Moog Muse

A Moog polysynth is always a big event, but is the Muse the inspiration you’ve been waiting for?

There was a time in the early 1970s when all synthesizers were called Moogs, no matter which company built them, just as all ballpoint pens were Biros and all vacuum cleaners were Hoovers. But despite being the world’s most recognisable synthesizer brand, Moog have hardly been prolific and, if you list their polyphonic instruments from the past 55 years, you’ll find that there have been just seven, two of which weren’t publicly released, and another of which wasn’t a polysynth at all. So today I have in front of me Moog’s eighth (or maybe fifth) attempt to capture the hearts and minds of the polysynth market. I wonder whether it will succeed.

While neither as large nor as heavy as the Moog One (with which it shares some features) the Muse is still a chunky beast weighing in at 14.5kg and, notwithstanding some wobbly faders that hark back to the days of the Rogue, Liberation and Opus III, it feels solid and robust. Its 61‑note keyboard offers both velocity and aftertouch sensitivity but neither polyphonic aftertouch, nor MPE, nor the more exotic performance capabilities that recently appeared elsewhere. So, if you want to control this with that while wiggling something else, you might need to look somewhere else.

Despite boasting almost 200 controls, its panel is clear and (in my view) very attractive, and its menus — many of which are accessed using the triangular buttons in the voicing panes — have just a single tier, which helps to keep things simple. All of the housekeeping is carried out in the Programmer section toward the centre of the panel and, while small, the monochrome screen is adequate if you don’t mind scrolling up and down a bit.

The Signal Path

The basic sound generation unit of a Muse sound is the Timbre, and this can be polyphonic, monophonic, or unison monophonic. A patch contains two Timbres and you can switch between these, split them or layer them. The eight voices are then distributed according to your choice from the various voicing modes.

At first glance, a Timbre appears to be based upon a dual‑oscillator‑per‑voice architecture, but the powerful Modulation Oscillator offers an audio mode and tracks the keyboard perfectly, so triple‑oscillator patches are never far away. Oscillators 1 and 2 are based upon those used in the Voyager and, in addition to the standard facilities, they offer sync and bidirectional FM for a wide range of clangourous noises. The Modulation Oscillator is also based on the Voyager design but generates a different set of waveforms and can be disconnected from the keyboard for constant modulation rates and drones. You can use it as a modulator in either range and it provides eight simultaneous modulation destinations directly from the control panel. Glide is also provided. It offers linear constant rate, linear constant time and exponential options, and you can direct it to any selection of the three oscillators as you choose.

The audio mixer lies next in the signal path. We often take such devices for granted, assuming that they sum the signals presented to them without imparting any sonic characteristics, but that’s not the case here. The Muse’s mixer is based upon the Moog CP3 module and mixes the outputs from all three oscillators, the osc 1/osc 2 ring modulator (which was inspired by the Moogerfooger MF‑102) and the white‑noise generator. At low levels, it does this without clipping, but you can overload it to distort and fatten the signal. Interestingly, you can create DC offsets in the mixer to cause it to clip more asymmetrically, and this creates a further range of distortions.

Next come two 24dB/octave filters based upon the 904a filter module. The difference between them is that filter 1 has high‑pass and low‑pass modes, whereas filter 2 is dedicated to low‑pass duties. Both offer the expected facilities, but there are just three keyboard tracking options available from the panel, although you can obtain anything ranging from zero to more than 100 percent using the menus. You can use either or both filters as additional oscillators, which means that you can have up to five sources in a Timbre, even before you invoke the ring modulator and noise. There are three signal routing options — serial, stereo and parallel — and you can link the filters so that Filter 2’s cutoff knob adjusts both equally, allowing you to use them as a single notch or band‑pass filter. Unfortunately, the Muse has no audio inputs, so you can’t use its filters as external audio processors.

The Muse’s VCAs are based upon the Moog 902 module, but implemented in stereo so that you can pan voices or Timbres within the stereo field. You can even link the phases of applied LFOs to the pan position, which is a novel idea, and the VCA menus offer a gain offset so that you can create drones and other effects. Following this, the output section offers independent levels for the main audio and headphone outputs, a mute for the main outputs, and a simple 6dB/oct high‑pass filter to remove deep bass that might interfere with other instruments in a mix.

Paralleling the path between the VCAs and the outputs section, the Diffusion Delay is a digital effect unit that, in addition to stereo delays and ping‑pongs, can create diffuse effects that approximate reverberation. But it’s unable to create choruses and other modulation effects because — at least at first sight — it contains no LFOs to sweep the treated signals. Nonetheless, the manual mentions ‘chorused diffusion’, so I asked the chaps at Moog what was going on. They told me, “The character knob is bipolar. If you turn it clockwise from 12 o’clock you wind up with non‑chorused diffusion (ie. the sound is fed through a diffusion network made up of 24 individual delay lines), and if you turn it counter‑clockwise you obtain chorused diffusion where six LFOs are round‑robin allocated to modulate the delay times of the 24 delay lines.” These LFOs could in principle offer all manner of opportunities for choruses, flangers and ensemble effects and, when I asked about this, they added, “We will be allowing the delay to be modulated in the modulation map soon. We will have a post‑launch update to address this before long.” That sounds promising. Other settings allow you to adjust the nature of the diffusion, select which Timbres are treated (or not) and whether the resulting effect is sent to the main outputs, the headphones, or both. If you bypass the Diffusion Delay, the audio signal remains in the analogue realm all the way from the oscillators to the outside world.

Modulation

At first sight, the Muse’s digitally generated contours appear simple, with a dedicated ADSR (whose default response is modelled on the Moog 911 module) directed to the filter cutoff frequencies and another directed to the VCA gains. However, for each contour, you can choose the shape of the curve for each of the attack, decay and release stages, the amount and curve of the velocity response, and the trigger type in mono and unison modes. In addition, you can determine whether new contours are initiated from zero or from the current value when you press the next key that uses a given voice, and you can loop contours to create additional cyclic modulators. With the attack and decay set at or near zero, the loop frequency strays into the audio range, which means that your next sci‑fi soundtrack need never be far away.

The two digitally generated LFOs offer three standard waveforms, sample & hold, plus your choice of one from 11 complex shapes to insert into a fifth slot. They can be unipolar or bipolar, synchronised to clock, key‑sync’ed and applied globally or per‑voice. Furthermore, you can determine their ranges individually on a per‑patch basis with a maximum of up to 1kHz, which means that yet more AM and FM effects are available, and you can choose different modulation depths for each destination. In addition, there’s a dedicated per‑voice pitch LFO optimised for vibrato that can also be used as a ‘one‑shot’ AD contour generator. From the panel, you can assign this to the oscillators, the modulation oscillator and, surprisingly, to the amount of detune, which creates a pleasing ensemble effect.

All of these modulators can be used as sources within the modulation matrix. This offers 16 slots per Timbre that you can populate with your choice from 17 sources (some of which offer both unipolar and bipolar modes), 17 modulation controllers, 14 mathematical transform functions that can affect both the source or the controller or both, and 50 destinations. You can direct multiple sources to a single destination, and a single source to multiple destinations, which is always a good thing. There are six physical controllers that you can use as both sources and controllers — keyboard CV, velocity, aftertouch, the mod wheel, the expression pedal and a macro knob — but there’s no ribbon or touch pad, which is a shame.

Many polysynths have a chord generator that allows it to learn a chord and then play it up and down the keyboard, but the Muse goes further, allowing you to save a different chord under each note.

Chords, Arpeggios & Sequences

Many polysynths have a chord generator that allows it to learn a chord and then play it up and down the keyboard, but the Muse goes further, allowing you to save a different chord under each note. Philosophically, this is the same principle as the pads on (say) a Korg OASYS when used in the same fashion, but with 61 chord triggers rather than eight, so you can record what you need for some pretty extensive pieces. You can even use the arpeggiator to play the chords in sequence, which brings us neatly to...

...the Arpeggiator. This is monophonic (or, rather, single‑key) and saved on a per‑patch basis. From the panel, it appears to be a simple affair with three modes, a maximum four‑octave range, a forward/backward switch, and latching. But when you enter its menu, you’ll find that you can also create patterns of up to 64 steps. You can insert rests, and even choose whether any given step might randomly become a rest. If it does, you can then decide whether the Muse plays no note on that step or skips to the next. There are controls for looping, octave shifts, pivoting (whether the first and last notes are repeated when the arpeggio changes direction), mirroring lengths of random sequences, alternating patterns, excluding repeated notes, and leapfrogging steps with the forward and backward lengths set to different values. If all of this sounds a bit complex, don’t worry... it is. Many of the parameters interact with one another, leading to unexpected results that may or may not be useful. But if you want to find the gold amongst the mud, you can record the arpeggiator’s output into the Muse’s sequencer to edit and develop it.

The small triangular buttons in each section are shortcuts to the relevant menu pages.

Ah yes, the sequencer. This starts out looking fairly conventional, allowing you to determine the length and then enter notes step‑by‑step or play into it while it’s running. You can also use the Value encoder to enter notes (Moog call this ‘Gesturing’) and things such as velocities and gate lengths. Overdubbing is possible, and you can extend notes across multiple steps if desired. In addition, you can record up to eight control panel values on each step so that, when you replay the sequence, each note or chord might have an individual filter setting, or use a different waveform, or experience different amounts of modulation... or whatever. It offers extensive editing capabilities so that you can fine‑tune your sequences. One such facility — and this came as a surprise — is the ability to shift the timing of each note to divisions between the steps. There are 23 subdivisions between every 16th note, and using these creates far more ‘human’ sequences. But then the really complex stuff starts. In addition to easily understood controls for direction, swing, scale, further quantisation, clock source and division options, and determining which note plays which Timbre, there are numerous parameters that use random probabilities to add fluctuations to the sequence. These can range from a single randomly changing step, to changes throughout the sequence, to complete uncorrelated mayhem. Then there’s Coin Toss. This adds another set of lanes to the sequence, and you can then determine the probabilities that the sequence will play the notes from the main sequence or from the second! Having created a sequence, you can then save it to one of the 256 available locations (16 banks of 16) and, if you wish, lock the current patch to it so that the appropriate sound is loaded when you recall it. You can also create one ‘Sequence Chain’ in each of the 16 banks. This can contain up to eight sequences for a maximum of 512 steps.

Finally, the dedicated Clock section allows you to determine the clock source — internal (with tap tempo), external analogue, or MIDI Clock over 5‑pin or USB — and whether the arpeggiator and sequencer respond to MIDI Start/Stop messages. Interestingly, you can program a MIDI Clock offset of up to 50bpm in either direction. There must be a reason why Moog did this, but I have no idea why you would want the Muse running at (say) 103bpm when fed a 100bpm clock! You can also choose whether a clock pulse is output each time that the arpeggiator produces a note. Given that you can program rests in your arpeggios, this means that you can cause external equipment to mirror more sophisticated patterns than might otherwise be possible.

Moog Magic

Unreleased synths often require firmware updates during a review because they tend to evolve as launch day approaches. So, upon receiving the Muse, I contacted Moog to ask whether the installed version was still current. To my surprise it was, although the engineers told me that there is a new one in preparation that should be ready for the launch which, at the time of writing, is still a couple of weeks away.

Having allowed the Muse to warm up, I wasn’t surprised to find that it had arrived out of tune, so I ran the Quick Tune procedure. Things were still a little out of kilter, so I initiated the full tuning and calibration routine. (Don’t do this immediately before a gig — it takes almost an hour.) Everything was then fine and although further Quick Tunes were necessary after switching on and during long sessions, they were sufficient to keep everything in order.

The Muse is a substantial instrument, measuring 990 x 420 x 110mm and weighing 14.55kg.

I was pleased to find the pitch‑bend and modulation wheels in their rightful places, and I liked the control panel, which has the right amount of variation and colour to break up what might otherwise have been a bit impenetrable. The chaps at Moog also pointed me toward a parameter in the menus called Show Param Changes, which allows you to see the saved and current values of the parameter that you’re editing, and this proved to be very helpful. But what I didn’t like was the keybed — or, to be more precise, its response to aftertouch, which felt jerky and abrupt. So, whenever I programmed a destination for this, I added Slew as the Function Source in the Mod Map to smooth the response and make it more musical.

It was straightforward to obtain the forceful and full‑bodied sounds that we associate with the large analogue polysynths of the late ’70s and early ’80s.

I created some typical analogue sounds using single Timbres: brasses, solo and ensemble strings, pads, pipe organs, basses, leads, and even a few Clavinets. As I had expected, it was straightforward to obtain the forceful and full‑bodied sounds that we associate with the large analogue polysynths of the late ’70s and early ’80s. This shouldn’t surprise anybody; the Muse is, after all, a Moog based in large part on vintage Moog technology. But, even with the gains in the mixer backed right off, it’s not what I would call naturally smooth or silky, and it didn’t take kindly to being asked to do things that it would rather not.

To investigate this further, I viewed its output on a signal analyser. With no overdrive, the filters wide open and the VCAs set to moderate levels, the raw waveforms proved to be spikey no matter what the mixer levels were. What’s more, when listening to them through high‑quality speakers and expensive headphones, the tone was audibly crunchy. Using the Diffusion Delay and — given the paucity of onboard effects — hooking the Muse up to some external effects units smoothed things out and increased the range of sounds that I could obtain. At one point, I stumbled upon a family of ethereal pads that I’ve loved since the 1970s, but the Muse is far from all‑encompassing, and there are some sounds that you’ll obtain more easily elsewhere. Don’t misunderstand me... there’s nothing wrong with that. Synths have characters, and it would be daft to want the Muse to sound exactly like a Prophet, or an OB‑something, or a PolyBrute, or whatever. Just make sure that it has the character that you want.

Since every Muse patch contains two Timbres, I moved on to creating some splits, and this was where I encountered my first serious concern. If you import existing Timbres into a split patch, it’s likely that you’ll want to be able to shift their octaves to play them at sensible pitches. I could find no way to do this, so I contacted Moog’s engineers again to ask whether I was being a numpty. I wasn’t. They told me, “This feature is not present in the current firmware but will be addressed in coming updates. For now, it would need to be handled via the oscillator controls, or through Mod Map routings.” I didn’t bother to point out that shifting the oscillators up or down by an octave or two is very different from shifting the whole Timbre up or down by the same amount.

Next, I created some patches with Timbre A directed 100 percent to the left output and Timbre B directed 100 percent to the right, and played them as independent synths, even applying the left‑hand delay to Timbre A and the right‑hand delay to Timbre B. Sure, this means that you can’t use the panning features in the VCA section, nor can you use the Diffusion Delay to spread the results across the stereo field, but I found this to be a more than acceptable compromise, especially since I was able to apply separate external effects to each output.

I also layered the Timbres to obtain some composite sounds. These often sounded excellent, but bear in mind that this reduces the polyphony to just four voices.

It was now time to take a deep breath before diving into the complexities of the arpeggiator and sequencer. To be honest, it would have taken far longer than a pre‑release review to explore the possibilities that they offer. Given all of the Muse’s connectivity options (and, in particular, its use of 3.5mm connectors) I suspect that Moog are encouraging us to see it as the centrepiece of an experimental electronic music studio and, if used in this way, I think that many of their features will make a great deal of sense. However, it was while I was combining the arpeggiator and the Chord mode that I discovered another serious shortcoming. I spent half an evening creating a complex setup with chords distributed right across the keyboard and then looked for a way to save this. I couldn’t find one, so I hoped that the map would be retained as a global setup. But when I power cycled, all of my work was lost. This renders the chord/key capability almost useless, so I again contacted the chaps at Moog. To my great relief they told me that, “The chord map will be non‑volatile with eight available maps you can store, but that’s not currently implemented. We’re working on it.” Phew!

Next, I hooked the Muse up to my MIDI studio to see whether it played nicely with the other children.

It did. Its MIDI specification includes an extensive range of MIDI CCs that allow you to control other equipment as well as automate the Muse itself, and there’s even a Multi mode that allows you to control the two Timbres in each patch using separate MIDI channels. You can also choose different MIDI output channels for the keyboard and MIDI CCs (on the one hand) and the pitch‑bend and modulation wheels (on the other). That’s odd, but interesting.

Finally, I hooked it up to my MacBook Pro and powered up the synth in its Disk mode, whereupon it appeared as an external drive on the Mac and gave me access to all of its 256 patch and 256 sequence memories. I could now move and rename them, build new banks and back up the results. Power cycling then returned the Muse to its usual operation with all of the changes made.

Final Musings

The Muse will evolve further after I’ve submitted this review, but some things are already clear: it was stable and well‑behaved throughout the review period, it’s pleasant to use, it has that big Moog‑y character that many of us love, it’s superb at generating some sounds but less so at others, and it will take time to learn how to use it to its full potential. Of course, no synth is perfect, but none of the shortcomings I’ve mentioned are dealbreakers, especially since one or two of them might be history by the time that you read this.

So, finally, we reach the issue of its price, and I think that there are two ways to look at this. On the one hand, it is expensive. Although £3000$3500 is half the price of the equivalent Moog One, nobody in 2024 would blame you for wanting more voices, poly aftertouch and even MPE for this much money. On the other hand, it’s cheap. For £3000$3500, you’re getting the genuine Moog sound for £375$438 per voice, which is a fraction of what it would cost if you were to try to obtain it by other methods.

Would I buy one? I would have to think about it. Would I use one? Of course I would.