Cut and high-pass before the compressor, boost after it. Whatever sits in front of a compressor decides what it reacts to, so that slot is for removing things you don't want it to hear. Whatever sits after it changes the tone of the finished, level-controlled sound.
If you only want to change what the compressor reacts to, and leave the sound alone, put the EQ in its sidechain. On a drum bus or a mix bus that usually means a high-pass, so the kick and bass stop making everything else pump.
At a glance
| Property | EQ first | EQ after | Sidechain EQ |
|---|---|---|---|
| The compressor hears | The EQ'd sound | The original sound | The EQ'd sound |
| You hear | EQ, then compression | Compression, then EQ | Compression only, no EQ |
| A boost | Pushes the compressor harder, which can hold it back down | Reaches the output as set | Makes the compressor react more to that range |
| A cut or high-pass | Stops the compressor reacting to that range | The compressor still reacts to what you removed | Stops it reacting, and leaves the range in the sound |
| Best for | Cleaning up: high-pass, boom, mud, resonances | Tone: presence, air, weight | Bus compressors, de-essing |
Does the order really matter?
Yes, whenever the compressor is doing real work. An EQ on its own is a linear process: it changes each frequency by the same number of decibels however loud the sound is. Two EQs in a row give the same result in either order, and so do an EQ and a plain volume change.
A compressor isn't linear. It only turns the sound down when the level crosses its threshold, so anything that changes the level going into it changes when and how hard it works. An EQ boost of 9 dB at 80 Hz makes every kick 9 dB louder in that range. A compressor placed after that EQ hears the bigger kicks and turns them down harder, while a compressor placed before it never hears the boost at all.
One more detail catches people out. A normal compressor turns the whole sound down, every frequency at once, even when only one range set it off, so if the low end is what crosses the threshold, the vocal, the hats and the pads all dip with it. That's why an EQ placed before a compressor can change the level of things the EQ never touched.
Reading the lab
The four cards along the top of the lab are the four places the EQ can sit. EQ first and EQ after are the two orders the question is about. Split puts the high-pass before the compressor and the bell after it, and Sidechain puts the whole EQ in front of the compressor's detector only, so it changes what the compressor listens to and nothing you hear. Press 1 to 4 on your keyboard to switch between them while the loop plays.
The big screen shows the loop going in (pale) and coming out (coloured) at the same loudness. Under it, the gain reduction screen runs the same loop through all four routes at once, one lane each, with the most each one takes off written on the right. The lane for the route you're hearing is the bright one. When the EQ is flat, the four lanes are identical, which is the quickest way to see that order only matters once the EQ is doing something.
The tone screen shows how much louder or quieter each frequency comes out than it went in, measured on the loop at matched loudness. The grey line is the EQ with the compressor switched off. The solid line is the route you're hearing and the dashed one is the route it's being compared with. Where those two part company, the compressor has changed what the EQ did. The panel on the right puts the same measurements into words, and A/B plays the original at the same loudness as the processed sound.
EQ before compression
With the EQ first, the compressor reacts to the sound as the EQ left it. That's exactly what you want for cuts. In step four the bell takes 9 dB off the drums at 80 Hz before the compressor, and the most it takes off falls from 11.2 dB to 8.9 dB, because the kick that was setting it off is quieter. The snare and hats stop ducking under every kick as much.
The same applies to a high-pass filter. Rumble, handling noise and the thump of a plosive on a vocal mic are all low and loud. Leave them in and the compressor turns the whole vocal down each time one comes through. Filter them out first and the compressor's gain reduction follows the singing.
Boosts are where EQ first gets awkward. Step two adds 9 dB at 80 Hz in front of the compressor, and the compressor takes up to 13.9 dB off instead of 11.2 dB. On a loop where the kick sets the compressor off, that extra work lands on every kick, so the whole loop pumps harder. On the vocal in step six, an 8 dB boost at 5 kHz in front of the compressor measures +6.6 dB at matched loudness, against +7.8 dB when the boost comes after it: the compressor reacts to the brighter consonants and holds part of the boost back down.
Sometimes that's the point. A broad mid boost before a compressor makes it react more to the mids, which is one way to make a part sit more evenly. Just know that the boost and the compressor are now arguing with each other.
EQ after compression
With the EQ after the compressor, the compressor works on the original sound and the EQ shapes what comes out of it. A boost here reaches the output at the size you set, and it doesn't make the compressor work any harder. Step three is the same 9 dB low boost as step two, moved after the compressor, and the gain reduction goes straight back to 11.2 dB.
On that drum loop the measured boost at 80 Hz is within 0.1 dB in either order. What changes is the movement. With the boost first, every kick pulls the rest of the loop down further, while on the vocal the difference shows in the tone as well, because the compressor was reacting to the very range being boosted. Expect order to change the dynamics on most sources, and the tone mainly when the boosted range is what's driving the compressor.
Cuts after the compressor have the opposite problem. In step five the 9 dB cut at 80 Hz comes after the compressor, so it still clamps down on every kick at full size. The kick you end up hearing is 9 dB quieter, but the rest of the loop is still ducking under it.
Cuts before, boosts after
Put the two halves together and you get the usual starting point for a channel: one EQ before the compressor for cuts, and one after it for boosts. On a vocal that might be a high-pass at 80 to 100 Hz and a narrow cut at a ringing room resonance in front of the compressor, then a gentle presence lift around 3 to 5 kHz and a high shelf for air after it.
The Split card in the lab does this with one EQ, the high-pass in front and the bell after. On the vocal in step seven the 5 kHz boost comes back up to +7.8 dB at matched loudness. This particular recording has very little energy below 100 Hz, so the high-pass hardly changes what the compressor does here; on a vocal recorded close to the mic, with proximity boost or stand noise, it would.
Several channel strip plug-ins put an EQ and a compressor in one window with a switch for their order, and the default is worth checking the first time you use one. In any DAW you can simply load two EQs, one in the slot above the compressor and one below.
Using EQ on the compressor's sidechain
The sidechain is the signal a compressor listens to. Normally it's the same sound it's compressing, but most compressors let you filter it. Put an EQ there and you change what the compressor reacts to without changing a single frequency in what you hear.
The everyday use is a high-pass on a bus compressor. On the full mix in step eight, a 150 Hz high-pass in the sidechain lowers the compressor's biggest move from 10.3 dB to 8.6 dB. The kick and bass stay in the sound at full weight, but they no longer drag the chords down with every beat. Somewhere between 80 and 150 Hz is a sensible starting range on a drum or mix bus.
Ableton Live's Compressor has an EQ in its sidechain section, Logic Pro's Compressor has a sidechain filter, and FabFilter Pro-C 2 has sidechain filters too. Boost the sibilant range in a compressor's sidechain instead and it reacts mainly to the esses and turns the whole voice down on each one, which is how a basic broadband de-esser works. The de-essing lesson goes further with that idea, and sidechain compression covers feeding a compressor a different sound altogether.
EQ and compression order for common sounds
These are starting points. Each one puts clean-up before the compressor and tone after it, and the last column says where the sidechain filter earns its place.
| Source | Before the compressor | After the compressor | Sidechain |
|---|---|---|---|
| Lead vocal | High-pass around 80 to 100 Hz, cuts for boom or harsh rings | Presence around 3 to 5 kHz, air above 10 kHz | Rarely needed; use a de-esser for esses |
| Kick | Cut boxiness around 300 to 500 Hz | Weight around 50 to 80 Hz, click around 3 to 5 kHz | No |
| Bass | High-pass the sub rumble below the lowest note | Low-end weight, definition around 700 Hz to 1 kHz | No |
| Acoustic guitar | High-pass, cut body boom around 150 to 250 Hz | Sparkle on top | No |
| Drum bus | Any corrective cuts | Broad tone moves | High-pass around 80 to 150 Hz |
| Mix bus | Usually nothing | Gentle broad tone moves | High-pass around 80 to 150 Hz |
How to set up EQ and compression on a track
- Load an EQ, then the compressor, then a second EQLeave both EQs flat for now. Having the slots ready makes it easy to put each move in the right place.
- Clean up in the first EQHigh-pass up to just below where the sound starts to lose body, and cut anything that rings or booms. These are the things you don't want the compressor reacting to.
- Set the compressorLower the threshold until the loudest moments get the gain reduction you want, then set attack and release. The compression lesson walks through each control.
- Shape the tone in the second EQMake your boosts here, listening to the compressed sound in the mix.
- Try moving one boost in frontIf a boost makes the part sound too jumpy, try it before the compressor and see whether you prefer it held down.
- Compare at matched loudnessBypass the chain and match the levels before you decide. Louder nearly always sounds better at first.
Where this pair sits among reverbs, saturation and everything else is covered in how to order effects in a chain, and a full worked example is in the vocal chain lesson.
Common mistakes
High-passing after the compressor
The filter removes the rumble from what you hear, but the compressor has already turned the whole track down every time the rumble came through. If the gain reduction meter jumps on low thuds you can't hear, move the high-pass up a slot.
Boosting into a compressor and wondering where the boost went
A big boost in front of a hard-working compressor gets partly turned back down, so you keep boosting more. Move the boost after the compressor and you'll usually need less of it.
Using the main EQ to stop pumping
If the kick or bass is making a bus compressor pump, cutting the low end out of the bus fixes the pumping and loses the low end too. A high-pass in the compressor's sidechain keeps the low end and stops the pumping.
Judging the order by level
Moving an EQ around a compressor changes the output level, and the louder version tends to win. The lab matches loudness for you. In your DAW, adjust the compressor's makeup gain or the second EQ's output until both orders play equally loud.
Training your ears to hear the order
The quiz under the lab puts a 10 dB boost either before or after a compressor working hard, on whichever sound you've picked, and asks where the EQ is. With the screens covered, listen for movement. When the boost comes first, the boosted range sets the compressor off, so the sound pumps on it and the boost feels held back. When it comes after, the boost is bigger and steadier. The drums and the vocal are the easiest sounds to start with.
Questions people ask
Should EQ go before or after compression?
Put cuts before the compressor and boosts after it. A cut in front stops the compressor reacting to boom, rumble or harshness you were going to remove anyway, and a boost after it reaches the output at the size you set. When one EQ has to do both jobs, put it before the compressor if it mostly cuts and after it if it mostly boosts.
Should a high-pass filter go before or after a compressor?
Before. Low rumble, handling noise and plosives carry a lot of energy, and a compressor in front of the filter turns the whole sound down every time one of them arrives, even though the filter then removes them. If you want the compressor to ignore the low end but keep it in the sound, use the compressor's sidechain filter instead.
Is it bad to use two EQs, one before and one after the compressor?
No, it's common, and it's the simplest way to put cuts before the compressor and boosts after it. The first EQ cleans up what the compressor would otherwise react to, and the second shapes the tone once the level is under control. A stock DAW EQ uses very little processing power, so the second one costs you almost nothing.
Does EQ order matter with a light compressor?
Less, as long as the EQ move is small too. With 1 or 2 dB of gain reduction the level hardly moves, so the two orders sound close. A big move changes that: a 6 dB boost in front of a light compressor can push it into working hard, so glance at the gain reduction meter whenever you move an EQ in front of one.
What does a compressor's sidechain EQ do?
It changes what the compressor listens to without changing what you hear. A high-pass in the sidechain stops a bus compressor pumping on the kick and bass. A boost around the sibilant range makes it react mostly to the esses, which is how a basic broadband de-esser works.
Where does a de-esser go relative to EQ and compression?
After any EQ that adds top end, so it catches the sibilance that the boost brought up. After the main compressor is a common place for it too, because compression turns the vowels down more than the short esses, and a slow attack lets the esses through untouched, so they can come out louder against the vowels than they went in.