ITB vs OTB
- Published: 2020-11-29 12:06
- Updated: 2023-03-27 18:51
Foreword¶
Once upon a time around 1991, a 14-years old me sat in a Techno producers' wet dream: a studio stuffed with all the relevant machines of the day, connected to a nice console at its center.
Back at home, I was only able to sample vinyls into my beloved Amiga, for later use in Protracker. Right off the bat, having experienced the difference of making Techno in its natural habitat vs a piece of software, limited to 4 channels of sample-playback, certain limitations became quite clear at an instant.
And I remember that impression of freedom at home around 98, that went along with the discovery of Buzztracker: a semi-modular tracking environment, stuffed with realtime synths and effects. That was before VST plugins were standard.
Meanwhile, on the hardware side of things, Stereo Jack and I had already produced a number of 12" for Superstition Records at Hamburgs Boogie Park Studios. Which was always equipped with the latest and most sought-after gear.
Despite finally being able to route chains of effects at home the way I was used to from the hardware world, there was always this impression that "something's off" with Buzztracker. Or the way I was using it. Things just didn't "feel" or "behave" as I expected.
And it made me wonder why.
Of course, there were a number of obvious things: the DBX 166s, 1176s and Distressors of the world were made by engineers who understood what they were doing. And forget about ever having a plugin that sounds like a 303.
So, what do you do when it's 1998 and you don't have Google, or YouTube and not even the faintest clue what you're looking for? You start digging on your own.
Here's what I puzzled together so far. Wrapped up for producers, from a Dub/Techno-ish production point of view. Mind you, even though I've had my share in developing a lot of audio-software, I still don't code.
Take this as food for thought. If you're a developer and find some flaws, please connect and correct me. So, where do I even start? Bees and flowers, I suppose.
Physics vs Maths¶
I get that this stuff is kinda boring, but I can't talk about Analog vs Digital without getting a few basic scientific concepts out of the way. In order to make any sense of the more interesting bits, I need to drag you through this.
Physics is a reflection of the practical world. For reasons beyond my comprehension, we're drifting through space on a rock around another rock, that's incredibly hot. And there's a shitload of more rocks around our rock. Some are flying in tandem building gargantous circles. While others aren't.
Some clever folks found out that all of this makes sense. They sat down, and observed, and calculated, and applied logic, and wrote down their trains of thoughts. And when something seemed to be coherent enough, thus agreeable by others: we had a new law of physics.
Sometimes, that process took more than an individuals' lifetime. Sometimes, the theories were ripped apart and replaced by somebody elses'.
My point is: boy, what time does it take to observe, and calculate and verify and think and discuss, to arrive at a point where physicians agree:
that observation of nature/physics is probably correct. At least, until we have a better one. Let's see if it stands the test of time. Kinda funny that one of these theories holding physicians brains together, basically claims that time doesn't exist (and I kinda get it).
What potentially keeps a scientists mind active like a nuclear reactor: instability or chaos. Or: what makes physics physics and nature nature. Because truly stable systems don't evolve. If it wasn't for chaos, none of this would exist in the first place.
To give you a teaser why I think above matters, let me rephrase chaos for musicians: Noise.
And maybe, if it wasn't for chaos, humans would have everything figured out by now. No, wait. sigh
Maths on the other hand, is a theoretical language, describing the observations of the practical world. Or a mean to predict the behaviour of a physical thing-abob. Or, if you're insane/genius, even a way of predicting and calculating the complex, and intertwined nature of the universe. And fuck with everyone's brains, including those of following generations, along the way.
At its core, I'd say maths is mode of thinking. Which couldn't exist on its own without its friend: Logic.
At times, math even seems to defy the gravity of logic in itself. I mean the moments, in which your math teacher didn't have anything else to offer than "that's just the way it is".
Fancy a taste? Here's a fave: "Why does our number system range from 0 to 9?" - You're welcome. Probably because humans have 10 fingers. What if we had 9, or 8 or 13? Yeah, well.
For sanitys sake, let's stop there, and take for granted that it works. Cause it does. And on top of that, I'll boldly claim that math, first and foremost, was build upon observation.
If I throw three medium sized rocks at the head of a predator, he's knocked out. Hm, that's exactly as much as this amount of fingers. So, I better make sure to always have one, two, no: three medium sized rocks in my, erm–yeah, pocket.
Fast forward a few thousand years and we arrive at a point in time, at which, through tedious observations, we know a lot about how things work. At least, through the lens of our own human bubble.
Analog vs Digital¶
To stick to above distinction between physics and maths/logic, or practical vs theoretical means, the biggest difference between analog and digital circuits, or software, lays in their instability or chaos.
Analog circuits deal with the arbitrary behaviour of current, the tolerances of its individual parts. Depending on its design, also including a realm of intertwined, intended or unintended side effects.
Example: run current through a Tube to amplify a signal, and get heat for free. If you put a simple, analog VCO in close proximity, its pitch would go all over the place as the thing is heating up.
Of course, not all side-effects are as dramatic as above. However, as long as you're dealing with analog components, nothing's ever 100% stable. That's simply how things work in in the practical world. (Thank you Joachim)
Digital systems however, are based on observations of physics, or the practical world. Which get wrapped into theoretical constructs of mathmatical algorithms. Emulating the observed behaviour to the degree of what has been observed.
Which then get transcribed into some form of programming language, which then run through an interpreter, so the stupid CPU can have its binary cereals for breakfast.
For the sake of my sanity, I'll simply assume that above process of chinese whispers works and doesn't introduce any artefacts on its own, driving the results even further away from what was intended. Ok? (Yes, I've heard about rounding errors, cough).
My point is: only what has been observed and deemed worthy of being written into code also ends up in code.
Mind you, I've been very close to the development of this 303 emulation since 2002 now. I still have no clue about writing code, yet learned a lot about algorithms on a more topographic level. And only ended up at AudioRealism, out of the the very common belief back in the day:
"A truly authentic 303 emulation in software? Impossible." Thank you for proving me wrong, Mike. <3
What I learned along the way: it takes an utterly absurd level of anal retentiveness when it comes to analyzing even the smallest intricacies of the physical world (or a likewise amount of trial and error).
Moreso, the 303 itself is a great testiment to the quirky nature of the analog realm. And I wish I could Tardis myself back in time to the meeting during which Roland decided:
"Yeah well, that's good enough." ¯_(ツ)_/¯
When it plays an accent, it draws enough power from the clock that the sequencer has a short hickup, its VCO drifts all over the place, it makes strange pffftsch and tschzzt noises depending on its settings. If the batteries are dying, it's going to remix all your pattern for free. On top of that, no two 303s sound exactly the same, despite using the same parts on paper.
Very non-japanese. But what a lovely piece of beauty. As well as proof to the sentient life of analog circuits. I like to think it's the result of not-giving-a-fuck engineering while trying to keep the development and production costs as low as possible.
And the only analogy to digital software that comes to mind:¶
Copy/pasting a number of algorithms from the web, If you want the same amount of weirdness in a piece of software, you first and foremost have to think it. Then code it.
Restricted by the laws applying to digital systems. The accuracy of an emulation is only as good as its observer.
A saturated signal literally occupies more space on the frequency spectrum, and thus indeed is "fatter".
If you'd like to dive deeper:: https://blog.mixanalog.com/plugins-why-analog-still-sounds-better