Compression is one of the most misunderstood tools in mixing. Every beginner knows they need to "compress" something, but the actual mechanics — what the knobs do, how to set them, how to know if you've gone too far — remain mysterious. The result is either over-compressed tracks that sound flat and lifeless, or under-compressed tracks that still have the dynamic problems compression should have fixed.
This guide fixes that. Not with a list of settings to copy, but with a clear understanding of what compression does and how to make decisions that serve the music. Once you understand the controls — really understand them — you can apply compression correctly in any situation.
What Compression Actually Does
At its core, a compressor reduces the dynamic range of an audio signal — it makes loud parts quieter and, optionally, makes quiet parts louder through makeup gain. Dynamic range is the difference between the loudest and quietest parts of a performance. Compression shrinks that difference.
The purpose: make everything easier to hear at a consistent level. A vocal that whispers at -30 dBFS and belts at -3 dBFS will disappear in the quiet sections and overwhelm the mix in the loud sections unless you even out those levels with compression. Compression makes the vocal sit in the mix consistently so it can be heard clearly throughout the song.
Beyond level consistency, compression adds character. The act of reducing peaks and bringing up the average level makes a track feel more present, punchy, and "in your face." This is why rock vocals, snare drums, and bass guitars almost always have some compression — it adds energy and aggression that the raw performance doesn't quite have.
The Four Key Controls: Threshold, Ratio, Attack, Release
A compressor has four main controls. Understanding each one is the key to using the tool correctly.
Threshold is the level (in dBFS) at which the compressor starts working. Signal above the threshold gets compressed. Signal below the threshold passes through unaffected. Set the threshold low, and more of the signal gets compressed. Set it high, and only the loudest peaks trigger the compressor.
Ratio is how much compression gets applied once the signal exceeds the threshold. A ratio of 4:1 means for every 4 dB that the signal exceeds the threshold, only 1 dB passes through. The remaining 3 dB is reduced. A ratio of 1:1 means no compression — everything above threshold passes through unchanged. As ratio increases, compression gets heavier. At infinite:1 (often called "limiting"), nothing above threshold gets through at all.
| Ratio | Intensity | Common Use |
|---|---|---|
| 1:1 to 2:1 | Very light | Glue, gentle level control |
| 3:1 to 4:1 | Light to moderate | Vocals, acoustic instruments |
| 6:1 to 8:1 | Moderate to heavy | Snare, electric guitar, rock vocals |
| 10:1 to infinite:1 | Heavy / limiting | Master bus, catching peaks |
Attack is how quickly the compressor engages once the signal crosses the threshold. A fast attack (1–10 ms) catches the initial transient of a sound — the initial spike when a drum is hit or a word is spoken. A slow attack (20–100 ms) lets some or all of the transient through before the compression kicks in.
For drums and percussive sounds, a fast attack is usually correct — you want to catch the transient and control it. For vocals and sustained instruments, a slower attack is often better — letting the initial consonant or note attack through naturally before the compression smooths out the tail.
Release is how quickly the compressor stops compressing after the signal drops below the threshold. A fast release (20–100 ms) lets the compressor reset quickly, ready for the next peak. A slow release (100–500 ms) holds the compression longer, creating a "smoothing" effect that can sound more natural on sustained instruments or create "pumping" if set wrong.
Fast attack + fast release = snappy, immediate compression that controls peaks tightly but can sound choppy and unnatural on sustained sounds. Slow attack + slow release = smoother, more natural compression that lets transients through but takes longer to recover. The right combination depends on the source and the style you want.
Step-by-Step: Setting Up a Compressor
Here's the process for setting a compressor from scratch:
1. Set ratio first. Start with a moderate ratio (3:1 or 4:1) and adjust later based on what you hear.
2. Set attack and release to starting positions. For vocals, start with attack at 10–20 ms and release at 60–100 ms. For drums, start with attack at 1–5 ms and release at 50–100 ms.
3. Set threshold last. With the ratio set, lower the threshold until you see 3–6 dB of gain reduction on the loudest peaks. This is the starting point — adjust based on what you hear.
4. Listen and adjust. Does the track sound more controlled but still natural? Keep the settings. Does it sound flat and squashed? Back off the threshold or lower the ratio. Does it sound like it's breathing or pumping in time with the beat? Adjust the release.
5. Adjust ratio. Once threshold is set, adjust the ratio. If 3:1 isn't controlling the peaks enough, go to 4:1 or 6:1. If the track is losing life at 4:1, back it down to 3:1 or even 2:1.
6. Use makeup gain. Compression reduces the overall level of the track. Use the makeup gain (if your compressor has one) to bring the output back up to match the input level — so you can A/B the compressed and uncompressed versions at the same loudness, which is the only way to fairly judge whether compression is helping.
Understanding Gain Reduction and Metering
Most compressors have a gain reduction meter showing how much compression is happening at any moment. This meter shows the "working" of the compressor — how much it's reducing the signal. Watching this meter tells you when the compressor is active and how hard it's working.
Reading the meter: If the meter shows 0 dB of gain reduction, the compressor is doing nothing — everything is below threshold. If it shows -6 dB, the compressor is reducing the signal by 6 dB. Typical compression for most sources is 3–6 dB of gain reduction on the peaks. More than 8–10 dB is usually too much for anything except special effect compression.
Matching input and output levels. When comparing compressed vs. uncompressed, the compressed version will always sound louder if you don't match levels. Use makeup gain or reduce the input level of the uncompressed signal so they're at the same loudness. Only then can you hear whether the compression itself is improving the track or just making it louder.
For most mix compression tasks — vocals, snare, bass, guitar — you're aiming for 3–6 dB of gain reduction on peaks. This is enough to even out dynamics without sounding processed. If you're reaching for more than 8 dB to make a source sound controlled, the problem is usually in the arrangement or performance, not in the compressor settings.
Parallel Compression — When and Why
Parallel compression (also called New York compression) is a technique where you blend a heavily compressed signal with an uncompressed (or lightly compressed) version of the same track. This gives you the punch and density of heavy compression while retaining the dynamics and natural character of the original signal.
The process:
- Set up two sends from your channel — one to a dry track (no compression), one to a wet track with heavy compression.
- Set the compressor on the wet track for 6–10 dB of gain reduction — heavy by normal standards.
- Blend the wet track in with the dry track until you get the right balance of punch and naturalness.
This technique is especially useful on drums and vocals. A heavily compressed drum bus adds size and weight to a mix. A parallel compressed vocal retains its dynamics and presence while gaining density and closeness. It's one of the most effective mixing techniques for adding power without losing life.
Common Mistakes to Avoid
Over-compression. The most common mistake. If you're seeing more than 8–10 dB of gain reduction, you've almost certainly gone too far. Compression should control dynamics — it shouldn't eliminate them. A compressed track should still have movement and variation, just less extreme.
Not matching levels when comparing. If you A/B compressed vs. uncompressed and the compressed version sounds better because it's louder, you've learned nothing. Use makeup gain or a gain knob to match the levels before comparing. The difference you hear should be about the sound, not the loudness.
Setting threshold by eye instead of ear. The gain reduction meter tells you how much compression is happening, but it doesn't tell you whether that compression sounds good. Always listen first. If the track sounds worse with compression, don't add it just because the meter shows activity.
Using compression as a fix for a performance problem. If a vocal has wildly inconsistent levels because the singer moved away from the mic, compression will make it more consistent but won't fix the underlying problem — which is the recording, not the level. Fix it in the recording next time, and use compression only for the normal dynamics variations that every performance has.
Fast attack on everything. A fast attack catches transients and controls peaks — but it also removes punch and impact from drums, bass, and any percussive source. If you want a drum to sound punchy and powerful, use a slower attack that lets the initial transient through. Save the fast attack for when you need to control peaks tightly, not as a default setting.
Learn how to apply this to specific tracks in a mix — read our compression for beginners guide which covers the same controls in the context of specific instruments.