Sidechain compression is a mixing technique in which a compressor reduces the level of one signal according to the level of a different control signal, or of a filtered version of the same signal, fed into its detector. In audio mixing, it is used to duck music under speech, to let a kick drum cut through a bass line, to create rhythmic pumping and to make compression respond to selected frequencies only.
The word sidechain refers to the detector path inside every compressor. Normally that path listens to the signal being compressed. When the path is fed from somewhere else, the compressor becomes a tool for interaction between tracks, which is why the technique is also called ducking or key-input compression.
How the sidechain controls gain reduction
A compressor contains two parallel paths. The audio path carries the signal through a gain element, such as a voltage-controlled amplifier or a digital multiplier. The sidechain path feeds a level detector that measures how loud its input is, compares it with the threshold and calculates how much gain reduction to apply according to the ratio, attack and release settings.
In ordinary compression, both paths start from the same signal. In external sidechain compression, the detector receives a separate key signal, for example a voice track, while the gain element still processes the music. When the voice exceeds the threshold, the music is turned down; when the voice stops, the music recovers at the release rate.
An internal sidechain filter is a related variant. The detector still listens to the signal being compressed, but through an equalizer or filter. Removing low frequencies from the detector, for instance, prevents a bass-heavy mix from triggering heavy compression, while boosting a narrow band around sibilance turns the compressor into a de-esser.
Types of sidechain processing
Sidechain processing takes several forms that share the same control principle but serve different goals. The table distinguishes the main variants used in mixing.
| Form | Key signal | Processed signal | Typical purpose |
|---|---|---|---|
| External ducking | Voice or another instrument | Music, ambience or competing track | Keep the key source intelligible |
| Kick and bass sidechain | Kick drum | Bass or synth bass | Separate low-frequency energy in time |
| Rhythmic pumping | Steady four-on-the-floor trigger | Pads, chords or full mix | Audible breathing effect in dance music |
| Filtered internal sidechain | Equalized copy of the same signal | The original signal | Frequency-aware compression |
| Sidechain gating | Rhythmic or percussive track | Sustained sound through a gate | Impose rhythm on a pad or ambience |
Related tools use the same idea with different processors. A noise gate with a key input opens only when the trigger is present, a de-esser is a compressor keyed by a high-frequency band, and dynamic or multiband equalizers can apply sidechain-controlled reduction to a single frequency region instead of the whole signal.
Key parameters and typical settings
The audible character of sidechain compression is shaped mostly by attack, release and depth of gain reduction. The ranges below are practical starting points, not fixed rules, and depend on tempo and material.
| Parameter | Dialogue ducking | Kick and bass | Pumping effect |
|---|---|---|---|
| Ratio | About 2:1 to 4:1 | About 4:1 to 8:1 | High, often 8:1 or more |
| Attack | About 5-20 ms | About 0.1-5 ms | Fast, near 0.1-1 ms |
| Release | About 200-800 ms | About 50-150 ms | Timed to the tempo, often 1/8 or 1/4 note |
| Gain reduction | About 3-10 dB | About 2-6 dB | 6 dB or more, clearly audible |
Release is the most critical setting. A release that is too short lets the ducked signal jump back between syllables or beats, which sounds nervous; a release that is too long leaves it suppressed and creates a hole in the mix. For rhythmic effects, matching the release to the song tempo makes the pumping sit in the groove.
Some compressors offer look-ahead, which delays the audio path slightly so that gain reduction can start before the key transient arrives. This makes ducking smoother but adds latency, which the host must compensate.
Sidechain compression compared with alternatives
Sidechain compression is not the only way to make room for one sound under another. Choosing between methods depends on whether the conflict is constant or momentary and whether it spans the whole spectrum or a narrow band.
| Method | How it works | Best when |
|---|---|---|
| Static EQ cut | Permanent reduction of an overlapping band | Overlap is constant throughout the mix |
| Sidechain compression | Whole-signal reduction only while the key plays | Conflict is momentary and broadband |
| Dynamic EQ with sidechain | Band-limited reduction only while the key plays | Conflict is momentary and narrow |
| Volume automation | Manual level changes drawn over time | Precise, irregular moments in post-production |
| Arrangement change | Different register or timing for parts | The conflict is musical, not technical |
In broadcast and podcast work, dedicated automatic ducking features apply the same principle with simplified controls. They are fast to set up, but careful automation usually sounds more natural for finished programs.
Pitfalls when using sidechain compression
- Overly deep ducking makes music disappear audibly under every word, drawing attention to the processing instead of the speech.
- A key signal that contains bleed or background noise triggers gain reduction unpredictably, so gate or filter the key first.
- Applying pumping to the whole mix can reduce perceived loudness and punch when the effect is not intended stylistically.
- Routing errors are common: the compressor must receive the key on its sidechain input, and the key track may need a pre-fader send so that muting it does not stop the ducking.
As with any dynamic process, judgment by ear in context is essential. Soloing the ducked track can mislead, because the effect is designed to be heard together with the key.
A plain-language summary
Sidechain compression lets one sound turn another sound down automatically. Think of a radio host speaking over music. Whenever the host talks, the music politely gets quieter. When the host stops, it comes back up.
Music producers use the same trick on purpose in dance music. Every kick drum briefly pushes the other sounds down. That creates the familiar breathing or pumping rhythm. The idea is simple: one track listens, another track reacts.
An example from practice
A video editor mixes a product explainer with a voice-over and an energetic music bed. With the music at a fixed level, several words are masked during dense passages. The editor inserts a compressor on the music track and feeds its sidechain from the voice-over.
With a ratio of about 3:1, an attack of 10 ms and a release of 400 ms, the music drops by about 6 dB whenever the narrator speaks and swells back during pauses. The narration becomes clearly intelligible on phone speakers, and the music still feels present between sentences without any manual automation.