Live Sound Compression: Threshold, Ratio, Attack, and Release

TL;DR — A compressor reduces level according to a threshold, ratio, and timing controls, making selected peaks or changes in dynamics easier to manage. It does not create headroom before an overloaded preamp or replace consistent performance technique. Set gain first, begin gently, listen in the full mix, compare at a similar perceived loudness, and document only settings that are verified and essential to the show.

Table of contents

1. What a compressor does
2. What the controls mean
3. Where compression helps
4. How to set a compressor
5. How to document compression
6. FAQ

What a compressor does

A compressor is a dynamics processor that changes gain when the signal meets conditions set by its controls. In a conventional downward compressor, level above the threshold is reduced according to the ratio. The result can make peaks easier to contain, bring a source forward in a mix, or shape how quickly an instrument feels.

Compression does not simply make every quiet sound loud. Makeup gain may raise the processed signal after gain reduction, which can also raise noise, spill, and feedback risk. A compressor also cannot repair clipping that happened at a source, preamp, converter, or earlier device. Establish a healthy signal path with the gain staging guide before using dynamics control.

Compression is different from limiting, although the controls are related. A limiter normally uses a high ratio and fast behaviour to restrain peaks near a ceiling. Exact definitions and timing vary by device, so check the console or processor documentation rather than treating a ratio value as a universal operating mode.

What the controls mean

Threshold and ratio

The threshold is the level region where compression begins to act. Lowering it generally causes more of the signal to be processed. The ratio describes how changes above the threshold are reduced: at 2:1, an ideal 2 dB increase at the detector produces about a 1 dB increase at the output, before makeup gain. Real processors may use a soft knee, program-dependent response, or other behaviour that makes the transition less literal.

Attack and release

Attack controls how quickly gain reduction develops after the detector responds. A faster attack can restrain a transient, while a slower attack can let more of the initial impact pass. Release controls how quickly gain returns after the signal falls back. Release that is too fast for the material may sound busy or distorted; release that is too slow may hold the source down into the next phrase.

There is no correct attack or release time based only on an instrument label. Tempo, performance, microphone, arrangement, compressor design, and the desired result all matter.

Knee, detector, and sidechain

A knee determines how gradually the ratio takes effect around the threshold. The detector decides what the compressor responds to; peak, RMS-like, and program-dependent modes can react differently. A sidechain filter changes what the detector hears without necessarily filtering the audible channel. An external sidechain can let one signal control another, commonly called ducking.

Gain reduction and makeup gain

The gain-reduction meter shows how much the processor is turning the signal down according to its own detector. It is evidence, not a target. Makeup or output gain compensates for level loss, but louder comparisons often sound subjectively better. Compare bypass and processed states at a similar perceived loudness before deciding that compression improved the source.

Where compression helps

Vocals and speech

Compression can contain strong syllables and keep intelligibility more consistent, especially when a performer moves or changes intensity. It also raises the audibility of breaths, handling noise, cymbal spill, and stage wash when makeup gain is added. Improve microphone technique, placement, and monitor geometry first; compression can reduce gain before feedback if it leads the operator to add more level.

Bass, guitars, keys, and playback

Compression may even out notes, manage patches with different output levels, or shape sustain. However, pedals, amplifiers, keyboards, playback rigs, and modelers may already apply significant dynamics processing. Listen to the complete source chain before adding another stage, and preserve intentional contrasts between songs or patches.

Drums and percussion

Attack and release can change transient impact, sustain, and spill substantially. A setting that makes a close microphone sound exciting alone may make the whole kit smaller or bring unwanted cymbals forward. Check polarity, timing, gates, and bus processing when several drum microphones combine.

Buses, matrices, streams, and recordings

Bus compression affects every source routed to that bus. Mix-bus processing may help control a familiar touring mix, while separate stream or recording compression can serve a different listening environment. Confirm the routing and tap points: a processor inserted for one destination should not silently alter monitors or another deliverable. See aux vs matrix routing for destination planning.

How to set a compressor

1. Define the job

Choose one audible problem or creative goal: restrain occasional peaks, steady speech, shape a drum envelope, or control a destination. “Make it better” is not a testable objective.

2. Verify gain and routing

Confirm the source, input, patch, preamp, insert position, channel routing, and destinations. Make sure nothing overloads before the compressor. If the processor is in a saved show file, confirm that it still matches the current performer and equipment.

3. Start from a neutral state

Bypass the compressor or reset it to a known state. Remove unnecessary makeup gain. Choose a moderate ratio and a threshold that only engages on the part of the performance you intend to control; exact values must follow the actual signal level and processor scale.

4. Set threshold and ratio together

Have the performer deliver representative quiet, normal, and loud material. Lower the threshold until the intended moments cause gain reduction, then adjust ratio for the required amount of control. Avoid forcing a meter to display a memorized number.

5. Tune attack and release by listening

Shorten attack if problematic peaks pass too freely; lengthen it if the source loses useful impact. Shorten release if gain stays reduced too long; lengthen it if the recovery sounds nervous, pumps, or distorts. Change one control at a time and listen in context.

6. Match level and compare

Use makeup gain only after the compression behaviour is useful. Compare bypass and processed states at similar perceived loudness. Listen for tone, envelope, noise, spill, feedback margin, and whether the musical contrast still serves the arrangement.

7. Check every destination

Verify FOH, wedges, IEMs, fills, stream, recording, and any matrix that receives the processed signal. A vocal setting that helps the audience mix may be distracting in an IEM; destination-specific processing may be clearer than one compromise.

8. Save and recheck

Save verified settings in the appropriate show file and test again after changes to performer, microphone, source output, console, patch, or venue. A threshold is tied to signal level, so even a good previous setting can become irrelevant when gain structure changes.

How to document compression

Most riders should specify the source, routing, responsibility, and desired outcome rather than prescribe channel processing for an unfamiliar room. Detailed values belong in the handoff only when they are a rehearsed part of the production, a show-file starting point, or an acceptance item that another engineer must verify.

Record the channel or bus, processor location, purpose, responsibility, linked channels, affected destinations, and the condition that requires a new check. If exact threshold, ratio, attack, release, knee, or sidechain values matter, name the processor or show file because control scales and behaviour are not interchangeable.

In Techrider.live, keep source names aligned between the stage plot and input list, add concise notes only where processing affects the handoff, and invite the responsible engineer to edit and save the same Rider. Inspect history when a starting point changes, then export a dated PDF when the venue needs an offline copy.

Compression handoff checklist

• Define the audible purpose.
• Verify gain and routing before processing.
• Identify existing compression in source devices or show files.
• Test representative quiet and loud material.
• Compare bypass at a similar perceived loudness.
• Check noise, spill, feedback margin, and musical dynamics.
• Verify every affected output and monitor mix.
• Document only repeatable requirements and recheck triggers.

FAQ

What does compression do in live sound?

It changes gain in response to signal level and timing controls. Engineers use it to manage selected peaks, reduce dynamic variation, or shape an envelope. Compression does not fix overload before the processor, poor microphone technique, or an unclear arrangement.

How do you set a compressor for live vocals?

Set input gain first, hear the singer's full working range, and begin with a moderate ratio. Move the threshold until only the intended phrases trigger reduction, then adjust attack and release so words remain natural. Match perceived level for bypass comparisons and verify FOH and monitors separately.

What do threshold and ratio mean on a compressor?

Threshold defines where compression begins to act; ratio defines how strongly level changes above that region are reduced. Lower threshold usually means more of the signal is compressed, while higher ratio means stronger control. Knee and detector design affect the exact transition.

What attack and release should I use on a compressor?

No universal times fit every source. Attack must balance peak control against useful transient impact. Release must recover soon enough for the next phrase without pumping, distortion, or audible level movement. Set both with representative material in the full mix.

Should compressor settings be in a tech rider?

Usually only when they are verified, repeatable, and important to the production. Identify the source, processor or show file, purpose, responsibility, destinations, and recheck trigger. For ordinary house mixing, clear source and routing information is more useful than copied preset values.

Make dynamics decisions reproducible

Build a reusable Rider with Techrider.live, align every source with its input-list row, and record the few processing constraints that genuinely affect the show. A clear purpose and recheck trigger travel better than unexplained compressor numbers.