How to Set Up a Compressor for the First Time

Production guides · Meme Plugins · 26 October 2026 · Updated 27 October 2026

Start a compressor by choosing a specific dynamic problem, then lower the threshold until the relevant parts trigger gain reduction. Set the ratio for the amount of control, adjust attack and release while listening to the phrase, and match output loudness before comparing with bypass. Change one parameter at a time.

A compressor does not need to work on every track. If one quiet word needs a deliberate level ride, editing or automation may be a clearer first move than compressing the entire performance.

1. Choose a phrase and a job

Loop a passage that includes the event you want to control and the sound immediately after it. “The strong entrance jumps out from the verse” defines a useful task. “Make it louder” does not tell you which dynamics should change.

Keep a clean version. Choose an ordinary downward-compression mode for this first exercise. Leave special character, parallel blending and external sidechain features for a separate experiment. If auto makeup or automatic threshold is enabled, understand what it is changing before you start.

2. Learn threshold and ratio together

Threshold sets the level region where reduction begins; ratio describes how strongly levels above that region are restrained. A softer knee makes the transition more gradual. FabFilter’s dynamics-control manual explains these controls and distinguishes ratio from a maximum-reduction range.

For a simplified hard-knee, steady-state arithmetic example, suppose the threshold is −18 dB and the input reaches −10 dB: eight decibels above threshold. At 4:1, that excess becomes two decibels, giving about −16 dB before makeup gain. Real attacks, knees and detector behavior make a musical performance more complicated than that static calculation.

Begin with a modest ratio, such as 2:1, as an experiment. Lower threshold until you hear and see the chosen event respond. Compare a little more and less reduction. Neither a ratio nor a gain-reduction reading is a universal target.

3. Hear attack before changing release

Attack governs how quickly gain reduction develops. Keep threshold and ratio steady, then compare a faster and slower setting on the same phrase. Listen to consonants or drum attacks as well as the body afterward. Faster control can restrain an entrance, but may remove an attack you wanted to keep.

FabFilter’s time-control reference describes attack and recovery behavior and notes that it depends on programme and compressor style. A time printed on a control does not mean every compressor follows an identical envelope.

4. Let the phrase explain release

Hold attack steady and compare release settings. Listen to the transition from the loud event into the next word or hit. Does the sound recover naturally, remain pulled down too long, or change level in a distracting way? Watch gain reduction to understand what you hear rather than aiming to reset the display exactly on every beat.

Example: one vocal line, three separate decisions

Imagine a line with a strong opening word, an even middle and a quiet ending. This is an illustrative listening exercise, not a measured recording:

  1. Set threshold so the strong opening is clearly affected. Check whether the middle is being reduced more than the intention requires.
  2. Change only attack. Keep the setting that controls the entrance while preserving intelligibility.
  3. Change only release. Check that the quiet ending is not held down by the earlier word. Use a local level ride if that ending still needs help.

If every change seems better only because the signal gets louder, stop and match the output.

5. Compare at matched loudness in the mix

Use output or makeup gain to make processed and bypassed playback comparably loud. Follow the plugin level-matching workflow, then ask whether the phrase sits more consistently without losing its character.

Check the complete arrangement and another section. A setting that suits the verse may flatten the chorus. If the source already has unwanted overload, use the clipping guide; ordinary compression does not restore a damaged waveform.

Compressor questions

Should I copy a preset’s threshold?

The result depends on the incoming signal level and detector. Keep the preset only as a starting point, then set it against your actual phrase.

Why is the compressed version quieter?

Gain reduction reduces level. Output compensation is a separate choice, and must not hide whether the dynamic change helps.

When a different instrument should trigger the reduction, continue with sidechain compression. The trigger path and the audio being processed have separate jobs.

For the next decision, compare compression with limiting when the task changes from dynamic shaping to peak restraint. For words, phrases and section balances, compare vocal automation with compression and check where each gain change sits in the signal path.


Browse all production guides → · Free browser audio tools