Secondly, it says that you can sync VCO 1 to either of VCO 2, 3, or 2 and 3. In fact, only VCO 2 is available as a master.
One idiosyncrasy of 2600V 's oscillators is that they go sharp when you apply audio-frequency FM. This is not in accordance with FM theory so, when notified of this behaviour, Arturia stated on their web site, "FM modulation may not produce such a pitch deviation, neither with a DX7 nor with an ARP 2600" and undertook to correct the fault. But when I tested this on my original ARP 2600, I found that it did exhibit the same behaviour as the current version of the software synth; increase the modulation depth, and the pitch goes sharp!
Unlike the hardware original, 2600V offers no fewer than five filter profiles: 24dB-per-octave low-pass, 12dB-per-octave low-pass, 12dB-per-octave high-pass, 12dB-per-octave band-pass, and Notch types.
Arturia's marketing information states that 2600V has two filters: one that emulates ARP's 4012 filter, and one that emulates the ARP 4072, but this is not the case. Furthermore, the manual makes much of the equivalence of 2600V 's 12dB-per-octave low-pass filter and that of the original synth, but that can't be right, as both the 4012 and the 4072 were 24dB-per-octave low-pass devices. Fortunately, tests show that the 24dB-per-octave setting on 2600V does indeed roll off the signal at approaching 24dB-per-octave, so make of the marketing and manual what you will.
To get a feel for 2600V 's filter, I first tested my ARP 2600. Measurements showed that the front-panel sliders adjust the cutoff frequency from 9Hz to 19.4kHz. With CVs applied to the control inputs, the top end shoots off the top of the audible scale. Furthermore, the self-oscillation at maximum resonance behaves as you would expect. It produces a low-ish output (as set up) of 500mV at 9Hz, around 1000mV throughout the audio range, and it falls away to 140mV at 19.4kHz.
I then tested 2600V 's 24dB-per-octave filter; Arturia claim a frequency response of 10Hz to over 21kHz for this. Making the filter self-oscillate and patching the VCF output to the main output, I found the range to be 32Hz to 18.6kHz. This is almost two octaves short of the claimed specification at the bottom end. Furthermore, the upper frequency of 18.6kHz appears to be an absolute maximum, and no amount of controllers applied to the filter's CV inputs will cause it to exceed this.
While taking these measurements, I noticed that the amplitude of 2600V 's filter oscillation is rather weird, with its highest output at the lowest and highest frequencies, and with a significant dip at mid-range frequencies. If, as I did with my ARP 2600, I set up 2600V 's filter to generate an output of 500mV at its lowest frequency, the output was just 60mV at 1kHz, but 190mV at 18.6kHz. This is not the normal response of an analogue synthesizer's resonant filter.
Once again, though, I realised that I'd seen this response before — and referring back to my review of Minimoog V confirmed my suspicions. Again, it's obvious to see what this implies.
By a happy coincidence, I was able to discuss this matter with Arturia at the recent Frankfurt Musikmesse and it turned out that, as I had suspected, Minimoog V and 2600V share blocks of common code. Arturia justified the use of a common 24dB-per-octave filter by suggesting that, as ARP had been sued in the 1970s for breaching Moog's filter patent, the response of the ARP 4012 would be identical to that of the Minimoog filter. When I suggested that this was not the case, they undertook to look into it and, if necessary, consider changing 2600V 's filter. The same positive attitude prevailed when we discussed the incorrect triangle waves and other observations I'll make below regarding 2600V. I am very encouraged by this, especially since significant upgrades — such as the forthcoming v1.5 for Minimoog V — are supplied free of additional charge to legitimate users.
Returning to the filter itself, Arturia claim that the other four filter profiles are emulations of the multi-mode filter in another ARP synth — the 2500. I have no way to test this, and I'm not going to try to make a judgement based on my quarter-century-old memories of playing a 2500. All I'll say is that these filters extend the range of sounds considerably.
The envelopes were perhaps the weakest aspect of the ARP 2600 because, without modifications, their maximum Attack, Decay and Release times were far from generous. Thankfully, Arturia have not been bound by these limitations, and, in recreating the envelopes and trigger/gate options, I'm pleased that they have extended the maximum times way beyond the original specification.
More improvements lie in the minimum contour times. The first graph above shows the click generated by passing white noise through the ARP 2600's VCA when controlled by the ADSR with all values set to zero. The second trace shows the result obtained from 2600V. As you can see from the third trace, which superimposes the previous two, the durations are similar, but the software synth's amplifier begins to 'close' much more rapidly. The result is a more precise click than that obtained from the analogue synth.
Another difference — although one that you may or may not view as an improvement — is the elimination of the famous ARP 2600 'thump' that can occur when you press and release notes when the envelopes are at their fastest settings. Most players viewed this as a fault, so at least one electronic 'fix' was developed for the original synth. By definition, therefore, the 2600V emulates a modified ARP 2600.
There is another significant change in the VCA section. On the ARP 2600, the horizontal slider at the top applies a permanent 'Initial' gain to the VCA, famously prompting 2600 user Joe Zawinul to ask, "how do you switch it off?" This is replaced on the software synth by a Global Volume control, which is very useful, although not quite as sensible as it seems, especially if you want to play external signals though the filters and effects.
At the end of the pre-patched signal path lie the audio mixer, spring reverb, and stereo outputs. Apart from some minor graphical changes, these appear to be identical to the original's, but the sound is not. To be specific, the 'boinnggg' of a three-spring reverb is hard to emulate, and although Arturia have retained the spirit of a spring, the sound is very different.
Hidden away behind the left speaker grille lie two additional effects: chorus and delay. The chorus is a single modulated delay line whose output is directed equally to the left and right channels, no matter where the output from the main signal path is panned. It's no substitute for the lush textures of a triple-delay chorus, but it's acceptable if used subtly. The basic delay unit offers independent delay times (from 1ms to 3s) and feedback (from zero to 100 percent) for each of the left and right channels, without cross delays. The nice touch here (and nowadays a common one) is the ability to synchronise the delay to MIDI Clock. Ratios are available from one-eighth to nine times the currently set MIDI tempo.
I have criticised Arturia in the past for attaching faulty effects to the outputs of their software synths and, unfortunately, this time is no different. This is because, if the delay is 'wet', merely the wet (ie. pitch-shifted) part of the chorus is passed to the delay. Only if the delay is completely 'dry' (ie. 'off') does the chorus wet/dry mix work correctly.
The Sample & Hold/Electronic Switch section was one of the additions that made the ARP 2600 special, and despite small graphical changes, this has been recreated precisely on 2600V. There's one extra: as elsewhere, you can synchronise the internal Clock to MIDI Clock.
But if the S&H section is true to the original, the same cannot be said of the Voltage Processor section (shown left). It's better! Rather than the original's seven inputs, limited mixing, dual inverters, single lag processor and three outputs, 2600V offers eight inputs configured as four linkable input pairs with CV-controlled mixing, four inverters with on/off switches, four lag generators, and four outputs. The possibilities — such as dynamic mixing of multiple sources to create evil modulation curves — are considerable, and remarkable.
If you wanted to get silly with an ARP 2600, you could present external signals to its inputs and see what happened. However, there was just one place where ARP intended that you should insert such signals. This was the Preamplifier, with its associated Envelope Follower. These are recreated on 2600V, and using them proves to be simple. In my case, it meant loading 2600V as a VST instrument under my chosen VST host (Plogue's Bidule), whereupon it appeared with two audio inputs as well as two outputs. I could then direct the output from other software synths to these inputs, patching the output from the Preamplifier and/or the Envelope Follower to the destination(s) of my choice.
A Ring Modulator lies alongside the Envelope Follower. Strangely, while Arturia's panel has the Audio/DC legending found on the ARP 2600, the switch to which this pertains is missing. Testing showed that 2600V modulator is 'AC coupled' (the Audio setting) which means that the carrier and modulator frequencies are not present in the output. Given a single option, I would select this one, but it would be nice to have the choice.
Up to this point, Arturia's programmers have stuck to the spirit of the ARP 2600, but they couldn't resist adding something from the 21st century. It's the so-called 'Tracking Generator' that's tucked away behind the right-hand speaker grille. The manual describes this as having four audio inputs, four modulation inputs (the on-screen jacks on the left of the module — although it's not clear what's being modulated) and four audio outputs (on the right). As you can see from the screenshot below, this is incorrect, as there are no audio inputs, just the modulation inputs on the left of the module, which affect the frequencies of the control waveforms.
In essence, the Tracking Generator allows you to create four control waveforms using a variety of tools available in the Edit window provided for each. You can then smooth these if desired, and direct the outputs to the destination(s) of your choice. You determine the frequencies of the outputs using the Freq knobs and, as elsewhere, these can be synchronised to MIDI Clock.
In other words, the Tracking Generator is nothing more nor less than four programmable LFOs. While this may seem a bit advanced in an emulation of a 1970s monosynth, the philosophy is not outrageous. On a smaller scale, this is what ARP did when they added an LFO to the ARP 3620 keyboard, thus freeing up one of the 2600's oscillators and extending the modulation capabilities of the complete system. Which brings us to...
ARP developed three keyboards for the ARP 2600. The first, the 3601, was available only on the early models. Next, the monophonic, 49-note 3604P offered portamento and variable tuning, but this was later replaced by the duophonic 3620, a far superior unit that incorporated the dedicated LFO that I've just mentioned.
The 3620 was quite a sophisticated piece of equipment. In addition to the LFO, it offered a two-octave up/down switch, a pitch-bend knob, portamento with a momentary on/off switch as well as a footswitch on/off input, single/multiple triggering, auto repeat, and dual pitch CV outputs.
For aesthetic reasons, Arturia have sited many of these functions elsewhere on 2600V. Most noticeably, the LFO has migrated to the centre cabinet, where it has grown an extra output (a sawtooth wave) and MIDI-synchronisation capabilities. However, the output from the LFO is very strange. Notwithstanding the fact that it goes up to 100Hz (despite the manual saying that it has a maximum frequency of 20Hz), its waveforms exhibit 'stepped' shapes and complex spectra. Counting the levels shows that the LFO is quantised with just three-bit resolution, but when you apply it to a destination, it does behave as it should, with smooth sines and correctly rendered triangles, squares and saws.
Other keyboard controls now found in the sequencer section include the Global Tune knob, plus the Trigger Mode and Auto Repeat switches. These are joined by modern-day functions such as the Mono/Unison/Poly switch and the unison Detune amount. The space thus liberated to the left of the keyboard is re-used to house a number of performance CV outputs — velocity, aftertouch and so on, plus a pair of pitch CV outputs; one at 100-percent tracking, and the other at 400 percent.
The original ARP 2600 appeared in four major guises, although many revisions were not honoured with a change of external design or a new paint job.
The first few (built in 1970/71) incorporated the simple, monophonic 3601 keyboard and appeared in a blue-painted metal case with a wooden carry handle. Despite their cachet as the earliest ARP 2600s, these 'Blue Meanies' were hand-built, notoriously unreliable, and difficult to keep in tune.
The next revision was the shortest lived, and only a handful of 'Grey Meanies' exist. These incorporated two handles and were very stylish, but they were deemed less suitable for ARP's target markets — performing musicians and schools — than the design that was to follow.
The most common ARP 2600s (1971 to 1978) are grey with white legending, and have separate synth and keyboard units. These came in two distinctive flavours. Early models, like the Meanies, used ARP's 24dB-per-octave 4012 filter, but this infringed a Moog patent and was replaced in 1976 by another 24dB-per-octave design, the 4072. Unfortunately, the 4072 had a fault (easily fixed, as it happens) that limited its bandwidth to just 12kHz. If you hear an ARP 2600 and wonder what all the fuss is about, you're playing one without the correction!
The final revision saw the synth adopt the black-and-orange livery of latter-day ARPs. This offered numerous internal improvements, most significant of which was the demise of the epoxy-encased circuitry of earlier models, which made these 2600s relatively simple to repair.
2600V 's sequencer is perhaps the area in which Arturia have kept most closely to the form and function of the past... yet it was never originally a part of the ARP 2600! In fact, it's a recreation of ARP's separate 1601 Sequencer. As such, Arturia's emulation offers 16 CVs, 16- and dual eight-step modes, sequential and random playback, three gate busses, the ability to set and modulate the clock pulse width, unquantised and quantised sequencer outputs, a novel Boolean AND function, and inputs to the quantisers (so that you can convert any two external CVs into semitone voltages).
There's more that's the same as it was on the hardware so, rather than list everything, I'm going to mention the two features that have changed. Firstly, there is no multiple. This is no longer necessary because — as elsewhere on 2600V — you can direct a single output to multiple destinations. Secondly, the voltage ranges have been modified. The CV sliders on the ARP 1601 swept across 10 octaves (ie. 10 Volts) at the unquantised outputs, and two octaves at the quantised outputs. Arturia's version offers ranges of ±4 octaves at the unquantised outputs and approximately ±2 octaves at the quantised outputs. The benefits of the improved quantised range far outweigh any loss at the unquantised outputs.
2600V offers three colours of patch cable: red ones denote connections made from audio sources, orange ones denote signals from modulation and other control sources, and green ones denote modulation signals generated by the Tracking Module.
Connecting could not be simpler. Drag from a source to a destination or hold down the Shift key and click on any socket to reveal the patching menu. A source can be directed to any number of destinations, but each input can only receive a single source at a time.
When you connect cables to one of the 10 sockets with a black surround, you can click on the 'nut', and it becomes a knob that allows you to set the input level between -100 and +100 percent. This is like having 10 additional VCAs at your disposal, and is extremely useful.
First things first. 2600V runs on 1GHz Macs or PCs with 256MB of RAM or more, provided they're running Mac OS 10.2 or higher, or Windows 98SE, 2000, or XP respectively. Loading 2600V onto my test Powerbook G4 was glitch-free, and if you're using a Mac, the software installs stand-alone, VST, RTAS and HTDM versions (if you're using a PC, you also get a DXi version). Having done this, you'll need to set up the toolbar correctly, selecting such things as the MIDI input source, the keyboard range, and the audio channels for input and output. This is also where you save and recall patches, determine the number of voices for monophonic, unison or polyphonic use, and select which part of 2600V you're going to view (the synth itself, the sequencer and keyboard, or everything). The last of these is important, because the full representation is big; 1156 pixels from top to bottom, which makes it too large for many monitors. You can also choose which 'skin' you wish to use: a Blue Meanie, a grey-face or a black/orange-face. However, don't think that the sonic characteristics of each are on offer; the choice is purely cosmetic.
One final aspect of the toolbar deserves a mention before I move on: it's the Magnet tool, which causes the patch cables to move out of the way of the mouse pointer when you want to adjust a control that may otherwise be obscured. If only my modular synths had this facility!
Having set everything up, let's start by noting a couple of things that are not implemented on 2600V. Firstly, the duophonic mode of the 2600+3620 configuration is not recreated. Whether this is a problem or not is moot, especially when you remember that 2600V is polyphonic. More importantly, you can't use the Voltage Processors to create static voltages, or as manually-controlled voltage sources. Happily, the additions are more significant: oscillator sync, the multi-mode filter, the tracking generator and variable keyboard CV tracking are very welcome, even though they bear no relation to the performance of the original synth.
Once I got used to the layout, everything fell to hand (mouse?) as I had hoped. Furthermore, you can assign a MIDI continuous controller to almost any control in 2600V, so I linked the knobs on my Korg Legacy MS20 (USB) controller to sensible equivalents in 2600V, and the software synth leapt into life. I wish Arturia produced a miniature ARP 2600 controller...!
So, what about the sound? Is this truly a reincarnation of the ARP 2600? To test this, I set up a simple, monophonic, single-oscillator patch on my ARP 2600 and on 2600V. With careful tweaking (and no filter resonance) I could create sounds that were almost indistinguishable from one instrument to the other. I could also tweak 2600V with no discernable zipper noise. However, as soon as I added filter resonance, the differences began to show. Furthermore, when using the USB controller, the stepping of the filter cutoff frequency became audible. As my tests suggested, the 2600V's 24dB-per-octave filter may or may not be a good one, but it does not emulate the 4012 of my ARP 2600.
As my patches became more sophisticated, the timbral differences became more significant, and it became apparent that, for some sounds, 2600V lacks the depth of a real ARP 2600. It's sometimes subtle, but there are other times when something has definitely been lost. For example, the Ring Modulator lacks the depth and 'grunt' of the device in the ARP 2600 itself. It's still a ring modulator, but — notwithstanding the loss of the DC-coupled option — this one just doesn't behave the same. I also discovered that it is important to keep an eye on the levels within 2600V, because too much signal will cause clipping and extreme distortion. If pushed, the ARP 2600 will clip by a smidgen, but nothing like 2600V, so you have to keep this under control.
Similarly different, if that makes sense, are the factory sounds supplied with 2600V. These are supposed to be recreations of the sounds supplied on patch charts with the original 2600, but with the original ARP patch book to hand, an audio comparison revealed similarities, but no more than that. Overall, the presets are useable sounds, but they are not the ones from the original patch book.
I experimented further with the sequencer and some of the more esoteric modules. The 1601 is far more flexible than its panel suggests, so Arturia should be commended for including this alongside the 2600 itself. I particularly like the way that the output jumps to the step you're adjusting whenever you move one of the pitch CV sliders. When patched properly, this makes setting up a doddle.
Next, I got a bit daring, using 2600V as a signal processor for all manner of other software synths, virtual Mellotrons and electric pianos. The results could be stunning, and I had great fun sequencing parameter changes, especially when synchronised to MIDI. Furthermore, I was pleased to find that my 1GHz G4 was capable simultaneously of supporting 2600V (with moderate polyphony) and other packages without glitches, pops, or dropouts. My only disappointment was that I could not automate patch changes or the insertion/removal of patch cables.
Finally, I created a range of polyphonic patches that included strings, brass, pads, electric pianos, percussion and effects. Initially, I found that 2600V excels at producing edgier and more percussive sounds such as electric pianos (at which it is superb) rather than strings and brass. Arturia's 'factory' patches seemed to bear this out; there's hardly a string patch among them. But wait... the filter and at least some of the oscillator code is the same as that in Minimoog V, so the same, fat timbres should be available. To test this, I loaded Minimoog V and 2600V simultaneously, and copied the parameters of one of my string ensembles from the former to the latter. Having done so, the sounds were almost indistinguishable. If nothing else, this shows that the user-interface is an important aspect of any synth, guiding the way in which you conceive and create sounds.
Arturia's 2600V takes the form of the original ARP 2600, but adds many facilities not previously available, so we should judge it in two ways. Firstly, does it emulate — and sound like — an ARP 2600? Secondly, ignoring the original, is it a desirable synthesizer in its own right?
The answer to the first question is no, unless your patches are quite simple. Nevertheless, the software retains much of the character of the ARP 2600, and I applaud that. The answer to the second question is more important. The 2600V is powerful, it's extremely flexible, and its sound makes it something that you might choose to use whether it looked like an ARP 2600 or not. Add the sequencer, polyphony, and all the other features and — despite a few niggles — it all adds up to a very attractive package.