Minimoog Softsynth Shootout: Diva MiniV3 Monark Legend Minimonsta vs Model D

[SMH]

Yeah, polyphonic synths have unique and separate monophonic synths for each voice including filters. The only thing they usually share between voices is LFO but some even have separate LFOs which is useful if they can be re started when the note is hit to make echo like effects. Easy for software to do, expensive for analog as it's like 6 or 8 entire mono synths -- though they all share the same controls.

[klinik]

What's the difference to "normal" polyphonic synths with oscillators, filters and envelopes per voice?

The FVS or eight voice "main" controls in the polyphonic programmer are just offsets to the setting in each SEM panel and all those can be different.
That combined with the different voice allocation modes makes the Obies something very special (apart of the sound of course!)
I had one for many years and still miss it sometimes... The Arturia SEM have some of that functionallity, but in reverse: the "main" control affects all "voices" and you can offset some parameters for each voice. Also the sound is nowhere close to the original. A four or eight voice system emu would be a dream poly synth... just after repro-5!!

[Urs]

@klinik Hehehe, I was aware of the idea of separate individual SEM units in a 2/4/8 voice, I was merely referring to the idea that "synths with individual envs and filters per voice" would cost more CPU. This is (by my definition) a very common thing, i.e. no big deal at all.

Even having multiple separate synths won't change that, the only difference being separate sets of parameters.

[ENV1]

Would it excessively bring up CPU usage, if you recreated one filter and envelope per voice? Probably yes, but, thought i'd aks anyway.

Id say it depends on how the synth was made.

Take Rapture. It has 6 so-called Elements (E1-E6) which are more or less 6 separate 1-osc synthesizers with internal voice-multiplication capability up to 9x. (3v, 5v, 7v, 9v.) Each Element has 7 dedicated EGs, 7 dedicated LFOs, 7 dedicated step sequencers à 128 steps, 1 dedicated Scala Tuning module, 6 dedicated keytrack modules, plus a variable-order DSP array consisting of 2 user-selectable filters and 3 more slots for distortion/bitred/etc. Each Element further has 3 single-band EQs plus an FX slot for various effects, many of which can be modulated by another dedicated LFO and/or filtered by another dedicated filter. And finally on the Global page there are 3 more FX slots, 3 more EQs and 2 more dedicated step sequencers for L/R panning.

Thats a lot of wood as it is. But: Each of the 6 Elements can be set from 1 voice to inf., which means in theory (of course your CPU will impose a limit here at some point) Rapture can provide all of the above @ infinite polyphony. Yet Rapture is one of the most low-CPU synths ive seen. You have to really push it HARD before it even registers on the CPU. With all 6x9 oscillators going at once plus a chorus module on each Element my CPU alternates leisurely between 1% and 2%, which is pretty much 'background radiation' around here. (I.e. what the OS uses when idling.)

So like i said. In the end it all depends on how the synth was made.

[egbert101]

<delete>

[Richard_Synapse]

You probably know this, but for others, the eight-voice SEM was truly polyphonic, with a filter and envelope per voice. I think that might be too much to do in sofware. Arturia's SEM does 32 poly which is pretty nice.

Actually eight-voice is quite nice CPU-wise, we can use AVX to work on all voices in parallel (this is what we're doing already with SSE in Legend, same principle with four voices). Or do you mean the interface would be no good in software if all parameters are cloned 8 times?

Richard

[AnX]

Isnt that what Dune2 does already (and it has an obie filter...?)

[egbert101]

<delete>

[egbert101]

<delete>

[Mutant]

As for interface hmm. It would be no problem for me if all voices were clones, but some might find it difficult to create presets. Perhaps a feature like copy/paste to other units, or some other workaround might be cool. Perhaps assigning how many units you want to link together as you change parameters? It is an interesting challenge.

The fact that none of the real hardware modules can be set exactly as any other module is one of the main things that sets this synth apart from almost everything else.
Omitting this feature would be like not including feedback in a Minimoog emulation.

[egbert101]

<delete>

[Richard_Synapse]

I kinda agree, if we really want that level of authenticity, although again, it's a matter of whether it can be done or not, retaining the kind of quality in emulations such as Repro-1 or The Legend. Also, it is going to be a pain in the arse to create presets unless some clever ideas are used to make it easier.

It could be done like in DUNE 2, where by default you edit all parameters, but then you also have the option to edit up to 8 individual parts as you please (and there's copy+paste too). For sound design this concept has proven to work well

Richard

[egbert101]

<delete>

[Ingonator]

You probably know this, but for others, the eight-voice SEM was truly polyphonic, with a filter and envelope per voice. I think that might be too much to do in sofware. Arturia's SEM does 32 poly which is pretty nice.

Actually eight-voice is quite nice CPU-wise, we can use AVX to work on all voices in parallel (this is what we're doing already with SSE in Legend, same principle with four voices). Or do you mean the interface would be no good in software if all parameters are cloned 8 times?

Richard

What I meant is that it would be a different filter and envelope per oscillator (or pair of oscillators in the SEM), rather than a single filter for all voices. If that's easily done, then all good. It's a bit like having eight monophonic synthesizers all joined together. So each voice is a separate synthesizer or SEM unit.

Arturia's version only does polyphony through a single filter for all voices, which is good enough (for me anyway), but not 100% authentic.

A single filter for all voices is a bit misleading as that describes a "paraphonic" synth. It would be more precise to tell that all voices got the same filter settings while they still use an own filter per voice (like "voice cards" in polyphonic syntsh as the OB-X that each contain the synth engine for one voice).
Even my Ensoniq ESQ-1 and Korg DW-8000 while using digital OSCs do have 8 filter chips included with one for each voice (those chips also seem to include an analog VCA).

As mentioend earlier In Arturia SEM V2 you also got the "8 voice programmer" (in advanced settings) to set up to 6 parameters (selectable from a drop-down list) differently for 8 voices. This also includes filter settings like e.g. Cutoff and Resonance. Different voice cycling modes are available too.
The original 8 Voice synth had a comparable feature to "override" certain parameters of the single SEM units and set them to similar settings for for all of them.

Arturia SEM with the mod matrix even allows modulation of the morphable filter modes (LP to Notch and then to HPF). The option of a morphable filter mode is quite unique with the SEM fiter and also included in the new OB-6 from Tom Oberheim and Dave Smith. The "Uhbie" filter mode/module in U-He Diva has that option too.
My Waldorf Pulse 2 and Novation Bass Station II hardware synths got analog multimode filters too but they do not have morphable modes.


In the original 8 voice having differnt soundns for each voice might be useful in Unison mode but with a plugin you could do the same using multiuple instances. With the plugin you coudl even payer polyphonic sounds while the Unison mode in the real thing was monophonic.
Personally i had used teh 8 voice priogrammer in SEM V2 mostly to use different fine tuning and panning for the voices to make the sound "bigger" when played polyphonic.

[egbert101]

<delete>