In previous articles, we covered different types of hardware compressors and why their behavior matters in this hybrid/digital world. We also covered typical mistakes engineers make in using compression. This article expounds on compression in a specific way: we are diving into the various compressors available in the iZotope ecosystem and relating them to real-world applications.
If you need a brief recap on what compression is, get a refresher here. Now let’s cover some basic reasons to use a compressor before we move on to our use cases:
Why do we compress?
Here are some basic reasons to use a compressor:
To control a signal’s overall dynamics: The signal varies too much between loud and quiet sections, and we want to restrict its dynamic range. Think of a vocal that grows too loud at the end of each phrase.
To shape the front end of a transient: Using the attack and release control, you can really change the flavor of how a compressor hits the first percussive signal in its path. You can make drums smack harder with the attack and release parameters.
To alter the groove or feel of an instrument: You can use a compressor to shape the groove of an instrument in a variety of ways. Sidechain compression is one way of doing this, though you can also shape groove without an external sidechain input, particularly on a looped phrase; you can change the feel of a drum kit or a bass part, for example.
To add color or tonal alteration: Hardware compressors often have a desirable, useful tone to them, as a result of their circuitry. A solid-state compressor for example, might predominantly add even-order harmonics, while a tube compressor could impart more in the flavor of odd-order harmonics. Many digital compressors model the tone of their analog forebearers.
Adding bass character with compression
Sometimes, for the electric bass, you want to beef up the tone without obvious distortion or equalization—but you must be sure to level out the dynamics so one note doesn’t bloom above another, and you want to lock that groove in as much as possible.
Here, I’d recommend Neutron 3 in Vintage mode. Try Neutron 3’s compressor in Vintage mode with a ratio of 4:1 and a medium attack time, then use the release to your advantage. If you want to emphasize each note, go with a quicker release, as the compressor will let the transient pass, and clamp down on the rest of the signal. If you want to fit the bass into the drum’s pocket, you’ll probably want a slower release time, though you must tune the release by ear.
Another variant would be Nectar’s optical circuit. Optical compressors often feel quite musical, because their circuit reacts to the input in a non-linear way. Nectar’s Optical mode aims to react similarly. Its fixed ratio of 4:1 ensures you won’t drive the compressor too hard unless you really lower the threshold far past its best usage. Its attack and release may give you smoother compression, but it may be less easy to predict; so use this one on a bass part whose dynamics are already pretty solid.
Bass control with multiband compression
Sometimes you may want to control various ranges of the electric bass. Perhaps you want to tame its low end more than the midrange or the highs, for example. Here, multiband compression on the bass might do you well—Neutron’s default compressor mode in multiband is a good option here. Try a simple two band split where everything below, say, 80 Hz is affected with a lower threshold. This might help the bass stay out of the way of the drums, without giving you the obvious loss of EQ.
Again, for this usage, I’d recommend Neutron 3 in multiband. It gives you up to three bands. Try setting each band’s attack, release, and ratio to the same settings. Then, only change the thresholds. This might give you a more natural response—one that doesn’t sound obviously carved up.
Compression on percussive hits (snares, drums, and toms, both acoustic and electric):
For the more percussive elements in a drum set—kick, snare, and toms—I either want to shape the transient, add an overall punch to the drum, or reign its in dynamics (if it is remarkably inconsistent).
In any case, I usually turn to Neutron 3 in its Vintage mode, for much the same reasons as described in our first entry: the slight coloration of the 1176, and the ability to use faster time constants. For transient emphasis, I’ll go with a relatively low ratio, threshold to taste, a medium to medium-slow attack, and a faster release.
If I want to add a lot of punch, I’ll take these settings, lower the threshold, max out the ratio and raise the output gain so that it’s loud and thrashing. Then I’ll edge the mix back so it’s operating in parallel.
Sometimes I’ll run two Neutron 3 compressors in series. The first will provide the settings for transient emphasis, and the second will add the settings for punch, inched up in parallel.
Parallel compression is great for adding thwack or punch—but it’s also quite nice for restraining dynamic range in a minimally invasive way. Sometimes we don’t want to alter the attack of a drum hit, but we don’t want it to sound inconsistent either. Here, we can use parallel compression to our advantage. Mark the attack and release times on this kick drum:
This is way too fast for straight implementation—it clamps down on the transient so hard that it can add distortion. But we’re using it in parallel, with a mix of 60 percent. We’ve also turned the output down.
When the kick is played soft, we’re likely to hear the parallel signal more; the compressor is working less hard, as its reacting to a softer note, and furthermore, its tone cuts a bit more over the quiet kick. But when the kick hits hard, the compressor reacts harder—we hear the compressed signal less because it's lower in level, and the original signal cuts through more. In this way, we’ve brought the quieter kicks up, left the louder ones relatively unaffected, and not altered the transient response nearly as much as insert-compression would have.
Compressing overhead mics
Compressing overheads is tricky; you can make the whole cymbal picture seem out of whack if you do it wrong. For this reason, I often avoid compression on the overheads, favoring transient shaping if I want to change the dynamics of the kit.
However, compression on overheads can be necessary. Think of a drum set recorded in a cheap studio, where the mics are picking up something bleary in the atmosphere, or the room does horrible things when the crashes sound. In these cases, a relatively fast attack can be your friend, as you’re trying your best to soften the transients. But don’t go too fast, or else you’ll be imparting distortion.
For this use case, I’d recommend the solid-state compressor in Nectar 3 or the vintage compressor in Neutron 3. You’d think you’d want to use a cleaner algorithm, one with less character—but not so! Go too clean, and you may feel a sensation slightly divorced from reality, the result of a compressor reacting to the overheads. The solid-state or vintage compressors can hide this apparent weirdness to a certain degree.
Be careful though: you usually don’t want to go too far in compressing an overhead cymbal, or else all feels weird and whooshy quite quickly. Save the squashing for the room mic.
Usually when compressing the room, you want to smash. Like, Hulk levels of smash. Here, there’s no better tool in the iZotope arsenal, I find, than Nectar 3’s compressor with a high ratio.
Observe how we can turn this:
With the following settings:
This is what I’m talking about when I say "smash."
You know the sound of sidechain compression: it’s on pretty much every EDM track. Think of the bass ducking in volume every time a four-on-the-floor kick hits, rising back up after.
Here, Neutron 3 has your back; it gives you multiple options for sidechain compression. You can duck the lowest frequencies of a bass to the drums in a multiband-compression situation, if need be, thanks to inter-plugin communication. Or, you can go broadband, and duck an entire instrument bus to the kick drum, giving you the classic pumping of an EDM track.
Here’s how you’d set up the former use. First, place Neutron 3 on the bass, and click into multiple bands, like so:
Make sure you sidechain to the kick drum, like so:
Then select EXT on the lowest frequency range, like so:
And you'll have this:
To sidechain a synth bass to a kick, it’s roughly the same—except there’s no splitting of the bands. You sidechain the entirety of the synth bass to the kick, not just the low frequencies.
Frequency-dependent sidechain compression works wonders on making the bass and kick drum work with each other without obvious ducking. Dramatic sidechain compression also works wonderfully for the EDM pumping trick. The latter is an example of how compression can change the groove of the music: when you compress in this fashion, the whole feel of the music changes. You can really change the intent of the music.
Gentle vocal control with compression
Nectar has an auto-level mode that, when paired with the subtle, musical algorithm of the optical-mode compressor, makes for a smooth-sounding vocal. Simply hit the ALM button shown below, and feed the optical compressor.
The ALM button will give you a transparent hand in achieving a consistent vocal, meaning the compressor doesn’t have to work so hard. In setting the compressor, don’t set the threshold too low. We don’t want to color the vocal too much on this application, and the optical mode does impart some coloration: it adds harmonics in a manner similar to a tube compressor. It may start to sound more obviously colored with heavy heaps of compression, and you might not want this.
Then again, maybe you do.
Tight vocal control with a soft touch
Sometimes it pays to run two compressors in series, one designed to catch peaks, the other for catching more sustained passages of loudness. In the hardware world, this is often done with a combination of the 1176 and the LA2A, both classics. The 1176 is very fast, and can catch peaks quite well. The LA2A is a tube-based optical compressor whose attack and release are slower, and more importantly, musically dependent.
Believe it or not, there is no prescribed order to which goes first! While the settings of the 1176 are generally faster, the order depends on the musical content—one could imagine needing something faster up front for a quickly-spit rap vocal than a slowly-sung power ballad; thus, the rap track would get the faster compressor first in the chain, while you might switch for the ballad.
In the iZotope world, Nectar provides both smooth compressors and fast ones. Optical mode, for example, can be your smoother circuit. Vintage and Digital will get you the same quickness in attack, though digital will have less coloration. Experiment on your own.
Multiband compression on a vocal
Before dynamic equalizers were so prevalent, we often used multiband compression on vocals, many times in four-band implementations. Why? Because vocals exhibit different issues in different frequency ranges—and a singer might move from one range to another within a phrase. Moreover, static EQ wouldn’t cut it in a satisfactory way, as it would force a listless and dull sound overall.
Instead, you’d use the attack and release parameters to fool the ear into thinking it was hearing more of the frequency than it actually was; it would let some of the offending band through, then clamp down.
These days, dynamic EQs facilitate this effect with pinpoint accuracy, but you still may want to utilize the multiband compressor.
Why? Because the specificities of controls offered within something like Neutron’s or Ozone’s multiband compressor are more extensive than the dynamic EQs. You not only get attack, release, and threshold, but also ratio, and oftentimes a knee control, which affects how strict the compression is around the threshold point.
Because of the panoply of control offered here, you can use a subtle hand to shape the offending frequency points, or you could get overly dramatic—so it pays to hear when you’ve gone too far.
Take this vocal:
I’d like to pull down a little between 300–900 Hz or so, because it’s too overwhelming. The brightness is also killing me on certain syllables. Finally, the high midrange needs some tamping down, but not as much as the low midrange.
I could see to these issues with a multiband compressor, but if I went too far, it would sound like this:
Back off the threshold and ratio a bit in each band, and we get this:
Compared to the bypass example, you’ll note it’s doing a subtle job of controlling the issues.
Undoubtedly, dynamic EQ is great if the vocal is too prominent in a specific set of frequencies. For de-essing, dynamic EQ can often work wonders. However, a vocal can sometimes travel too heavily throughout the entirety of the low midrange—from 250 Hz through 600 Hz, for example. In these occasions, multiband compression is the better fit, for it will tamp down in response to a wider range of frequencies.
Neutron 3 will give you three bands of multiband compression. However, I usually favor four bands for vocals, so I’d rather go with Ozone. In this case, it would look like this:
A single-use of Ozone here shouldn’t add latency in this circumstance. Use RMS mode for the cleanest operation, as ENV and Peak modes can add a small amount even ordered harmonics.
Ozone is the only compressor in the iZotope ecosystem that gives you mid/side operation right out of the box. If you wanted to use Nectar or Neutron, either you’d need a DAW that supports mid/side encoding of plug-ins (Wavelab 9 and above, for example), or you’d need to do your own multing and routing beforehand, to ensure a proper mid/side split.
That having been said, I’ve not found too many utilitarian uses for mid/side compression; it can feel quite unnatural, changing the stereo image in a program-dependent way that calls attention to itself.
Sometimes, a creative implementation of mid/side compression can be quite interesting: you could sidechain all stereo elements to the kick drum, so that the signal swells in width after the kick hits, rather than in level alone. This can be really cool, but you must know that it is a creative usage—it will have an audible, noticeable effect.
Conclusion: where did the other instruments and buses go?
You may have noticed we haven’t spoken of how to compress synth parts, pads, guitars, horns, or other instruments. We also haven’t talked about drum buses, instrument buses, or the stereo bus as a whole. What gives, eh? Why the glaring omission?
Simply put, most instruments and instrument submixes are wildly different; there is no panacea. No one set of instructions apply.
Furthermore, Nectar, Ozone, and Neutron have quite different tonalities and modes of operation—heck, you’ll find so many variances among the specific algorithms of Nectar alone that it’s impossible to give you a blanket judgment.
Consider the attack and release parameters of Neutron. This is what the Modern mode looks like, when put through Plug-in Doctor:
This test pushes three test signals through the compressor over time; the second is much louder (or stronger) than the first or third. Thus, the graph shows us the attack curve of the compressor at 1 second, as it responds a signal louder than the threshold. Then, it shows us the release curve at 3 seconds, as the signal trends back to its original level.
In the following screenshot, we’ll compare the Modern setting to Vintage. Note the differences:
Here Modern is displayed in green, while Vintage is pink. We can see a prominently different attack shape.
The settings here are as follows: a threshold of -20, an attack of 20 ms, a release of 100 ms, and a ratio of 4:1 (the ratio of 4:1 lets us compare more accurately to Nectar’s Optical mode.)
Speaking of Nectar, let’s see what happens when we compare Neutron’s Vintage mode to Nectar’s Vintage mode:
The attack curves are, again, quite different, and there’s some discrepancy in the release curves as well, which you can see when we zoom in:
Indeed, when we look at all the modes of Nectar, we can see a great variance:
And we’re only talking about attack/release here. Bring harmonic content into the mix, and things change even more.
Remember when we talked about how different hardware compressors exhibit different flavors and tones based on their circuits? We can see this plainly in the difference between Neutron’s Modern and Vintage modes. Here’s Modern:
And here’s Vintage:
It’s clear that to our 1 kHz test tone, Vintage mode adds even-ordered harmonics, which is a pleasant saturation that reinforces the fundamental of a signal.
Nectar has its even-ordered modes as well. Observe Solid State mode:
But in Optical mode, you’ll note more activity:
This is more in-line with many tube-based circuits, which often add more complex distortions.
All of this analysis is meant to show you just how much choice you have in the matter—and how this choice becomes your weapon to wield when you have the proper knowledge at your disposal. If you know the differences in attack/release shapes for each compressor, as well as the type of harmonic saturation it may induce, you’ll make more informed decisions when compressing instruments.