Here’s a free sound bank with 16 patches – sound demo badly played, but hopefully indicative enough.
Inside pics, circuit descriptions, and patching tricks
Better pics coming soon – I didn‘t want to fiddle too much with my Muse yet but understand that not everything is as detailed as it could be!
Stacked Voice Cards
Summing and Output Section
Voice Card Back: VCOS and Mixer
Voice Card Front: Filters and VCA
Stacked voice cards from behind
View from your Manhattan loft down onto the midday traffic (or something…)
Circuit Description
Sources: Moog Voyager schematics; Moog 904A schematic; Moog 902 VCA schematic; MW Moog Matriarch Mod thread.
NB this is merely from what I see in relation to the Moog Voyager schematics and findings by Tony Allgood and others about the Moog Grandmother and Matriarch. I’ve not probed or measured anything, so this is grain of salt stuff.
The three VCOs (yes, the MOD OSC is also a fully fledged VCO) are based on those of the Voyager, while the VCFs are related to the 904A and the VCAs to the 902 (hello, Grandmother and Matriarch!).
Each VCO has an expo converter (transistor pair + opa, and, very well visible orange tempcos). The integrator is based around an IC switch rather than a Fet transistor. In contrast to the Voyager, you cannot continuously blend the wave forms, and accordigly, the respective circuitry is not used here. As far as I can tell, Pulse/Square and Tri generation is like on later revisions of the Model D (the so-called New Osc Board), as featured on Grandmother and Matriarch as well. Pulse is built around a 393 comparator IC and TRI seems to be done like on the Mini by means of two npn transistors and an amplifier.
In the following you see pulse/tri shaping…
The Mod OSC routing and amount control is done per SSI2164 VCAs and digital switch ICs, as is the FM (xmod) amount between OSC1 and OSC2. While Waveform selection from MOD OSC is done via IC switches (noise, I suspect to be digitally produced), OSC1 and OSC2 use one 2164 each for blending between Saw and Tri as well as between these and Pulse. This leaves two 2164 VCA cells on this side of the board, which, I assume, are used for the ring modulator and probably also mixer overload.
On the very right hand side of the board you see a LM13700, and a couple of dual transistor packages in addition to an OPA with a little blue tempco. Judging from the Voyager filter schematics, this part of the circuitry is used to control the filter frequencies.
Let’s flip the board and check out the filters.
You can easily spot the two ladders that make the core of your two filters. There are more IC switches (for setting the serial/stereo/parallel configuration, as well as for switching VCF1 to high pass) and there’s another 2164, two VCAs of which are for resonance control (no clue yet about the remaining two VCAs on that IC).
As mentioned, this filter circuit is also used in the Grandmother and Matriarch, and, whatever you make of it, also shares a kind of flaw that’s prominent in these. Keep reso at zero and pull up cutoff, and you will hear quite clearly a passband drop as soon as cutoff exceeds 12khz:
Here’s another one wirth cutoff low-ish but filter env sweeping in from above:
Tony Allgood, in his excellent Moog Matriarch Modifications thread, analysed this and found a cure, as it were, for the Matriarch. Long story short: all Moog filter have this a wee bit but this one has it very strongly – probably on account of taming resonance at higher cutoff rates. Following the Moog Matriarch analysis, I think I found the corresponding troublemakers on the Muse voice cards, but please don’t fiddle with that w/o further investigation/confirmation. Alternatively, just accept that thing for what it is. Mind you, there are neat mod-matrix-based workarounds if you are concerned about this (see below)
Moving on to the remaining part of the signal path: The two main VCAs are transistor based and look like derivations of the 902A to me (I think these are usedin the Grandmother and Matriarch too). Voice summing etc is then done on the main PCB, but I’ll leave that for now.
Snapulator Sound Design
In relation to the passband drop aspect of the filter, but also more generally in a “can this Moog replace other Moogs” sense, some people wondered whether the Muse can sound snappy. I think it can. Here’s a semi-tame bass in mono mode, for instance that’s not half bad for starters, I think:
Yet, since this synth has a more than decent mod matrix, you can, of course also do nasty stuff 🙂 In the following example, I blend in mixer overload in proportion to how much envelope amount is routed to filter cutoff (first pass with volume amount, second pass just env to cutoff):
Another neat trick is to have your timbre volume (i.e. the final volume of your sound) modulated by a little snappy env. In this example I set timbre vol to ca 80% and have filter ADSR modulate the final volume – to add some spice, I have the same filter ADSR modulate the amount of volume modulation (again, this is an A/B example, this time first without volume modulation, then with):
An two more sounds using the same ADSR to timbre volume technique. Th first saw bass sound also utilises the Key reset function of the mod osc to highlight snap even further.
These examples are fairly moderate but I hope to have shown that there are easy ways to make the muse not only sing but also snap slap.
Moog Muse and Prophet 5 Rev 5 Module Comparison
Well, these comparisons are very limited in terms of like, everything, but I have them next to each other and was curious if they could get close in terms of basic sounds. Here’s a stereo WAV file: for best listening experience download it and split L/R in a DAW or audio editor.