The short guide
Understand the idea
Audio compression, or dynamic range compression, turns down the loudest parts of a sound so the gap between soft and loud gets smaller. It is not file compression: making an MP3 or ZIP smaller is a different idea. A compressor watches the level, and when the sound goes above a threshold it reduces it by a ratio, reacting at the speed set by attack and letting go at the speed set by release.
What you can explore
- Compare the same soft and loud pulse as Original and Compressed.
- Set a threshold and ratio, and explain which way the numbers move.
- Tell the compressor’s attack and release apart from a note’s own shape.
A worked example
One pulse, soft and loud, with a −24 dB, 4:1 compressor
A single triangle note, A4, plays on beats 1 to 4 of one looping bar at a fixed 80 BPM: soft, loud, soft, loud. Each note is held for 412.5 ms of each 750 ms beat. The pulse is identical whether Play uses Original or Compressed; only the compressor changes. The worked example uses a −24 dB threshold, a 4:1 ratio, 10 ms attack and 180 ms release. The numbers below are illustrative input levels on an ideal hard-knee curve, not measurements of the pulse.
- Illustrative −18 dB input
- 6 dB above the threshold. At 4:1 those 6 dB become 1.5 dB, so it comes out at −22.5 dB.
- Illustrative −12 dB input
- 12 dB above the threshold becomes 3 dB, so it comes out at −21 dB.
- The gap
- A 6 dB difference between the two inputs becomes 1.5 dB, before automatic makeup and time response.
- Below the threshold
- An illustrative −30 dB input is below −24 dB, so the ideal curve leaves it unchanged.
Threshold and ratio: where and how firmly
Levels here are negative numbers in decibels, where 0 dB is the reference at the top of this illustration. A lower, more negative threshold such as −30 dB catches more of the sound; a higher one such as −6 dB catches less. In the ideal static curve, sound at or below the threshold is unchanged. During playback, earlier reduction can continue through release, and automatic makeup also changes the level.
The ratio sets how firmly sound above the threshold is turned down. At 4:1, every 4 dB above the threshold becomes 1 dB. At 8:1 the reduction is stronger. At 1:1 there is no dynamic reduction at all: sound above the threshold passes through unchanged, which is a useful comparison rather than a mistake.
Attack, release and automatic makeup
Attack is how quickly the compressor reacts once the sound goes above the threshold, and release is how quickly it lets go when the sound drops. Attack is not a delay before anything happens, and release is not a countdown that ends exactly on time. They are also separate from the note’s own shape, the ADSR fade-in and fade-out, which never changes in this lesson.
The compressor in your browser adds automatic makeup gain, and both versions share the same fixed output trim, so Original and Compressed are not loudness matched. One may sound louder; louder is not better. Compare by character: is the gap between soft and loud notes smaller, and do the loud notes keep their punch at the start?
Finishing on Original, at 1:1 or with a high threshold is valid, and the review describes what you actually chose. Mixing Stems explores whole-part levels and EQ explores frequency balance. Compare this level-responsive compressor with the Sidechain lesson’s kick-timed volume dips: that lesson uses planned ducking rather than a level detector. Sound Design shapes a note’s envelope; Automation makes planned changes over time.
Common questions
Is audio compression the same as file compression?
No. Audio, or dynamic range, compression changes how loud parts of a sound are. File compression, like MP3 or ZIP, makes a file smaller. This lesson is only about the first.
Why does a lower threshold catch more?
Thresholds are negative numbers. Moving from −18 dB down to −30 dB puts the line lower, so more of the sound is above it and gets turned down.
Why can Compressed sound louder or quieter than Original?
The compressor adds automatic makeup gain after turning the peaks down, and the versions are not loudness matched. A louder version is not a better one, so listen for the gap between soft and loud notes instead.
What can this lesson not do?
It uses one fixed pulse and does not record, upload or export audio. The pictures illustrate your settings rather than measuring live sound. The recipe download is a text description of your settings, not a sound file. Your work is saved only in this browser.
Published by TunePal. Interactive examples for beginner music producers.