Parallel Compression on Drums

Parallel Compression on Drums: How to Blend in Punch Without Squashing the Transients

Drums that sound polite usually have the right notes and a decent balance, but the hits lack weight and the tails die too quickly. Compressing them harder in the usual way makes the hits smaller, because a compressor on the main signal turns down the loudest part, which is the transient. Parallel compression avoids that trade. This article builds a parallel drum bus from scratch in any DAW, sets the compressed copy so it adds body and sustain, and covers the phase and latency checks that most walkthroughs leave out.

What parallel compression does to a drum kit that serial compression can’t

A compressor inserted directly on a drum track (serial compression) processes the whole signal. Loud peaks are reduced, the quiet material between them stays put, and makeup gain then lifts everything. Pushed hard, the result is a flatter, denser kit with less difference between the crack of the snare and the room around it.

Parallel compression keeps the original drums untouched and adds a heavily compressed copy underneath. The dry signal supplies the attack and dynamics. The compressed copy supplies what heavy compression does well: it raises the low-level detail, such as the body of the snare, the decay of the toms, the ghost notes and the room ambience, to a level close to the peaks. Blended together, the kit gains density and sustain while the front edge of each hit is still the original.

This also explains why the technique suits drums in particular. Drums have sharp transients and quick decays, so their quiet parts are easy to lose in a mix. A parallel bus lets you raise those parts without touching the hits that already sound right.

Send or duplicate: setting up the parallel bus the right way

There are two ways to build the bus, and both work in every major DAW.

Send method (recommended). Create an auxiliary or return bus, put your compressor on it, and set each drum track’s send to feed it. Keep the bus fader at unity and control the amount with the send levels or the bus fader. Sends let you choose which drums feed the bus and how much each contributes.

Duplicate method. Duplicate the drum group or the individual tracks, compress the copies, and balance them against the originals. This takes more CPU and more tracks, but it allows per-track editing. It suits anyone who wants to alter the compressed copy of a single track without affecting the others.

For the send method, check these before you touch the compressor:

• Set the sends to post-fader. If the send is pre-fader, moving a drum fader changes the dry level but not the amount going to the bus, and the balance between dry and wet shifts without warning.
• Route the bus to the same destination as the dry drums, usually the drum group or the master, so both signals pass through the same downstream processing.
• Set the compressor’s own dry/wet control to 100 percent wet. If it has a mix knob and you leave it at 50 percent, the bus already contains dry signal and you are blending three things instead of two. The exception is a compressor that has a built-in parallel control you deliberately want to use instead of a bus.

Name the bus something obvious like “Drums Crush” so you can find it quickly later.

Dialing in the compressed copy

The compressed copy should sound too compressed when soloed. If it sounds good on its own, it is probably not doing enough to add anything the dry signal lacks. These are starting points, and the right values depend on the kit and the tempo.

• Ratio: start between 4:1 and 10:1. Higher ratios make the copy denser and bring up more room and decay. Limiting-style ratios work as well.
• Threshold: lower it until the compressor is reducing gain on nearly every hit. Aim for roughly 10 to 20 dB of gain reduction on the loudest hits. That is more than most people would use on a normal insert, which is the point.
• Attack: a medium attack, around 10 to 30 ms, lets the first part of each hit pass before the compressor clamps down. The compressed copy then still has some snap of its own, which reinforces the dry transient. A very fast attack removes that snap, giving a duller copy that leans entirely on the dry signal for definition. Both can work, so try both and listen to which one the mix needs.
• Release: set it so the gain reduction recovers before the next hit lands. Watch the gain reduction meter while the loop plays. If the needle is still low when the next kick arrives, the release is too slow. If it snaps back instantly and the sound flutters, it is too fast. Somewhere between 50 and 150 ms is a common starting range for mid-tempo material.
• Makeup gain: raise it until the compressed copy sits at about the same perceived loudness as the dry drums when each is soloed. That gives you a neutral starting point for the blend.

Compressors with different circuits react differently at heavy settings, so use whatever you have. A stock compressor in your DAW is enough to learn the technique. Third-party compressors with character, such as FET or optical-style models, change the tone of the copy but do not change the method.

Blending by ear: finding the level where punch arrives before the mix gets smaller

Start with the compressed bus fader all the way down and play the loop. Raise the bus slowly. At some point the kit will seem to grow: the snare thickens, the kick gains weight and the decay of the drums lengthens. That is the zone you are looking for.

Keep raising it a little past that point and you will hear the cost. The transients begin to feel blunted, the cymbals wash out, the room noise comes up and the whole kit seems to step back in the mix. Then ease the fader back down until the punch is still there and the kit has stopped shrinking.

Some habits make this judgment more reliable:

1. Match loudness when you compare. Adding the parallel bus makes the drums louder, and louder usually sounds better. When you bypass the bus to compare, turn the drum group or master down by the amount the bus added, so both versions play at the same level. Without this, you are choosing louder over better.
2. Listen for what changed, not for what got bigger. Ask whether the snare has more body, whether the kick tail is easier to hear, and whether the groove feels steadier. If you can only say that it is louder, the blend is not working yet.
3. Check it with the rest of the mix playing. A blend that sounds right on soloed drums can fight the bass or the vocal. Set the final level with the full arrangement running.
4. Take breaks. Ears adapt to heavy compression within minutes. Come back after a short rest and toggle the bus on and off again.

The right amount is often lower than you expect. If the bus is clearly audible as an effect, the blend is probably too high.

Checking phase and plugin latency so the dry and wet signals don’t smear

When a dry and a processed copy of the same signal are summed, any timing difference between them causes comb filtering. This makes the low end thin, the snare hollow or the transient soft, and it happens even when the compressor settings are fine. Two checks catch it.

Latency check. Some compressors add latency, especially those with a look-ahead function, oversampling or a linear-phase mode. Modern DAWs apply plugin delay compensation (PDC) automatically, so the dry and bus signals should line up, but PDC can be switched off or limited in some projects and some plugins report their latency incorrectly. To test it:

1. Solo one drum hit, such as a single kick, on the dry track and on the bus.
2. Zoom in on the waveform of the bounced or recorded signals, or use the DAW’s meter, and compare where the transient begins in each.
3. If the compressed copy starts later, switch off look-ahead or oversampling on the compressor, or nudge the dry track later by the reported number of samples.

Polarity check. Flip the polarity (the phase invert button) on the parallel bus and listen. If the low end gets noticeably fuller with the flip, the two signals were partially cancelling. Leave the polarity in whichever position sounds heavier and more solid. Use the same test on a mono kick: if the bass thins when the bus is added, the signals are fighting.

Run these checks after you finish adjusting the compressor, because changing its settings can change its latency. If the rest of your mix has phase problems beyond the drums, the guide on making a mix sound wider without phase issues covers them in more detail.

Parallel on the whole kit versus kick and snare only

Sending the full kit to the bus is the quick option, and it gives a cohesive, room-like sound. The cymbals and hi-hats contribute to the compressor’s gain reduction, though, so each cymbal hit pulls the whole copy down. This can make the bus pump around the hats and emphasize cymbal wash.

Feeding only the kick and snare gives a cleaner result. The compression follows the hits that carry the weight of the beat, and the hats stay crisp. Toms can be added if they matter to the arrangement. If the kit has a live-sounding room or overhead microphones, the room sound is often worth including because that is where the sustain lives, but send it at a lower level than the close mics.

A practical compromise is two buses: one for kick and snare, and one for room or overheads with different settings, usually a slower attack and a gentler ratio. You can then balance body and ambience separately. For programmed drums with no room sound, the kick and snare version is usually the better place to begin.

Adding color to the crushed bus with saturation or EQ

Because the bus is blended in rather than heard alone, you can process it more aggressively than the dry drums without the usual consequences. Two additions are common.

EQ before or after the compressor. A high-pass filter on the bus, set somewhere in the low hundreds of hertz or lower depending on the kick, stops the compressed copy from adding more low-end build-up if the kit is already heavy. Boosting the low mids or the upper mids on the bus can add body or bite that the dry kit lacks. If the heavy compression has exaggerated cymbal hiss, a gentle high-frequency cut on the bus calms it. Place the EQ before the compressor if you want the compressor to react to the shaped signal, and after it if you only want to adjust the tone of the result.

Saturation. A saturator, tape emulation or soft clipper on the bus adds harmonics that make the copy sound thicker and let it be heard at a lower level. That is useful when the bus at a good blend level otherwise feels too quiet. Keep the amount moderate, because distortion on top of heavy compression can turn the copy harsh.

Make these changes with the bus soloed, then recheck the blend with the dry drums, since the level that was right before the added color may no longer be right.

Common mistakes that make parallel drums sound pumpy or lifeless

• The release is too slow or too fast. A slow release keeps the copy turned down when the next hit arrives, so the bus pumps in a way that does not fit the groove. A very fast release makes the sustain flutter. Adjust the release against the loop until the movement locks with the tempo.
• The bus is blended too loud. The most frequent result of this is a kit that sounds big in isolation but lifeless in the full mix, because the transients have been buried. Lower the bus until the dry hits lead again.
• Level matching is skipped. Without it, almost any amount of bus sounds like an improvement, and the blend ends up too high.
• Phase and latency are never checked. A thin low end or a soft snare after adding the bus often traces back to a timing offset, not to the compressor settings.
• The compressed copy is made to sound good on its own. If it is pleasant soloed, it is likely not compressed enough to add anything new.
• Every drum goes to the bus. Hats and cymbals can make the compressor pump. Try sending only the kick and snare.
• The bus is left on while the arrangement changes. A sparse intro and a full chorus need different amounts of parallel compression, so automate the bus level or the send level rather than keeping one setting for the whole song.

Try it on a loop you already have

Take a drum loop you already have, build the parallel bus described here, and compare it with the bypassed version at matched loudness. Then read how to make a mix sound wider without phase issues to keep the rest of your mix as solid as your drums.