Compressor: the parameters

What it is and how it works

A compressor automatically lowers the volume when the signal exceeds a threshold. A detector measures the level (peak or RMS), a calculator decides how much reduction to apply (threshold, ratio, knee), an envelope sets how fast (attack, release), and a gain element (VCA, FET, opto, tube, or a digital multiplier) applies the reduction. Afterward the volume is recovered with the make-up gain. Result: less difference between loud and soft parts (reduced dynamics), a denser and more controlled sound, and, with the right attack and release, a change in the shape of the envelope (punch, sustain, groove).

Gain reduction (GR) above the threshold: GR = (level − threshold) × (1 − 1/ratio). Example: threshold −20 dB, ratio 4:1, peak at −8 dB (12 dB above) → output 3 dB above threshold, GR 9 dB.

Parameters one by one

• Threshold: the level beyond which the compressor works. Lower it until the GR meter shows the desired amount. In Fruity Compressor it ranges from 0 to −60 dB.
• Ratio: how many dB of input above the threshold correspond to 1 dB of output. 1.5–2:1 gentle (bus, already controlled vocal), 3–4:1 standard (vocal, bass, snare), 6–10:1 firm (peak control, effect), 20:1 and above = limiting. In Fruity Compressor it ranges from 0.4:1 (below 1:1 = expansion) to 30:1.
• Knee: the transition around the threshold. Hard knee = compression starts abruptly (precise, aggressive); soft knee = it starts gradually over a 6–15 dB zone (natural). In Fruity Compressor the knee is chosen with Type: Hard (0 dB), Medium (6 dB), Vintage (7 dB, a curve that eases off on strong peaks), Soft (15 dB).
• Attack: how quickly the reduction arrives after the signal exceeds the threshold (usually the time to reach about 63% of the final reduction; definitions vary between brands). Short (0.1–3 ms) = squashes transients; medium (5–15 ms) = lets the initial part through; long (20–50 ms) = obvious punch, compression works on the body. In Fruity Compressor 0–400 ms.
• Release: how quickly the reduction returns to zero when the signal drops. Short (20–80 ms) = more density, risk of distortion on the lows and “breathing”; medium (100–300 ms) = natural; long (400 ms–1.5 s) = soft leveling, risk that the compressor doesn't reopen between notes. Golden rule: the release must bring the GR indicator close to zero before the next hit and move in time with the song. In Fruity Compressor 1–4000 ms; the “/R” types enable automatic release (Transient Controlled Release).
• Make-up gain / output: recovers the lost level. Set it to get the same perceived volume as the bypass, otherwise you are judging the volume and not the compression. In Fruity Compressor ±30 dB.
• Detection: peak or RMS: peak reacts to instantaneous peaks (transient control); RMS reacts to the average level over a window (10–300 ms), closer to perception, more musical on vocals and buses. Fruity Limiter has the SUSTAIN control (0–1000 ms) that sets the RMS averaging window.
• Hold: keeps the reduction for a time after the signal has dropped (useful against distortion on bass with short releases).
• Lookahead: delays the audio signal relative to the detector so that compression starts before the transient (0.5–5 ms). Introduces latency; FL compensates for it. In Fruity Limiter the limiter's ATT also sets the look-ahead.
• Feed-forward / feedback: the detector reads the input (feed-forward, precise, modern) or the output (feedback, softer and more “vintage”, like the 1176 and LA-2A).
• Sidechain and sidechain filter: the detector can listen to another signal (ducking) or a filtered version of the same one: an HP at 80–150 Hz in the sidechain keeps kick and bass from making a bus “pump”.
• Stereo link: with 100% link, L and R are reduced together (the image stays stable); with low link each side compresses on its own (wider but the image shifts).
• Mix (dry/wet): internal parallel compression (see the dedicated guide).
• Range: limits the maximum reduction (e.g. never more than 6 dB).

How much to compress: reference values

Release in time

A quarter note lasts 60,000 / BPM ms. At 120 BPM: 1/4 = 500 ms, 1/8 = 250 ms, 1/16 = 125 ms. At 140 BPM: 1/4 ≈ 429 ms, 1/8 ≈ 214 ms. At 90 BPM: 1/4 ≈ 667 ms, 1/8 ≈ 333 ms. On drums and buses try a release equal to 1/8 or 1/16 and watch the GR “breathe” with the groove.

Order and serial compression

• Two light compressors in series (e.g. fast FET 3–5 dB + slow opto 2–3 dB) sound more natural than a single one doing 8–10 dB.
• Corrective EQ first (the compressor doesn't react to rumble or resonance), color EQ after.
• The de-esser usually after compression (compression raises sibilance), sometimes before as well.

Starting settings by genre

• Pop/K-pop: heavily compressed vocal (series of 2–3 compressors, 8–12 dB overall), mix bus 1–2 dB.
• Trap/rap: aggressive vocal (4:1, medium-fast attack, 6–8 dB), 808 lightly compressed and heavily saturated.
• Rock: drums with parallel compression and drum bus 3–4 dB; distorted guitars already compressed by the amp (little compression).
• Jazz/acoustic: 1–3 dB, ratio 2:1, slow attacks.
• EDM: compression and sidechain as a rhythmic element; bus with 2–3 dB of glue.

Common mistakes

• Judging with the wrong make-up (louder = seems better).
• Attack too short on drums and bass: flat sound with no punch.
• Release too short on bass: distortion (the compressor follows the individual waveform).
• Release too long: the compressor doesn't reopen, the sound “sinks” and the groove is lost.
• Compressing without a purpose: always ask yourself “what needs to change?” (level, punch, sustain, density).
• Threshold set in solo and never rechecked in the mix.

How to hear it

• Exaggerate first: ratio 10:1, GR 10–15 dB, then move attack and release and listen to how the transient changes; then return to sensible values.
• Watch the GR meter: it should “dance” in time, not stay still.
• Bypass at equal volume, in context, at low volume.