High-pass filter

Also known as: low-cut filter

On this page 7

A high-pass filter is a filter that passes frequencies above a cutoff point and progressively attenuates everything below it. In audio Mixing, it is the most common corrective tool for removing rumble, handling noise and unnecessary low end from tracks, and engineers often call it a low-cut filter because it describes the same action from the opposite side.

Within Audio production, the high-pass filter appears on mixing console channels, in every parametric EQ plug-in and on many microphones and preamps. Its value comes from a simple fact: most sources contain low-frequency energy that adds nothing musical but still consumes headroom and muddies the combined mix.

How a high-pass filter shapes the spectrum

A high-pass filter works by defining a cutoff frequency, conventionally the point where the output has fallen by 3 dB, and a slope that sets how quickly attenuation increases below that point. Above the cutoff the signal passes almost unchanged; below it, the level drops by a fixed number of decibels per octave.

The slope comes from the filter order. Each order contributes 6 dB per octave, so a second-order filter falls at 12 dB per octave and a fourth-order filter at 24 dB per octave. Steeper slopes remove unwanted content more decisively but introduce more phase shift and can add a slight resonant bump near the cutoff.

The following table lists the slopes found on typical mixing tools and how each one behaves in practice.

SlopeFilter orderCharacterTypical use
6 dB/octaveFirstVery gentle, naturalSubtle thinning of a full mix element
12 dB/octaveSecondModerate, musicalVocals, guitars and most console channels
18 dB/octaveThirdFirm, controlledDrum overheads and room microphones
24 dB/octaveFourthSteep, surgicalRemoving rumble close to the music
48 dB/octaveEighthBrick-wall likeCorrective work and sound design
Common high-pass filter slopes

Key parameters and filter designs

Beyond cutoff and slope, the filter design determines the exact shape of the curve near the cutoff. Analog prototypes such as Butterworth, Bessel and Linkwitz-Riley responses are reproduced in digital equalizers, and many plug-ins also offer a resonance or Q control that raises a peak just above the cutoff.

These parameters interact, so the table describes what each one controls and what goes wrong when it is set carelessly.

ParameterWhat it controlsRisk when misjudged
Cutoff frequencyWhere attenuation beginsThin, weak tone if set too high
SlopeHow fast the level fallsAudible phase shift if very steep
Resonance or QPeak near the cutoffBoomy bump or ringing
Filter designFlatness and transient behaviorOvershoot on kick and bass transients
Phase modeMinimum or linear phasePre-ringing with linear phase
High-pass filter parameters

Most digital equalizers use minimum-phase filters, which behave like analog circuits and shift phase near the cutoff. Linear-phase modes keep phase relationships intact but add latency and can produce pre-ringing, a faint smear before sharp transients, so they suit mastering and parallel paths better than everyday channel work.

A high-pass filter is one member of the EQ toolkit, and it is easily confused with tools that also reduce low frequencies. The difference lies in whether energy is removed entirely below a point or simply reduced in a region.

ToolEffect below the targetBest for
High-pass filterContinuous attenuation toward silenceRemoving content that should not be there
Low shelf cutFixed reduction at a set depthReducing warmth without removing it
Bell cutReduction in a narrow band onlyTaming a single resonance or boom
Dynamic EQ bandReduction only when level is highControlling proximity effect that varies
Low-pass filterLeaves the low end, cuts the highsDarkening or removing hiss
Tools that affect the low end

A high-pass filter also differs from a noise gate: the gate silences a signal when it falls below a level threshold, whereas the filter removes frequency content at all times. When a filter is used in a sidechain path, for example to stop a kick drum from triggering a compressor, it shapes only the detection signal and leaves the audible track untouched.

Choosing cutoff points in a mix

Choosing a cutoff means finding the highest frequency that removes clutter without audible thinning of the source. The usual method is to sweep the cutoff upward while listening in the context of the full mix, then back it off slightly once the tone starts to lose body.

The starting points below are common in practice, but every source, microphone and arrangement differs, so they should be treated as rough guides rather than rules.

SourceTypical cutoff rangePurpose
Male lead vocal70-100 HzRemoves plosives and stand rumble
Female lead vocal90-140 HzControls proximity effect
Acoustic guitar80-120 HzReduces body boom that masks bass
Drum overheads150-400 HzLeaves low end to kick and toms
Electric bass25-40 HzRemoves subsonic energy only
Kick drum20-35 HzCleans inaudible rumble
Typical starting cutoffs by source
  • Filter in context, not in solo, because a source that sounds thin alone often sits perfectly in the mix.
  • Use gentler slopes on melodic parts and steeper slopes on purely corrective tasks.
  • Apply a filter early in the signal chain, before compression, so the compressor does not react to rumble.
  • Consider a filter on the audio bus for groups of tracks rather than identical filters on every channel.

Common mistakes with high-pass filtering

The most frequent mistake is filtering every track by habit with high cutoffs, which leaves a mix that measures clean but sounds small and lacks warmth. Low-mid energy between 100 and 250 Hz carries the body of many instruments, and removing it everywhere weakens the whole arrangement.

  • Thin vocals and guitars, caused by cutoffs set while soloed rather than in the mix.
  • Lost punch on kick and bass, when steep filters with resonance smear transient energy.
  • Phase problems on multi-microphone sources, if different filters are used on microphones that capture the same drum.
  • Uncontrolled low end on playback systems, when no filter removes subsonic content below 20-30 Hz.

Filtering cannot repair a poor recording either. A filter removes rumble and proximity build-up, but it does not fix room resonances above the cutoff, masking between instruments or a badly chosen microphone position, which need other tools such as bell cuts, panning or arrangement changes.

A plain-language summary

A high-pass filter lets the higher sounds through and turns down the deepest sounds. Think of it as a sieve that keeps the useful part of a recording and catches the low rumble you do not need, such as traffic noise, a bumped microphone stand or the hum of an air conditioner.

In a song, many instruments carry low sounds that no one really hears but that still pile up and make the mix muddy. The key idea: trimming those unneeded lows makes room for the bass and kick drum, so everything sounds clearer. Trim too much, though, and instruments start to sound thin.

An example from practice

A mixing engineer is working on an acoustic folk recording with a vocal, two acoustic guitars, upright bass and a pair of room microphones. Played together, the low end sounds boomy and the bass loses definition, although no single track seems faulty on its own.

The engineer applies a 12 dB per octave filter at 90 Hz on the vocal, about 110 Hz on each guitar and 180 Hz on the room microphones, setting each cutoff while the full mix plays. The upright bass receives only a gentle 30 Hz filter to remove subsonic stage rumble. The outcome is a clearer low end in which the bass notes are distinct, the mix gains about 2 dB of usable headroom and the guitars still sound full.