I’ve never used a sampler in 35 years and counting, except to mangle the sound to absolutely unrecognisable and a little bit more on top of that; I want my synths to sound like synths, just like no one perhaps expect your Steinway to suddenly start playing a monophonic 2600. Ah, um, sorry, I’m lying to you: I once owned a Mellotron VSTi. My comparison, naturally, will touch only tangentially on the sample oscillators.
I have the fewest experience of the three with Arturia, so I’ll start with it. I have no idea how to compare a collection of pre-built instruments with a single powerful synthesizer. You can build instruments with Halion 7 (henceforth H.) and Falcon 3 (F.), but you can only play with what you pull out of the bottomless box of the Arturia collection. Whichever is your thing, chalk it up as a positive or negative.
Arturia V Moog Modular is totally moogsy-modularsy; I feel that they have done an excellent analogue modelling job. If you have a little more than infinite patience, you can any sound that plays in your head, just what Dr. Moog prescribed. If the patience of those who do honest to the last resistor analogue modelling, let it be a solace for you. I did. I know. Standing on one leg for a week is perhaps easier. Which shortcuts did they take, I have no idea, but it sounds moogsy, if you know what I mean.
IMO, Korg 2600 is better sounding and easier to use. But when it comes to the comparison of 2600s, it’s inevitably IMO. Whoever wants to chime in with an objective argument, by all means do!
The closest to H. and F. comes Arturia Pigments, which is also a synth builder machine. IMO, IMO, and again IMO, it manages to suck and blow at the same time when situated within a direct line of sight of either H. and F. Not being a vacuum cleaner, this puts it on a lower tier than either of the two. But since they threw into the full collection, nice. You’re compelled to neither use nor ignore it. If it makes that sound that buzzes in your head, why not?
For the rest, I’ll compare H. and F., as I’m familiar with them. Let’s try to chalk up [non-]differences, strengths and weaknesses of each.
Both are multipart synth construction machines. Personally, I rarely use multitimbrality, only when comparing A/B/C/D/E… sounds: it’s easy to multiply a part, tweak each and then change MIDI channel in Cubase to select the final version. H. has 64 programs and up to 64 mono outputs, and one 5.1. F. has 32 program slots, 32 stereo outputs and no 5.1. I’m not sure if it’s capable of 5.1-synthesis, and how important it is. Otherwise, both have more program slots than you’ll ever need anyway.
A minor difference that will be important is that H.'s program tree looks like Program > Keygroup {… > Keygroup } > Zone, while F. uses Program > Zone > {Heaped Keygroups}. Both approaches are sensible. The largest difference between the two is the granularity of their components. F. is a Lego. When you drop an oscillator onto Keygroup (the leaf of the tree!), it receives an ADSR envelope by default, and that’s it. If you want a filter—and you perhaps do!—you must add it explicitly, in the chain of effects. In either of the two, you add effects into a bus associated with the Zone, Keygroup or Program. In F., you don’t get free 2 LFOs and other modulations; you add them by hand. This is clumsy; I’d rather delete what I don’t want, my time isn’t free. H.:0.5, F.:0. No biggie at first sight, but here we get the first difference in scoring:
H.:1, F.:1. Falcon effects are controlled by the modulation effects; in H., the only way to control them is by assigning a CC to the source and using it to control the destination. However, F. has its own idiosyncrasy: not all modulations modulate everything! For example, an ADSR envelope at the zone level does not modulate the filter’s drive knob. Layer-level envelope does. Just enough to get crazy before you figure it out. Let’s adjust it to H:1, F.:0.5. This makes zero sense.
Next, filters. H. has 3 or 4 (IDR). But these are good filters. F. has a lot, but most of them just meh. I’m a BsEE, so I know what “Sallen & Key filter” is; you, a sound designer, rather don’t have to, you want to hear it and say “yup, that’s what I want” (Sallen & Key topology is quite neutral: that was the design idea behind it, not musical at all). Quantity doesn’t mean quality. H.:2, F.:0.5.
Effects. Oooh, Falcon’s Thorus! Oooh! I want it! It’s not The Real Thing, it’s The Surreal Thing! +1.5. H.:2, F.:3.
There are a few minor but nice effects that help make the sound more analogue, if that’s your goal, or more this, or more that. For example, smooth random generator: not S/H, but rather a slow random walk, attracted to the centre. I made the same effect using H.'s Step Modulator with smoothness set to eleven, and randomly assigned positive and negative steps. But it’s not random. But you can’t tell anyway. +0.25. H:2., F:3.25.
I would say that the rest of the effects are on par between the two. F. has a nice optocoupled compressor/limiter (a switch), but as soon as you turn on tube emulation (another switch), each sounding voice eats 10% CPU. H., on the other end, is very mindful about performance. A draw.
F. is more stable than H. I have a large screen, and have an open (for months) a bug related to a crash on a large pixel count screen. F. has yet to crash on me. H:2., F:4.25.
H. has 8-op FM synth, F. 4-op. Not a big deal most of the time, until it is. Let’s give H. its due. H:2.25., F:4.25.
F. has an additive oscillator, H. doesn’t. For the life of me, I couldn’t make a similar sound using any of the H.'s oscillators. So, big deal. F+1. H:2.25., F:5.25.
H.'s wavetable is a Wavetable. No, it’s a WAVETABLE!!! All you can do in F. is to have prepared a file of equal-sized wavecells in an external editor and load it. Fugget a formant filters per cell. Fugget a filter envelope per cell. (Maybe, H.'s manipulation with the complex amplitudes is a bit of an overkill; I could never make any good out of it, but I made a singing robot out of 5, five, yes FIVE wavecells in H. Try that in F—no, don’t, that was a sarcastic remark) Compared to H., F. has no wavetable. H.+1.5, wavetable is not such an exotic thing to not do it right. Or usable. H:3.75., F:5.25.
H. has 8 assignable QC knobs per program; F. has no (reasonable) limit. I like how it’s done: you right-click on anything and make it a quick control, and it pops up somewhere on the Macro page for the program. All the SVG mucking optional. You may simply arrange them into a grid or whatever on the default background. +0.5 at the very least. H:3.75., F:5.75.
Granular synth is a tad more powerful in H., +0.5. There are a couple of other sound sources in F.: plucked string, bowed string and noise factures. Well, the former two don’t sound like what it says on the tin (IMO, IMO, IMO), and I can make noisy factures in a dozen ways on the H., analogue, granular or wavetable source. Meh. IMO, of course.
I already mentioned that I never use sampled instruments. It is said that F. has IRCAM sample oscillators of an exceptional quality but are on the CPU hogging side. But maybe H. does no worse, I just dunno.
As for the libraries, to my taste, they are way, way, way far too modern in both synth. I may peek how the sound is made, but when I make my own, my recipe does not include 1/3 Habanero peppers, 1/3 Agostino bitters and 1/3 sugar; I prefer something more palatable. So I leave the comparison of the libraries and samplers to someone who understands these matters more than I do.
Usability in F. is not well-thought at all, or maybe I’m just dumb (or both). I could not find a simpler way to move an LFO from a zone to program level than
- Copy the LFO via a hamburger menu (2 clicks);
- Create an identical LFO where I want it (many clicks);
- Paste copied LFO data to the new one;
- Delete the original;
- Connect the new LFO to the knob (one but extremely awkward drag gesture).
And the thing shows the tree not unlike H.! Usability is a thing. This and other horrors, I accumulated at least -1 of gripes only on F.'s usability side. H:3.75., F:4.75.
As you see, both instruments are excellent, both can make a lot of fantastic sound—with blood, toil, sweat and tears invested, but sound design is a work that’s not supposed to be easy. My final scoring for what I do with them and what I hit my poor head against in the process; they’re fairly close, when converted to a single number. But really they are two different synths, as, say, Buchla and Moog are two different modular synths. The East Coast and The West Coast are both coasts, but that’s not even a point to make. It’s all down to nuances, IMO, IMO, IMO.
And is Arturia Collection your thing? It is if it is; otherwise, it isn’t. Sorry, but this is the only piece of advice that I can comfortably share. All three should have an evaluation period, you may usually call and ask during evaluation, and many makers even extend it if you explain that you need a bit more time to decide. But the decision is yours in the end. Good luck!
