Learn Music and Audio Production | iZotope Tips and Tutorials
4 Types of Analog Compression—and Why They Matter in a Digital World
Yes, another article on compression! But this one’s a little different. It’s more of a history lesson, with some applicable takeaways to help you in your modern mixing practice.
If you've only operated compressors in plug-in formats, you'd be forgiven for thinking their operation is a matter of simple math—a bunch of binary acting in computerized concert. Many modern GUIs, often displaying the signal as it hits the threshold in a sort of easy-to-read graph, help solidify this perception.
But what about those fancy compressors we come across sometimes? Those emulations of 1176s, LA2As, Fairchilds, DBX 160s, and the like? What about iZotope’s vintage compressor, which seems to behave differently?
You’ve no doubt noticed that these fancy compressors react differently; heck, some of them don’t even have attack and release controls like the earliest compressors. Some of them aren’t even technically compressors; they’re limiters—or even leveling amplifiers!
How about the stock compressor plug-in in a DAW like Logic Pro? It emulates all sorts of compressors! FET? VCA? Opto?
What does it all mean?!?
If you’re new to the game, you’ve probably stumbled around these kinds of compressors, played with some knobs, and hoped for the best. I know I certainly did, so I aimed to educate myself on their differences.
I maintain it’s important to understand these differences, even here, in our digital world, for understanding them helps us make quicker, more efficient, and ultimately better choices.
So we’re going to talk compressors. Specifically, four kinds of analog compressors. We might not have the space, bandwidth, or frankly, the attention span to go into the details of circuitry, but we’re going to leave you with what this writer believes are the essential takeaways; in so doing, we shall aim to transmute this history lesson into a simple goal: to make your next compression decision a more informed one.
This is arguably the most commonly used compressor phenotype in the corporeal universe. It tends to sport all the controls you’re used to seeing (attack, release, threshold, ratio, and sometimes knee). VCA stands for “Voltage Controlled Amplifier,” a type of mechanism found in many musical applications. Here’s one: if you’ve seen VCA groups in Pro Tools and Logic, these get their name from the same technological concept that powers VCA compressors. Namely, a control signal dictates whether or not the level is brought down.
In the case of a VCA fader on an analog console, the control signal (which is you moving the VCA fader) “tells” all of the tracks in the group to diminish in level by an equal amount. In a VCA compressor, however, signal is split through an IC chip (another semiconductor device) into the detector path (which controls the compression effect) and the output path (which is what you hear).
This control path can be manipulated by a host of parameters that make these processors operate in an exacting way: attack, release, threshold, ratio, and knee are often tunable to a tee, making the process more granular than other compressors, as a general rule (there are always exceptions). The result might not be linear—in many cases its logarithmic—but it’s still predictable and dependable in a way you don’t always get from other compressors.
VCA compressors can be found on SSL channel strips, API compressors, and gear from Rupert Neve Designs. Many engineers love them for their predictability and repeatability. You quite often see them on the master bus, on groups of instruments, on guitars, basses, drums, even vocals.
Now, here's where it might get a little confusing: It can technically be said that most compressors employ voltage control amplifiers somewhere in their construction. However, the designation of VCA compressors still carries a specific meaning in the pro-audio world, because in these compressors, the voltage controlled amplifier is housed in an integrated circuit. This integrated circuit (or IC, as it's commonly abbreviated) helps with with the tunable aspects of the compressor mentioned above, and also with keeping unwanted distortion to a minimum.
Recording setup featuring a dual channel tube VCA compressor
This one might just be my favorite—not to use, necessarily, but to describe. Why? Because it depends on light, or more specifically, light-dependent resistors!
But wait, what’s a resistor? To properly get into this stuff requires talking about the nature of electricity. This would result in a heady discussion that would bore the pants off of both of us, so let’s skip the science and go right to a metaphor commonly used to describe resistance—that of water going through a pipe.
Just picture it: water flows through the pipe, and the pipe carries the water where it needs to go.
So far so good, right?
But what if we put a cap on that pipe, one with a few small holes? Yes, water may only trickle through the holes, but the resistance of the water on the other end of that pipe—the pressure of it—has increased as it builds up. Thus, when a pop-science article on resistors states that “if you turn the volume down, you're actually turning up the resistance”, the metaphor helps us to understand why this the case. We now begin to see the function of resistors in the circuit of a compressor—they help put the squeeze on the signal we need to tame.
But how does this impact the sonic characteristics of optical circuits?
In an optical compressor, the resistors are light-dependent: the audio signal feeds a lighting element (such as an LED), which shines upon a light-sensitive resistor. The resistance of this light sensitive element informs the compression circuit how much and how quickly to attenuate the audio signal.
The wrinkle here is that this interplay between the light source and the resistor, while fast, is not instantaneous. Furthermore, different types of light sources illuminate at different speeds, and a resistor can react differently depending on the material from which it is made. For this reason, an optical compressor’s sonic behavior is highly dependant on the types of materials used in its construction.
But here’s a commonality between them, no matter the make: the attack and release of an optical circuit is (at least most of the time) definitely not linear, often involving a bit of delay before the attack kicks in, and additional delay as the release drops off.
For instance, the harder you hit an optical compressor, the quicker its initial release time can be—but the slope back to a normal, uncompressed sound will not fall in a linear fashion. It will “curve.” So if the circuit gives you 10 dB of gain-reduction, the first five decibels might release much more quickly than the following five.
This bit of behavior is something you can hang your hat on when it comes to most plug-in emulations: the specific timing will certainly change depending on emulation, but the attack and release will act in a way that is a) often slower than many other compressors, and b) more meandering as it starts and stops.
The behavior of these time constants can result in a compression that is quite musical and often smooth. In general, vocals, lead lines, and other elements that need an intangible “rounding out” (not so much a hard squash as a general evening out, or a shapely bolstering) can benefit from optical designs and optical emulations. It’s not as functional for transient shaping, though of course there are exceptions to this rule given the wide variety of types of light sources and resistors available.
To me, optical compression is quite poetic, as it involves the communion of light and sound. It elucidates their wave-like commonalities, getting to the core of how they can influence each other.
Explore compression in a digital world
Try Music Production Suite Pro for free and explore iZotope compressors in Ozone Pro and Neutron. Access more than 30 industry-standard plug-ins, production courses, custom presets, and royalty-free sample packs.
If you’ve ever hit all-buttons in on an 1176 (mod or hardware), then you, my friend, are experiencing the glory of FET compression. FET stands for “Field Effect Transistor,” which begs the question…what is a transistor?
I’ll just say this: a transistor is a semiconductor that can both amplify and attenuate signal in accordance with the settings you dial in (the word “transistor” is technically a portmanteau of “transmitter” and “resistor”.)
FET vs. VCA Compression
A lot of people wonder about the difference between FET and VCA compression. Indeed, there's confusion, some claim FET is a subset classification of VCA. However, they serve unique purposes. VCA compression is meant for clean, transparent mixing where you want control over a large range of parameters. FET compresion is better suited for aggressive and heavy sounds like electric guitar and drums. However, there is a fundamental difference in the transistors: In a VCA compressor, the transistor is housed within an integrated circuit (an IC) which responds to the voltage of your incoming signal. The FET, however, works with the electrical field as a whole, and gain changes are the result of electrical charges in addition to voltage.
This is complicated, confusion stuff, to be sure. But know this: The designation between FET and VCA matters, because of the one of the most famous, widely used compressors of all time is an FET compressor—the previously mentioned UREI 1176. FET compressors like the 1176 are capable of exceedingly fast attack times, though not without color. I tend to avoid FET emulations for mastering, but pile them on for guitars and drums.
They often boast a feedback design, which contributes to the program-dependent nature of the compression achieved (for instance, have you ever seen a threshold knob on an 1176? Because I haven’t). We’ll get into feedback/feed-forward designs a little later, after we cover…
Feedback and Feed-Forward
Some VCA compressors, like the API 2500 and the Master Buss Processor from Rupert Neve Designs, have a switchable feedback/feed-forward circuit, which brings us into another discussion—namely, what the heck does feedback or feed-forward even mean?
I confess, it took a lot of gazing at schematics to understand it myself, so don’t worry if you’re confused by the following:
See, analog compressors split signal into two parts, as discussed above—the detector circuit and the ultimate audio path. On a feed-forward compressor, the detector is receiving the same signal that will wind up being effected. Pretty straight forward. (Hey, that’s a pun…)
Not so with a feedback compressor—its circuit is being fed by signal that’s already been through the works of the compressor; in effect, it’s reading an already compressed signal. This results in an arguably “smoother” compression, though still controllable.
What tricked my mind about this was the time machine effect; how could a signal compress itself with an already compressed signal—from the future? If obeying the laws of physics, wouldn’t such compression result in a slower attack time at the very least, one with severely audible clamping? And yet, the 1176 is a feedback compressor capable of a quick, quick attack time.
It turns out this isn’t an issue—your music, in the electrical world, is split instantaneously between the detector path and the audible, output path.
Trademarked by Manley as “Vari-Mu”, and found in the venerable Fairchild, “delta-mu” compressors rely on tubes. Indeed, a re-biased tube becomes the mechanism by which the compressor knows when—and by how much—to reduce the gain.
The term Delta-Mu is by no means the catchall, but I like it, because it involves the Greek letters we use for “change” and “gain” (i.e., gain reduction), and also, Greek is fancy.
How does this circuit work? Let’s skip through the heady science and land on what I believe to be the most important sentence: as the signal feeding these compressors increases, the actual current sent to the grid of their tube decreases, culminating in a reduction in overall level. In other words, the tube is the main engine driving the gain reduction; other compressors—I’m looking at you, optical LA2A—might have tubes in their construction, but this is for color; they do not rely on the tube itself to tame the dynamics.
This is a gross oversimplification of course, and there are many other factors at play. What matters to you is the use-case—the sound of the thing. Words to describe this sort of compression often include “smooth,” “thick,” and “creamy;” this is due, in my estimation, to two factors: the quality of tube-circuitry (i.e., the pleasant distortion of tubes), as well as the programmatic attenuation achieved by this kind of compression (they’re highly reactive to the material feeding them). Such compressors have been known to handle generous amounts of gain-reduction before unwanted artifacts accrue.
These compressors are particularly adept for “gluey” mix busses, as transients tend be handled musically, instead of with an iron fist; whereas VCA and FET compressors can give you a firm, corrective hand, but these are more like a gentle pat on the back.
On a different tack, it’s interesting to note the following: Each compressor—except for VCA—tends to have its poster-child; on these poster-children, one or more controls tend to be notably absent, but the control changes depending on the model. LA2As, LA3As, and many of their clones don’t boast attack/recovery constants, and they are optos. The 1176 doesn’t give you a threshold parameter, and that’s an FET. Neither the Fairchild nor the Vair-Mu have ratio knobs, and they’re delta-mu. VCA compressors are the standout here, and they often boast all of the controls.
Of course, there are exceptions—you’ll see opto compressors with attack/release parameters in both the hardware and the software world, for instance. But if you see a compressor in your DAW that you’re unfamiliar with, and happen to notice it looks like one of those iconic pieces mentioned above, there’s a good chance you can use what’s missing from the control set as a shorthand to tell you what you’re working with.
A coworker of mine recently asked me why I was writing this article. He said all that really mattered is the sound, not the tech. And to some extent, he might be right:
You don’t need to know how photocells work—or which is the common light element in modern optical designs—to understand the typical release characteristics of an optical model, and why it would be good for certain situations. Likewise, you don’t need to know how a variable-mu compressor uses the tube itself; you just need to keep in mind how highly music-dependent this kind of compressor can be.
Still, I maintain that it’s important to know these distinctions, and to understand on a basic level what they are. Why? Because the principles behind these compression-types are everywhere. Not only are they everywhere in the software world—they’re all over hit records.
Many are the engineers who define their vocal sound with a combination of FET and optical compression—or who demand an aggressive VCA compressor on the drum buss for both thwack and glue. These sounds have hallmarks, and we as engineers are paid to get these sounds. Understanding these sounds allows us to achieve them quicker, especially in unfamiliar environments, which can often be part of the job.
In engineering, efficiency is the name of the game as much as anything else. If you’ve got a vocal that you need right in your face, you can play around with compressors to your heart’s content. Or, you can know that an emulation of an FET compressor might put the vocal right where it needs to be, and get you there that much quicker. The choice is yours—and now that you’ve read this article, you get to make that choice with some roughly sketched facts at your disposal. What happens now, of course, is up to you.