AAS Multiphonics CV-3 realizes the potential of this modular software as a synthesis and physical modeling powerhouse. It’s the biggest update to date, with 3-op FM, new physical modeling components, effects, and sound tools. And it’s polyphonic and expressive, with full support for 8-voice polyphony and glides and expression with MPE.
Where CV-3 fits
Multiphonics is AAS’ software modular environment. It’s got the requisite knobs and patch points, but it’s not designed to feel like a Eurorack rig in the way that VCV Rack, Softube Modular, Reaktor Blocks from Native Instruments, or Cherry Audio Voltage are. It’s more like a curated toolbox of top-grade sound creation tools you can patch together. Those tools often reveal more capabilities as you work with them and try different patches – this is all about doing more with less. But there’s also an extensive sound library to get you started if you need inspiration.
The polyphonic and expressive additions alone make for a huge update. Switch to Poly mode, and any patch can become an 8-voice instrument. Plug in an MPE controller, and you get full support for glides and expression, with the ability to patch it wherever you wish. That also includes, incidentally, polyphonic and continuous pressure support from controllers like Ableton Push and Move. And MPE support also means that even though Multiphonics CV-3 lacks tuning support on its own, it instantly works with Ableton Live 12’s Tuning System support. It’s beautifully easy how this works; you just double-click on a tuning and the mono or poly patch magically follows the new tuning. (I should add – other modulars do this; VCV Rack specifically has both polyphony and MPE support. Because Rack is open, though, you have to manage different modules and whether or not the developer supported polyphony and how they handle pitch bend.)
CV-2 already took Multiphonics from an intriguing niche patching environment to a versatile tool, thanks to the addition of audio input/effects operation and a more well-rounded set of modules. But CV-3 is looking more like a real contender. It’s still a far more focused environment than some of its rivals, and you can’t add your own modules – but that’s always been the point, and now it’s likely to find a permanent home in your sound creation toolset. For all their variety, you won’t find the quality of AAS’ physical modeling tools anywhere else, and the patching tools and workflow in Multiphonics hold their own, too.
New modules
I’ll be honest: CV-1 felt incomplete. But AAS have been building up the module offerings in Multiphonics and this release brings a ton of useful stuff. A lot of these modules are now complete enough as a voice that you can build a working patch with just one or two modules (the FM voice, especially), but of course, the advantage of a modular is the ability to expand.
In CV-3, you get a new digital FM oscillator, one new stacked analog oscillator, multiple new physical modeling oscillators and filters, and a bunch of helpers and effects:
FM and analog:
- 3-operator Digital FM voice with some pre-configured modulation options (like pitch-to-decay) and internal envelope, plus FM feedback. It really is a full voice – you even have a topology switch, clearly taking some friendly inspiration from Ableton Operator (well, and the panel of the original DX7).
- Stack VCO Combine up to 7 detuned sawtooths, square, or triangle waves (with separate outputs for each and detune and Stack controls).
Physical modeling:
- Objeq Oscillator This acoustic resonance model is the heart of a lot of AAS’ physical modeling synthesis – we had it as a filter in Multiphonics before, but now it’s also an oscillator, with all the controls for material, density, exciter, and so on, making this a more complete voice.
- Tube Oscillator That’s tube as in the object, not the electronics component. In addition to Objeq, you get this specific model of a cylindrical tube and exciter with a lot more options.
- Tube Filter The tube again, here an open or closed cylindrical tube you can use as a filter – expanding on what Objeq already offered.
- Comb Filter Comb filter with feedforward and feedback – great for producing flanging and metallic effects.
And the rest – new envelopes, new effects, and an oscilloscope:
- Pluck Envelope gives you, well, plucked envelopes. That obviously is a great match for these other tools.
- µADSR Compact version of the ADSR.
- Oscilloscope – two-channel visualization – expanding on a simple, compact one-channel scope.
- Reverberator Room (“studio”), club, hall, and arena presets combine here – okay, great. But the important thing here is, by providing a Decay knob rather than just “Time,” you can generally modulate this Reverb in real-time without nasty sonic results.
- Chorus (Output Effect) This one is only an output effect, but then you can easily build one as a patch – this is just a handy convenience, and it sounds great.
- Distortion (Output Effect) Same deal; you already had the recently added shapers and whatnot. Now you get a simple output distortion, a bit like having an effects pedal.
I want to add some sound examples here but – ha, I’ve only had this build for a few days so I’m not quite ready to commit yet. Tune back in, David Abravanel and I are both patching away, so we should have some videos and presets for you shortly.
Workflow improvements
There are subtle but significant workflow enhancements, too.
Output effects now have their own pop-out panel, which in addition to the distortion and chorus make it easier to access parameters. (I’d love to see the mixer panel get a similar pop-out with solo, etc.!)
There’s a new command tray that pops out of the bottom of each module adding shortcuts for delete, duplicate (with and without connections), and online help. If the module has a clock input, there’s a shortcut for sync – that’s something I haven’t seen in any other software modular environment, and it’s great.
You can also auto-connect modules. I spoke a bit with AAS about this – even in advance of full details in the online documentation. Just double-click any CV input on a module, and Multiphonics will automatically select the most likely source. That means the typical task of repetitively selecting pitch or clock connections can happen instantly. (The feature scene to the same row to the left first, then the top row.) Some examples from the folks at AAS:
- Pitch input auto-connecting to a Gate+CV Sequencer’s CV output, a Quantizer, or Glide
- Gate input auto-connecting to Sequencer, etc.
- Trig connecting to Pulse output of a Sequencer
- Tempo, Clock, and Play connecting to a Clock module (or the Master Clock)
These will also then default in the cases above to the Keyboard if another source isn’t available.
It’s a funny example of something being useful both to newcomers and advanced users – the advanced users just to cut down on the mouse strain and go faster!
Other improvements:
- Macro knob renaming
- MIDI setting preset saving (so you can have your LinnStrument or Osmose or Ableton Move preset!)
- Improved icons
- Keyboard Mod has a MIDI Learn and the ability to assign modulation sources right on the keyboard menu
Deep enhancements under the hood
I expect from the moment you try CV-3, you’ll hear a lot of the work that was done on physical modeling, and you’ll experience right away that MoscilPE and polyphony are seamlessly integrated. (Both David Abravanel and I had that experience immediately.)
But it was also great to get to talk to AAS’ Benoit Charland, the lead developer for Multiphonics, who has managed the project since its debut with CV1. He talked about how they added polyphony and MPE without interfering with the way users interact with the software:
From the start, we decided against introducing new concepts like polyphonic patch wires or special polyphonic modules. We wanted users to have the freedom to connect anything freely, as they expect in
Multiphonics, whether a patch is monophonic or polyphonic. While one-click polyphony was not the easiest solution to program, we believe it offers the best experience for users who want to build
polyphonic synth patches. It also allows making original CV-1 and CV-2 factory patches polyphonic with a single click.Regarding MPE, we once again focused on user experience rather than programmer convenience. We wanted our factory library to be enjoyable and expressive whether using a standard MIDI controller, an MPE controller, or programming a sequence manually in a DAW. The Keyboard module handles MPE glide automatically, with Mod 2 and 3 mapped to the other two per-note MPE dimensions by default. This design choice allowed us to include at least one per-note MPE modulation in every patch of the CV-3 factory library.
Here, to me, is the most significant part and why this is such a landmark for AAS – and such a moment for any of us who enjoy crafting patches with physical models:
Another challenge was improving the classic AAS physical models to support real-time modulation from any CV signal on all parameters without glitches or volume spikes, in line with the Multiphonics philosophy. Our sound designers have created impressive patches with these new physical models, and we believe we have only scratched the surface of their potential.
I mean, this is why I like talking to developers; now I’ll go back and appreciate messing with all of those parameters using CV.
In use
As a teaser, here are some gorgeous sounds from David Abravanel – he’s already built three excellent presets which I’ll share shortly. Pay attention to two other features that are unique to Multiphonics: I can seamlessly switch presets without a glitch. And I can tweak the reverb and even switch reverb modules without any issues as Benoit predicted; I’m just using the UI but CV inputs will do the same.
Conclusions
It’s a great time to be a software modular addict. Even with other options, Multiphonics offers two things you can’t get anywhere else: AAS’ singular sound tools and an integrated environment. I love the open-ended wilds of VCV Rack. But this is something different; in order to get seamless functionality and that ability to patch any modulation anywhere without glitches, you do need a closed collection of modules. Multiphonics was already impressive; with CV-3, I think it’s simply the best available single set of software modules on the market.
Now that Multiphonics has grown, I’m missing a few conveniences offered in other software modular tools. It’d be great to be able to type in module names – that big menu worked well for CV-1, but not with the growing module library here. (Come on, I just want the LPG; I don’t want to scroll to the bottom. I want to type “LPG”!) MPE Tuning support is great, but there’s no internal tuning table support, and fine-tuning is otherwise tricky – some oscillators and resonators lack even a fine knob, and shift-dragging the coarse tuning knob is fairly imprecise.he ability to type in parameter values would come in handy.
But I can’t overstate this enough – we’re near to the full set of AAS physical modeling tools, and the magic happens once you start combining modules. Put Pulse and the filters together, work with Pluck envelopes, mess with these different models, and try feedback. It’s also actively developed (something which sadly can no longer be said of one package that helped define the category, Reaktor).
Multiphonics is efficient, it sounds brilliant – and there really is nothing else like it. It’s amusing to read a Sound on Sound review from 2004 that said “AAS are victims of their own modesty.” I agree. AAS’ software in 2025, as in 2024, offers revolutionary sound capabilities. Even with a wonderful supply of modular software, this is one you shouldn’t miss.
CV-3 is available at a 25% off intro pricing, including from Plugin Boutique (who also now bundle Steinberg Backbone drum designer for free). It’s a paid upgrade, though that’s also on sale:
If you buy something from a CDM link, we may earn a commission.
Multiphonics CV-3 Upgrade from CV-1 or CV-2
Promo pricing is also available directly from AAS.
