Silence and Logarithms
Before we go deeper into dynamics, a quick detour into the math. Decibels use a logarithmic scale — and logarithmic scales are how humans perceive the world. A sound that’s twice the physical intensity doesn’t sound “twice as loud” — it sounds slightly louder. To sound twice as loud, you need roughly ten times the power (about 10 dB).
This matters for compression because the numbers on the gain reduction meter don’t behave linearly. The difference between 0 dB and 3 dB of gain reduction is subtle — your ear might not notice it. The difference between 3 dB and 6 dB is more noticeable. The difference between 6 dB and 12 dB is dramatic. Each additional dB of compression makes a proportionally bigger audible impact.
This is why conservative compression (2-4 dB of gain reduction) often sounds transparent and heavy compression (10+ dB) sounds obviously processed. The logarithmic scale means the first few dB of compression are nearly invisible, and each subsequent dB becomes increasingly audible.
Another way to think about it: you can cut a lemon in half, then cut that half in half, then cut that half in half — and you never end up with nothing. That’s a logarithmic scale. The fader on a mixer works the same way: it goes to negative infinity, not zero, because you can always cut in half again. Silence isn’t a number on this scale — it’s a direction you approach but never reach.
The Knee
The knee controls how the compressor transitions from no compression to full compression around the threshold.
Hard knee: The compressor engages abruptly at the threshold. Below the threshold, no compression. Above it, the full ratio kicks in immediately. This creates a sharp, defined compression character — you can hear the compressor “grab” the signal. Hard knee is good for controlling dynamics precisely, and it’s the default on most compressors.
Soft knee: The compressor eases into compression gradually around the threshold. Compression begins slightly below the threshold and reaches the full ratio slightly above it. The transition is smooth and less audible. Soft knee is good for transparent, musical compression where you don’t want to hear the compressor working.
Think of the threshold like the height requirement at an amusement park — below this line, the compressor doesn’t care about you. The knee widens that line. Instead of a hard cutoff, there’s a transition zone where signals near the threshold get progressively more compression. This matters because real signals — a snare hit, a vocal phrase — don’t all cross the threshold at the same level. Some hits are barely above, some are well above. A soft knee handles that inconsistency gracefully.
The knee is subtle — many beginners don’t hear the difference at first. Try an extreme comparison: hard knee with fast attack on a drum loop. You’ll hear the compressor clamp down on each hit. Now switch to soft knee with the same settings. The compression is still happening, but the onset is gentler, less pronounced.
Two transfer function graphs: hard knee (sharp angle at threshold) vs soft knee (gradual curve around threshold).
Attack and Release: Going Deeper
Chapter 16 introduced attack as “how fast the compressor reacts” and release as “how fast it lets go.” That’s accurate but incomplete. The creative power of compression lives in the nuance of these two controls.
Attack as a Transient Shaper
A fast attack catches the transient — the initial burst of energy at the start of a note. Catching the transient means compressing it, which means reducing the punch. The body of the sound stays relatively unchanged, but the leading edge gets softened.
A slow attack misses the transient — the compressor doesn’t react fast enough to catch it. The transient punches through at full level, and the compressor only engages on the sustained body of the sound. The result: the body gets compressed (leveled out), but the transient keeps its impact.
This is why attack time has more effect on the feel of a sound than any other compressor parameter. A drum with a fast attack sounds controlled and smooth. The same drum with a slow attack sounds punchy and aggressive — even though the compressor is doing similar amounts of gain reduction.
Drum hit envelope showing fast attack compressor catching and flattening the transient vs slow attack letting transient through, compressing only the body.
Release and Groove
Release time interacts with the rhythmic content of the music. If the release is too fast, the compressor resets between hits and you hear the level “breathe” — pumping up between notes. If the release is too slow, the compressor never fully resets and the sustained compression dulls the dynamics. When the compressor breathes — a subtler, more musical version of pumping — the gain reduction follows the song’s rhythm like a living thing.
The sweet spot depends on the tempo and the rhythmic density. On a slow ballad, a longer release can feel smooth and natural. On a fast funk track, a shorter release lets the compressor follow the groove. Some engineers set the release by ear to match the tempo — they adjust until the compression feels like it’s “in time” with the music.
Gain reduction meter showing 'tickle' technique: three states — release too fast (pumping), too slow (never resets), sweet spot (returns to zero just before next beat).
When Pumping Is a Feature
Pumping — the audible rise and fall of the signal as the compressor engages and releases — is usually considered a problem. But in some contexts, it’s the point.
The classic example: sidechain compression on a synth pad triggered by the kick drum (more on sidechaining in Chapter 20). Every kick hit pushes the pad down; between kicks, the pad pumps back up. That rhythmic push-and-pull is the sound of modern electronic music. It’s compression as a creative effect, not a corrective one.
Even without sidechaining, a compressor with the “wrong” release time can create a rhythmic feel that works for the track. The pumping of a bus compressor on a drum mix — slightly too fast, audibly breathing — can add energy and excitement. Whether it’s a problem or a feature depends entirely on context.
Beyond level control, compression can:
Add sustain. Compress a guitar and the quiet tail of each note gets pushed up relative to the attack. Notes ring longer. This is why virtually every electric guitar signal chain includes some form of compression (often from the amp itself, as tube distortion is a form of dynamic compression).
Change the envelope. Fast attack + medium release on a snare transforms a sharp crack into a rounder, fatter hit. Slow attack + fast release preserves the crack and tightens the body. You’re reshaping the sound’s ADSR — the same concept from synthesis (Chapter 4), applied to recorded audio.
Create density. Moderate compression on a bus — drums, vocals, the entire mix — reduces the dynamic range and makes everything feel more “glued together.” The individual elements lose some independence but gain cohesion. This is “bus compression” or “glue compression,” and it’s one of the most common uses of compression in professional mixing.
What to Practice
- Compare hard and soft knee. Compress a vocal or drum track with a hard knee, then switch to soft knee with identical settings. Listen for the difference in how the compression engages. The soft knee should feel smoother, the hard knee more defined.
- Shape a drum with attack time. Put a compressor on a snare track. Set a moderate ratio and pull the threshold until you see 4-6 dB of gain reduction. Now sweep the attack from fastest to slowest. Hear how the snare goes from soft and controlled (fast attack) to punchy and sharp (slow attack). Find the point where the transient sits right for the song.
- Listen for the release. On the same compressor, set a very fast release and listen for the breathing/pumping between hits. Now slow the release until the pumping disappears. Keep going until the compression feels “sluggish.” The musical sweet spot is somewhere in the middle — where the compressor follows the rhythm naturally.
