What Is Audio Compression?
Audio compression is a processing effect that automatically turns down the loudest parts of a signal to make the overall level more even. When a vocal jumps from a whisper to a shout, a compressor reins in the shout so the two sit closer together — and if you then raise the whole thing back up, the quiet parts come forward too. The result is audio that stays at a steadier level without you riding a fader by hand. It's one of the most-used tools in mixing and mastering, and it's a completely different thing from making a file smaller.
That distinction trips people up constantly, so let's clear it first.
Two things share the name "compression"
- File-size compression shrinks the file — fewer megabytes on disk — by lowering the bitrate or sample rate. If you came here to email a track or save storage, that's the other kind, covered in how to compress audio files.
- Dynamic-range compression (this article) changes the sound, not the file size. It narrows the gap between the quiet and loud parts so the level stays more consistent.
They share a word and nothing else. Everything below is about the second kind: the dynamics processor.
What a compressor actually does
Every compressor works on one idea. You set a level — the threshold — and whenever the signal rises above it, the compressor reduces the volume of that part automatically. Anything below the threshold passes through untouched. So the loud peaks get pulled down toward the quiet stuff, and the distance between them shrinks.
That distance is a track's dynamic range — the span in decibels between its softest and loudest moments. Compression's whole job is to reduce that span in a controlled way. Turn the peaks down, then use makeup gain to lift everything back up, and the quiet passages end up louder than they started while the peaks stay put. The track now sits in a tighter, more even band.
None of this touches the file. A compressed WAV is exactly the same size as an uncompressed one — the bytes are unchanged, only the levels inside them moved.
The controls, and what each one does
Compressors look intimidating because they have several knobs, but each one answers a simple question: how loud does it have to get before I act, how hard do I act, and how fast do I react.
| Control | What it sets | In plain terms |
|---|---|---|
| Threshold | The level above which compression starts | How loud a part must be before it gets turned down. Lower threshold = more of the signal affected. |
| Ratio | How much the part over the threshold is reduced | 2:1 is gentle, 4:1 is firm, 10:1 and up is heavy. A 4:1 ratio means 4 dB over the threshold comes out as 1 dB over. |
| Attack | How fast it clamps down once the threshold is crossed | Fast attack catches sharp peaks and softens punch; slow attack lets the initial hit through, keeping snap. |
| Release | How fast it lets go after the level drops back | Too fast can sound pumpy; too slow and it never recovers between hits. |
| Knee | How gradually compression engages near the threshold | A soft knee eases in for a natural feel; a hard knee snaps on abruptly for tighter control. |
| Makeup gain | A volume boost after compression | Raises the now-quieter signal back up, bringing the soft parts forward with it. |
Attack and release are the two that shape feel rather than just level, because they decide how the compressor treats transients — the brief, sharp spikes at the front of a drum hit or a plucked string. A slow attack lets the transient punch through before the compressor grabs the tail; a fast attack rounds it off. That single choice is the difference between a snare that cracks and one that thuds. (The reductions themselves are measured in decibels, so it helps to be comfortable with the dB scale before you start dialing.)
What compression is used for
A few jobs come up again and again:
- Controlling vocals. Singers and speakers naturally swing between loud and soft. A compressor evens that out so every word stays audible in the mix without the loud ones jumping out — the single most common use there is.
- Adding punch. With the attack set to let the initial hit through, a compressor can actually make drums feel more aggressive by clamping the body while leaving the transient alone.
- Glue. A gentle compressor across a whole mix or a group of instruments nudges them to move together, making separate parts feel like one cohesive thing. Engineers call this "glue" for a reason.
- Consistency. For podcasts, audiobooks, and streaming, a steady level matters more than drama. Light compression keeps a voice sitting in the same place so listeners aren't reaching for the volume.
The common thread: compression trades some dynamic range for control. Used with a light hand, that trade is almost always worth it.
Compression vs limiting
A limiter is just a compressor pushed to an extreme. Set the ratio very high — 20:1, or effectively infinite — and put the threshold near the top of the signal, and instead of gently turning peaks down you get a brick wall that refuses to let anything cross a ceiling at all. Nothing gets through above the limit, period.
That makes a limiter the tool of choice at the very end of the chain, where the job is to raise a track's overall loudness right up to the maximum without letting any peak break through and clip. Think of it as a spectrum: gentle compression evens things out, and limiting is the same process taken to its hardest setting. Leaving a little headroom below the ceiling gives the limiter room to work cleanly.
The over-compression trap
Because compression plus makeup gain makes a track louder, it's tempting to keep pushing — and for about two decades the industry did exactly that. The "loudness war" saw masters crushed flatter and flatter, each one chasing the last for attention on the radio. The cost was dynamics: the quiet parts and the loud parts were squeezed into the same narrow band until the music stopped breathing and started fatiguing the ear.
Two things make that a bad bet today. First, over-compression is audible — flatten the contrast between a verse and a chorus and the chorus has nothing to push against, so the "big" moment lands soft. Second, streaming platforms now normalize loudness on playback, turning loud tracks back down to a reference level. Spotify targets around -14 LUFS, Apple Podcasts around -16 LUFS, and broadcast under EBU R128 uses -23 LUFS (some of those are published spec, others measured convention). Crush your master past the target and the platform simply attenuates it — you gave up your dynamics and got nothing louder in return. The track that kept its contrast often wins at matched loudness.
The lesson isn't "never compress." It's compress to shape the sound, not reflexively to make it louder, and check what you did at matched level. Note that this is different from normalization, which just applies one overall gain change to hit a loudness target — it moves the whole track up or down without changing the dynamics inside it. Compression reshapes the internal balance; normalization repositions the whole thing.
Doing it in the browser
You can hear all of this without installing anything. On vocalcut.com, every tool runs entirely on your own device — your audio never leaves it (the processing engines download to you, not your files up to a server), and there's no account or upload step. The in-browser mastering chain applies multiband compression and limiting with live LUFS metering, so you can watch the level even out and compare against the original at matched loudness instead of being fooled into thinking louder is better. If all you actually need is a finished file brought to a consistent target level, the volume normalizer measures the track and applies gain to hit it — no dynamics squashing involved. And for a control-by-control walkthrough with starting settings for vocals and full mixes, see how to use a compressor on audio.
Frequently asked questions
Is audio compression the same as compressing a file? No. Dynamics compression evens out the loud and quiet parts of the sound and leaves the file size unchanged. File compression shrinks the file in megabytes by lowering bitrate. Different tools, different goals — the file-size guide covers the other one.
What is the threshold on a compressor? The threshold is the level a signal has to exceed before compression kicks in. Parts below it pass through untouched; parts above it get turned down by the amount the ratio dictates. Lowering the threshold means more of the track gets compressed.
What's the difference between a compressor and a limiter? A limiter is a compressor with a very high ratio and a threshold near the ceiling — it acts as a hard wall nothing can cross. A compressor gently reduces peaks; a limiter refuses to let them through at all. It's the same process at its most extreme setting.
Does compression make audio louder? Indirectly, yes. Compression itself only turns peaks down, but because it frees up space at the top, you can add makeup gain to raise the whole signal — which brings the quiet parts up and makes the track feel louder overall. Pushed too far, though, it flattens the dynamics that give music impact.