Parallel Compression

There’s a paradox at the heart of compression: the settings that make a drum kit sound powerful and present are often the same settings that make it sound flat and lifeless. Push the ratio high, clamp the attack, and you get control — but you also lose the crack of the snare, the initial burst of energy that tells your nervous system something just happened. Parallel compression is the solution to this contradiction, and once you understand it, you’ll wonder how you worked without it.

The principle is simple. Instead of running your signal through a compressor and using only that processed output, you blend the compressed version back with the dry, uncompressed original. The compressor can be set as aggressively as you like — high ratio, fast attack, heavy gain reduction — because you’re not relying on it alone. The uncompressed signal preserves the transients, the initial attack, the natural dynamics. The compressed signal fills in the body underneath. Together they produce something neither could achieve separately: density and punch at the same time.

New York Compression

The technique became known as New York compression through its use in the mixing rooms of the 1970s and 80s, particularly on drums. Engineers working on large-format consoles would send drums to an auxiliary bus, crush that bus with a compressor — often an 1176 or an SSL bus compressor pushed hard — and blend that crushed signal back under the main drum mix at a relatively low level. The effect was felt more than heard. The room opened up, the snare acquired a weight it hadn’t had before, and the whole kit seemed to sit more confidently in the track without the transients being sacrificed.

It’s worth naming what’s actually happening sonically. Heavy compression raises the average level of a signal, bringing up the sustain, the decay, the ambience, the parts that naturally sit below the peaks. When you blend that back with the original, you’re effectively adding those elements back in — more room sound, more body, more sustain — while leaving the initial hit intact. The transient from the dry signal cuts through; everything the compressor excavated comes from the wet signal behind it.

Practical Settings

On a parallel drum bus, start with a ratio somewhere between 4:1 and 10:1 — don’t be timid. A fast attack (1–5ms) and a medium release (50–100ms, or programme-dependent if the compressor has it) will let the compressor grab the signal quickly and pump in a way that actually adds energy rather than sounding like a mistake. Push the gain reduction hard: 10–20dB is not unusual. This would be destructive applied directly, but in parallel it simply contributes texture and weight to the blend.

The blend itself is where the technique lives or dies. Bring the compressed channel up slowly from silence until you feel the drums change — until they sit down into the track rather than floating on top of it. Often this point is quieter than you expect. The compressed signal might sit 6 to 12dB below the dry signal in the mix, but you’ll hear its absence immediately if you mute it. If you find yourself pushing the compressed signal as loud as the dry signal, something has gone wrong: you’re no longer doing parallel compression, you’re just mixing two versions of the signal and the effect will sound bloated.

The approach transfers naturally beyond drums. Parallel compression on a vocal can add presence and consistency without the over-processed, choked quality that comes from heavy direct compression. On a bass, it can bring up the note body and sustain while preserving the pick or finger attack. On a full mix bus, used subtly, it can add cohesion in the same way the old New York engineers used it on their drum submixes.

One practical note: when using parallel compression, be aware of phase. Most digital compressors introduce some latency, and if your DAW doesn’t handle delay compensation automatically across parallel paths, you can get comb filtering that makes the blend sound thin and hollow. Most modern DAWs manage this transparently, but it’s worth checking — align the peaks of the dry and compressed signals on the timeline if anything sounds off.

The deeper lesson here is about what compression is actually for. It’s not just a tool for reducing dynamic range — it’s a shaping tool, something that changes the character and envelope of a sound. Parallel compression makes that character-shaping available to you at any intensity, without forcing you to sacrifice the natural life of the signal. You set the compressor to whatever extreme achieves the texture you want, then you decide how much of that texture to introduce. Control and dynamics stop being opposites.

Experiment with these techniques using the compression tools in Resonillator’s Mixing Guide module.