Compression gets sold as a way to make things louder, and that framing causes most of the confusion around it. A compressor does not make anything louder on its own. It reduces the distance between the quietest and loudest parts of a signal, and that narrower range is what then gets turned up without clipping. Everything a compressor does follows from that one idea: it is a tool for controlling range, not volume.
The four core controls, threshold, ratio, attack and release, are usually taught as isolated settings. They only make sense together, and the sections below build them up through what you actually hear rather than what each knob is labeled.
Threshold and ratio: when it kicks in, and how hard
Threshold sets the level a signal has to cross before the compressor does anything at all. Below threshold, the signal passes through completely untouched. Above it, the compressor starts pulling the level down.
Ratio decides how hard that pull is. A 2:1 ratio is gentle, for every 2dB the signal goes over the threshold, only 1dB gets through. A 10:1 ratio is aggressive, 10dB over threshold becomes just 1dB of actual increase, which is close to the territory where a compressor starts behaving like a limiter. Low ratios are for smoothing a signal without it being obvious. High ratios are for catching peaks that would otherwise blow past everything around them.
Threshold and ratio together decide how much gain reduction happens. Neither one means anything on its own, a low threshold with a low ratio can produce the same amount of reduction as a high threshold with a high ratio, just shaped differently.
Attack: the most underrated decision on the compressor
Attack is how long the compressor waits, after the signal crosses the threshold, before it actually starts reducing gain. This single setting decides whether the initial transient, the sharp, percussive front edge of a sound, gets through untouched or gets caught by the compressor along with everything else.
A fast attack grabs that transient immediately, which tames sharp peaks but can also drain the punch out of a drum hit if pushed too far. A slow attack lets the transient pass through uncompressed before the compressor engages, which is exactly how many drum bus compressors add perceived punch: the initial hit stays sharp, and only the sustain gets pulled down afterward. There is no universally correct attack time. The decision is always about which part of the sound you want to protect and which part you are willing to control.
Release: letting go, and the risk of pumping
Release controls how long the compressor takes to stop reducing gain once the signal drops back below the threshold. A fast release lets the signal snap back to full level almost immediately, which can sound energetic on a bassline but audibly pumps if it happens in time with the beat in a way you did not intend. A slow release holds the gain reduction longer, smoothing things out but risking a compressor that never fully lets go between hits, flattening the dynamics of an entire passage.
This is also the setting most responsible for the classic sidechain pumping effect used deliberately in house and techno, where a bass or pad audibly ducks and swells in time with the kick. Used on purpose, it is a rhythmic tool. Used by accident, on a vocal or a full mix, it just sounds like the track is breathing when it should not be. Attack and release on a compressor are conceptually close to the attack and release stages of an ADSR envelope, they both describe how something changes over time, just applied to gain reduction instead of a note.
Case study: leveling an uneven vocal
A vocal that jumps between whisper-quiet and suddenly loud is the textbook use case for compression as a leveling tool. A moderate threshold, a ratio around 3:1 to 4:1, a fast-ish attack to catch the loud moments, and a medium release that lets the gain recover between phrases rather than mid-word, brings the quiet and loud parts closer together without anyone noticing the compressor is there. The goal in this case is invisibility: if a listener can point to the exact moment compression is happening, it is usually working too hard.
Case study: punchy drums without killing them
Drum bus compression asks for the opposite instinct. Here, a slower attack, enough to let the initial transient of the kick and snare through uncompressed, paired with a fast to medium release and a moderate ratio, adds glue and perceived punch instead of squashing everything flat. The compressor is doing real work, gain reduction is happening, but it is acting on the tail of each hit rather than the strike itself, which is why the result sounds more powerful instead of smaller.
Set up the vocal-leveling case first: moderate threshold, ratio around 3:1, fast attack, medium release, and watch how consistently the gain reduction meter moves. Then, without resetting, slow the attack down significantly and switch to a fast release. Notice how the same source now pumps and breathes instead of sitting still, even though threshold and ratio never changed.
Compression is a decision, not a volume boost
Every setting on a compressor is really a decision about which part of a sound to protect and which part to control. Threshold and ratio decide how much happens. Attack and release decide when it happens, and it is that timing, more than the amount of gain reduction, that determines whether compression sounds transparent, punchy, or obviously pumping.
Compression sits naturally after EQ in most mixing workflows, since a cleaner frequency balance gives a compressor less unwanted material to react to. Our article on EQ basics covers that earlier step, and the attack and release logic here connects directly back to ADSR envelopes if you want to see the same time-based thinking applied to synthesis instead of dynamics.
Electronic Music Production for Beginners
Mixing fundamentals like this one, taught in sequence, with hands-on assignments and the full library of interactive tools.
