Arturia CMI V – Review

Arturia CMI V – Review

When I began writing my review of Arturia’s V Collection 6, it soon became apparent that, due to my deep interest and involvement with Fairlight CMI’s, the CMI V aspect of the package was going to get some deeper scrutiny. After about an hour of writing, and with the end still not in sight, I decided to break out that content and post it as a standalone review. So here it is…

Ok, so a virtual CMI has been spoken about for many years. I’ve become embroiled in the most incredibly nasty and misleading internet arguments about the possibility, likelihood and feasibility of such a thing. I’m not proud of those debates as they often descended into nasty threads with me being pilloried for my point of view. A point of view based on my knowledge of how a CMI works, informed by lengthy conversations with its inventor, amongst other learned folk. The main crux of the conflict has always been around a single premise. That premise is that because the CMI is old technology, so old that the RAM per voice was 16kb (256kb System) in a IIx, sampling was at 8bits and it all ran off a couple of Motorola 6809 processors, a modern computer will be able to model this easily, many times over. So, my opening gambit is usually, “Well, if that’s the case, why hasn’t it been done?” I mean, the demand is clearly there, so if it is so easy to do, why have we not, until now, seen anything that actually attempts to recreate a CMI as a plug in for modern day music making?

Then the debate quickly turns to the iOS app that Peter Vogel produced back in 2011, a project I had some minor involvement with. The conversation usually goes along the lines of, “Look! You can run a CMI on an iPhone!” to which I reply with the simple facts that the iOS app is nothing more than a sample playback device, using WAV conversions of the original .VC sample files. There is absolutely zero processing of the audio outside of the built-in audio chips of the device. Sampling into the app is via the same audio signal path used in the basic iOS Voice Recorder. There are no modelled filters or ADC/DAC’s to shape the sound the way a IIx would. And then the argument moves back to desktop computers and how easy it is for them to replicate a Series II CMI.

Which it isn’t.

You see, aside from the basic nature of a Fairlight’s computer specification (by today’s standards), there is a lot more going on inside than the uninitiated care to realise. For example, on top of the dual 6809 CPUs, each voice card is running its own CPU, as well as Curtis/SSM filters, all responding to custom code that applies the low pass filter on to every note as it is played to roll off the anti-aliasing caused by stretching a single, sub-2 second sample across a 73 note keyboard in either direction. Then there’s the ADC/DAC process and a very high quality audio chain to think about, and some very clever jiggery pokery to make the whole thing work seamlessly. By the time you have modelled that lot in a regular Windows or Mac PC, you’ve started to eat up more than an acceptable amount of CPU and RAM, enough to affect the overall performance of the host computer, the host DAW and any other plug in or application you might be running. It is for this reason that when Peter Vogel set out to recreate the CMI for its 30th anniversary, he used a bespoke Fairlight Crystal Core CC-1 FPGA DSP card to do all the heavy lifting. Those cards retailed for a fair few thousand pounds and featured 32bit floating point DSP processing, specifically designed for use in professional, high-end audio hardware, such as the Fairlight EVO range of post-production consoles. Only then could a consumer level PC faithfully recreate the antiquated design of its forebears.

There have been people who have told me, with huge amounts of conviction, that they are in the process of creating a fully functional CMI plug in, but at the time of writing, none of these have ever materialised. And then, out of the blue, Arturia make this announcement! So, did they actually do what I have said is currently not feasible? Was I wrong?? Is it all smoke and mirrors?

Well, yes and no!

CMI V is, in my humble opinion, the best, and possibly only, attempt at recreating the CMI experience as a plug in. However, it is most definitely not a carbon copy. What Arturia have done is taken the three key features of a CMI IIx and fashioned something that recreates elements of the CMI experience in a respectful, yet modern way. It’s not without some “issues” or that might just be me being picky, but I do want to point out, even at this early juncture, that I am a big fan of this plug in already!

When you fire it up, you’re presented with a lovely image of a CMI. Well, actually, it’s not. It’s Arturia’s hybrid version. Sure, the styling is incredibly familiar, but they have taken some artistic licence. In fact, they did make a bit of a boo-boo with the design of their mainframe. When developing the CMI V, Arturia took loan of a Series II provided by a French acquaintance of mine, Monsieur Jean-Bernard Emond. JB (as he is known) is a well known restorer of Fairlight’s, amongst many other great synths of the time, and given his location, Arturia could not have done any better when recruiting a collaborator with an actual CMI. But JB has adapted his Series II with a flash drive kit, because the original 8″ floppy drives are notoriously flakey and unreliable, and these flash kits are a nothing short of a miracle cure, allowing an entire library to be stored and no need for constant disk swapping. And if you look at Arturia’s visual recreation of a classic piece of 35 year old equipment, it actually has a very modern flash drive kit! A small detail, that will only ever likely bother purists like me! Other changes to the original design include a tripling of switches and a doubling of faders on the left of the music keyboard, no numerical keypad on the right, as well as a brand new fader module (aka a macro control unit) atop the keyboard with direct access to filter, vibrato and other settings. Clicking on the monitor screen brings that to the fore and this is most likely where you will spend most of your time.

Whilst the screen design and layout bears no resemblance to the original whatsoever, the features and layout are very clearly set out and brilliantly easy to navigate. There are four main tabs, each relating to Sound, Sequencer, Mixer and Tune/Map functions. All pretty self explanatory. The Sequencer is a cleaner, slightly expanded version of the CMI’s revolutionary Page R. Page R was an 8-track monophonic pattern based sequencer that allowed you to create your compositions on screen either by using the light pen to place notes on the tracks, or by simply playing the notes in to a click track. Sounds pretty straightforward, but back in 1982, this was a huge deal. Arturia’s version gives you an extra two tracks and 32 steps per pattern, twice as many as the original. Now some might find this incredibly basic, primitive even, but this is what we had to use back in the day and it was this kind of simplicity that shaped the feel of a huge amount of songs in the CMI’s heyday. In a world where we have seemingly limitless possibilities within a DAW, Page R, or its CMI V equivalent, is quite refreshing. Naturally, most will probably use it to program beats and the like, but try fully restricting yourself to this method and figure out ways of squeezing as much music out of it as possible and you will have a ton of fun.

The Mixer and Tune/Map tabs do much as you’d expect, the former allowing you to control levels and add some FX, both track insert and aux return, and the latter allowing you to map samples across they keyboard, again, very much like the original. Remember, multi-sampling only came with the Series III in 1985!

But it is the sound generation which most people will be interested in and CMI V doesn’t disappoint. Whenever I speak to people who harbour a desire for a CMI, I ask a few simple questions and it becomes apparent in all but a few cases that what people REALLY want is those classic CMI sounds. You know, the orchestra stab (Orch5/Orch2), the breathy vocal patches (Sararr), the swanee whistle (Swannee) or plucked Koto (Kotobend), maybe even the cello patches (Cello 2) used by Kate Bush to great effect. Most people just want the sound of the Fairlight CMI. Nothing else. A few of these would love the sampling capability, to be able to sample something and have it gritted up by the low-fi 8bit processes. Very few want the sequencing and barely anyone has ever said to me that they harbour a desire for the CMI’s additive synthesis capability. That last feature is so overlooked by so many people that it is positively criminal. The CMI was initially, before Peter Vogel hit on the idea of digital sampling, a purely additive synth. It’s genetic predecessor, the QASAR M8, and the machines that preceded that, were designed by Tony Furse to be additive synthesisers that would replicate acoustic instruments through the manipulation of harmonics. Sadly, or in many people’s eyes thankfully, it never worked that well, which is what led Fairlight to explore the concept of digitising sound and storing it electronically for playback. But the additive synthesis aspect never went away, and thanks to Arturia, we can now begin to explore it again.

Additive synthesis in the Fairlight, as I have previously mentioned, was mostly overlooked, especially in the Series III where it never really worked that well at all. I tend to regard it as the CMI’s “appendix”. It’s part of its physical makeup that is somewhat redundant and never really used. However, that doesn’t mean to say that you can’t come up with something quite unique and inspiring. Plus, on the Series II, it always looked cool, drawing in waveforms and having the CMI calculate and morph the two together. If you want a good example of how the additive synthesis of the CMI can sound, check out this track by Ben “Benge” Edwards from his album ‘Twenty Systems”…

So it is really exciting to see that Arturia didn’t just focus on sampling with the CMI V and have had a really good attempt at coming up with additive synthesis that both honours and respects the Fairlight’s methods, whilst putting their own twist on things. The Time Synth mode allows you to “draw” in up to 32 harmonics or wavetables, chosen from a selection of 6 of the former and 4 of the latter. This isn’t quite the same as drawing using the light pen, but it does the trick. You select your harmonic then begin to “draw” amplitude envelopes for each, adding points that you can vary to affect that amplitude. You can simply combine a bunch of sine waves or get really experimental and use a combination of sine waves, PWM and wavetables to come up with some very unique sounds. You can then convert this to a sample and work with it from there. Alternatively, you can take any sample, be it from the library or one of your own, and analyse that as a set of sine waves that you can then manipulate in the Time Synth mode by tweaking individual envelopes and then resampling it. I won’t begin to profess that I have much of a clue about the power of additive synthesis, but this is a pretty unique method of sound design and could, with a little time and effort, yield some riveting results.

Then there’s the Spectral Synth which is another, more simplified approach to additive synthesis that did not exist in the original at all. Using a spectral display that shows 32 harmonics, you can shape a tone by adjusting what Arturia call the Evolution settings, which adjust the shape of the spectral display in a number of ways. You can do this up to ten times, once for each slot, and then use the mixer to adjust levels and add FX as you see fit. Personally, I’ve not had any spectacular results with this area of the instrument, but it can be interesting to play around with and throw things back and forth between the synth and the sampler.

And what about the sampler? Well, this is where most people’s attention will be drawn to and to start you off, Arturia have provided a wealth of presets in the usual location that feature classic Fairlight sounds as well as a raft of patches with a more modern feel to them, all combining one or more of the available sound generation methods. But CMI V also comes with a later iteration of the original Series II library. Now, bear in mind that the Fairlight factory library was incredibly fluid and organic, and despite there being a number of documents that seemingly catalogue the content floating around online, was never quite the same from week to week or month to month. That said, what is provided is pretty much as definitive as it gets and you will soon find yourself going, “Ooh!! I know that one!!”. Load these up into available slots, layer them up or split them across the key range and away you go. It won’t be long before you’re constructing patterns in the sequencer that ape some of the classic rhythm tracks of the 80s. Editing these samples is limited to a couple of filter settings, sample rates (from 2kHz to 44.1kHz) and bit depth. You can mess around with tuning, loop points and directions and amp envelopes as well as portamento, glissando and bend levels.

But the one thing CMI V doesn’t do is actually sample. That’s not particularly unusual these days, although it sticks out like a sore thumb to me, but you can, of course, import your own WAV files. You can even import original CMI .VC files, although some people have reported issues importing .VC files from the final revisions of the CMI II. Arturia tell me they are aware. But this means that if you can find any old .VC files, you could potentially bring them back to life in CMI V. There are some other iterations of the Series II library knocking about online, and some of these feature some different samples, so this is a very welcome feature.

Remember this, though. CMI V does not fully emulate an original CMI’s hardware. I am told that they accurately modelled the filter chips, which were/are fairly common (Curtis/SSM), and that Arturia modelled the ADC/DAC to a degree, but there’s no true emulation of the actual CMI process, hardware or signal paths. So if anyone tells you that this is a proper CMI plug in, it isn’t. It is an approximation and a very good one at that. Until this point, we have had to make do with sample libraries of varying degrees of quality. Some of us are fortunate enough to be able to play with the real thing, but this is the first time anyone has put out a commercial plug in that has a proper stab at doing more than just sound like a CMI. I think many will compare this to UVI’s Darklight IIx, but that is an unfair comparison. Darklight is a sample library with a nice, scripted front end and some very useful features, such as the pattern sequencer and some useful sound shaping tools but it is nothing more than a sample library. CMI V goes beyond that by delivering the factory library in full, which Darklight doesn’t do, and having a really good go at Page R. But the fact it embraces the additive synthesis and resampling is where it earns more points in my book.

For now, it’s the best way to have a CMI experience outside of a CMI itself. It’s also a heck of a lot cheaper and, quite possibly, easier and quicker to use.

And there you have it. In spite of what many might have thought, I really do think Arturia did a great job with this plug in, and I sincerely hope they continue to develop it. What could they add? I’m not too sure. A lot of people have bemoaned the fact that it is based on the Series II and has no Series III features or content. Aside from the Series III factory library, which is indeed very good, I really don’t think there’s anything that could be emulated from the Series III that would be of any use. The Series III introduced 16bit sampling, with sample rates up to 50kHz (Stereo)/100kHz (Mono) so samples were lovely and pristine, just like they are today. The Series III featured multi-sampling, something we’ve had for many years now. It gave us the CAPS sequencer which, whilst superb in 1985, has been massively outdone in the intervening years. I honestly don’t think it could bring anymore to the game. So maybe, if Arturia can get to the bottom of the copyright and licensing of the Series III factory library, that would possibly be the only worthwhile upgrade. And I’d pay to have that, even though I have an extensive factory, 3rd party and end user Series III library at my fingertips!

In my humble opinion, CMI V is the best way to get your hands on a really good CMI recreation outside of getting the real thing. It delivers an authentic sound along with a modern take on the features that made it the legend that it is today.

  • Posted on February 24, 2018 - 3:55 pm
  • By Rob Puricelli
  • Posted in
comments so far
  • Adrian Hendy says:

    Nailed the review – it’s basically how I feel they should have done it

  • Raymondh says:

    Great review! Do you have any thoughts on comparing this to Bitley’s Fairlight refill? If a Reason user has WBF, would there be much benefit in the CMI V? cheers. (I’m not sure if WBF samples were raw samples or taken after the DAC/filter stages)

  • They’re kinda two different beasts, really. The Bitley library is superb value for money and contains the factory library as well as hundreds of really well crafted patches based on those samples. But that’s it. Nothing more. And for most, that’s just fine.

    CMI V offers all of the factory samples, a large amount of very nice presets but also delivers additive and re-synthesis modes and more “authentic” control over the samples and sounds. It also comes with a Page R equivalent and deep integration with Arturia’s KeyLab controllers.

    So, if you’re just after the library and some great patches, WBF might be better value, but if you want more, then CMI V is the way to go.

  • I was considering the CMI V as it is half off right now. I’m nothing near a CMI purist. Nor am I after the exact same patches of any particular 80s band. I am actually interested in the sound of it as an instrument. It’s cool to hear that it ‘passes muster’. I’ve been checking it out from the Demo. But sadly arturia limits the demo to 20 minutes!

Leave a comment