Mr Oktalski (my graphic designer) is a very smart guy. So I tend to take it very seriously when he makes an observation.
He’s been working his way through the controls, of which there are many, and posed the question: which hardware compressors will this model well (i.e. indistinguishably from the hardware). Immediately I can list a few classics that have influenced the DSP design that I can foresee no problems with emulating. Fun list of presets.
But this wasn’t the focus.
The original idea for this compressor was to create “the ultimate compressor” – something you could insert on any track, and have the flexibility to always get the job done without resort to another compressor.
As you’re aware, most compressors have about five controls; maybe a sidechain EQ, and occasionally a “processing mode” option, or a few other character tweaks. This compressor has tonnes of extra controls to allow you to configure it to your exact taste. You know- changing the effect it has on transients, that kind of thing.
I think you all largely know where I come from with dynamics; you probably remember I did Forte; I was, shall we say, “heavily involved in” the design of the Liquid stuff; I built TBK3; and brought the Sonalksis 315 from Mk1 to Mk2. All fairly modern designs.
So now I’m left with a dilemma which is this:
I hadn’t originally intended this compressor to turn into a replacement for all the modelling compressors out there, but Krzys pointed out that with all this technology, it might be a wasted opportunity not to do that.
Which is a design dilemma, because right now, it’s a fucking fantastic unit for a mixing/tracking/mastering engineer who wants a fantastic compressor/limiter/expander/gate whatever for a track… but the focus is very much on getting the job done, rather than being able to swap in SPECIFIC bits of kit.
Opto characteristics; check. Feedback behaviour (which can actually be modelled surprisingly well with a feed-forward design, the maths reveals); check. All kinds of oddities that one finds in an analogue model (knee types, over-easy, unusual transient shaping, bandwidth limits, sidechain biases, weird hystheretic behaviour, etc); check.
Certainly you’d need to know which controls to tweak to get these behaviours to come out, and that’s very much my job in the preset design stage… but perhaps a list of specific pieces of kit would make for a more useful preset list than task-focussed presets?
… and if I do go down that route, how far do I go? I can add in special modes for modelling things like overdamped feedback designs (some very obscure bits of kit ‘wobble’ when they attack. It’s weird but fun), really odd release behaviour (autorelease catches a lot of this in practice, but I could go deep into it).
EQuality was something of a free ride for me in terms of replacing hardware, because (neglecting distortion, which is fairly easy with EQs that aren’t being driven absurdly hard) all hardware EQs basically follow the same rules. In fact the maths forces them to, which is why you can get EQuality to sound like any random EQ you fancy.
Compressors aren’t so well-defined, nor so constrained mathematically.
So now I’m left wondering where the line is. Do people want another (better) Liquid Compressor? Or do people want a fast, efficient workflow tool, that gives them “just enough” control to make sure they can always get the job done?
I can add options all day long; I really can. I kinda like it even. But I’m painfully aware that every control I add makes someone’s life a little bit worse, because there’s more clutter that has to be on screen. Even with the smartest UI design (which Krzys will do, because he’s an evil genius) if the controls are there, you can’t escape them.
Advice please people 🙂
TL;DR: New compressor: Every compressor ever, or every problem solved?