Compression Ratio: What 4:1 Actually Means in Practice

Here's a pattern I see constantly: a producer slaps a compressor on a bus, sets the ratio to 4:1, adds some makeup gain, A/Bs against the dry signal, and says "yeah, that's better." Of course it's better. It's louder. Louder always sounds better. This is not an opinion, it is a well-documented psychoacoustic bias. A 1 dB increase in level is enough to make most people prefer one version over another, even when nothing else changed.

So when you compress a signal and then add 4 dB of makeup gain, you're not hearing the compression. You're hearing the volume difference. Every judgment you make from that point is contaminated.

This is the single most common compression mistake, and almost nobody talks about it because the result "sounds good" and that feels like enough.

The ratio describes how much the compressor reduces signal that exceeds the threshold. That's it. A 4:1 ratio means that for every 4 dB the input goes above the threshold, only 1 dB comes out the other side.

The formula is:

`output = threshold + (input - threshold) / ratio`

This only applies to signal above the threshold. Everything below the threshold passes through unchanged (assuming a hard knee, which we'll get to).

Here's a concrete example. Threshold is set to -20 dB. Input signal hits -8 dB. Ratio is 4:1.

`output = -20 + (-8 - (-20)) / 4`

`output = -20 + 12 / 4`

`output = -20 + 3`

`output = -17 dB`

The input was -8 dB. The output is -17 dB. That's 9 dB of gain reduction. The signal went 12 dB over the threshold, and the compressor squashed 9 of those decibels away, letting only 3 through.

Now the producer adds 9 dB of makeup gain to bring the peak back to -8 dB. The loud parts are the same level, but the quiet parts (which were below threshold and weren't compressed) are now 9 dB louder. That's how compression reduces dynamic range: it doesn't pull the loud stuff down permanently, it pulls the loud stuff down and then you push everything back up.

This table shows what happens to a signal that peaks at -8 dB with a threshold of -20 dB across different ratios:

Notice how the jump from 2:1 to 4:1 is large (6 dB to 9 dB of reduction) but the jump from 8:1 to 20:1 is small (10.5 to 11.4 dB). There are diminishing returns as you increase the ratio. At infinity:1, the output never exceeds the threshold, which is what a limiter does. A limiter is just a compressor with an infinite (or very high) ratio.

For most mixing work, ratios between 2:1 and 4:1 handle the job. I rarely go above 6:1 on individual tracks unless I'm going for an obvious effect. Anything above 10:1 starts behaving like limiting, and limiting on a vocal or guitar track usually sounds like you're strangling it.

Hard knee means the compressor kicks in immediately at the threshold. One dB below threshold, no compression. One dB above, full ratio applied. It's abrupt.

Soft knee means the compressor gradually increases the ratio as the signal approaches and passes the threshold. The compression starts a few dB before the threshold and reaches the full ratio a few dB after. This sounds more transparent on most sources because the onset of compression isn't sudden.

I default to soft knee on vocals and buses, hard knee on drums when I want the compressor to grab hard. There's no rule here though. Use your ears.

This is the part that matters more than ratio, honestly.

Attack time is how long the compressor waits before it starts compressing after the signal crosses the threshold. A slow attack (30 ms or more) lets the initial transient of a drum hit pass through uncompressed, then clamps down on the sustain. This preserves punch. A fast attack (under 5 ms) catches the transient itself and rounds it off. On a snare, fast attack makes it thud instead of crack. On a vocal, fast attack can make consonants disappear.

Release time is how long the compressor takes to let go after the signal drops below threshold. This is where things get musical, or ugly. Too fast a release and the compressor pumps: it grabs, lets go, grabs again, creating a stuttering distortion that sounds like the track is breathing. Too slow a release and the compressor never recovers between transients, so it stays compressed through quiet sections and kills all the dynamics you were trying to control.

The right release time depends on the tempo and the rhythmic content. On a drum bus at 120 BPM, I'll usually start around 100 to 200 ms and adjust until the compressor "breathes" with the groove. You can actually feel when the release locks into the tempo. The mix starts bouncing. Get it wrong and it fights the rhythm.

One thing that helped me: watch the gain reduction meter while adjusting release. The needle (or bar) should return to zero (or close to it) before the next transient hits. If it's still showing 4 dB of reduction when the next kick arrives, your release is too slow.

This is the most honest thing you can do with a compressor. Takes about five minutes and will recalibrate your relationship with compression.

Sometimes it does sound better. The drums feel more cohesive, the balance between kick and snare is more consistent. But sometimes it sounds worse: flat, lifeless, like the air went out of the performance. You would never notice this without level-matching because the louder version always wins the comparison.

I've had sessions where I removed compression from half the tracks after doing this test. The mix opened up. That's not an argument against compression. It's an argument for actually hearing what compression does versus what volume does.

The mistake is judging the compressed signal at a louder level than the dry signal and calling it an improvement. Makeup gain is not a feature of compression. It is a compensation for the level loss that compression creates. If you treat it as "the part that makes the compressor sound good," you're fooling yourself.

Every dB of makeup gain is a dB of bias in your judgment. Match the levels. Then decide.

Ratio is how aggressively the compressor squashes signal above the threshold. Lower ratios (2:1, 3:1) are gentle. Higher ratios (8:1 and up) approach limiting. The math is simple but the musical result depends almost entirely on attack and release, which control how the compressor interacts with the rhythm and transient content of the source.

Before you decide if compression is helping your track, level-match the output to the input. Remove loudness from the equation. What's left is the actual sonic effect of compression: the tonal change, the transient reshaping, the dynamic control. Sometimes that effect is exactly what you need. Sometimes it's taking away more than it's giving. You can't know which until you remove the loudness bias.