Guide · Equalisation
High-Pass Filters: What They Do and When to Use Them
Almost every track you mix has a low end problem: a rumble, a hum, a kick bleeding into a microphone, or simply energy in the bottom octave that nothing else needs. A high-pass filter is the cure, and it is also the easiest way to ruin a mix.
What the Filter Actually Does
A high-pass filter is a corner and a slope. The corner is the frequency where the filter starts to act, usually given in hertz. The slope is how quickly it gets there, given in decibels per octave. A 12 dB per octave filter at 100 Hz is doing almost nothing at 200 Hz and has removed 12 dB by 50 Hz. A 24 dB per octave filter at 100 Hz has removed 24 dB by 50 Hz and is well down by 25 Hz.
This is worth understanding properly, because "set the high-pass to 100 Hz" does not mean "remove everything under 100 Hz". It means "start removing at 100 Hz, and keep going". The difference between those two readings is the difference between a clean mix and a thin one.
Where to Set the Corner
The rule is: set the corner below the lowest frequency you want to keep, and above the highest frequency you want to remove. The gap between those two numbers is where the work is.
| Source | Typical corner | Why |
|---|---|---|
| Kick drum | 30–40 Hz | Keeps the fundamental, removes the subsonic thump that eats headroom. |
| Snare | 80–120 Hz | The body lives above 150 Hz; below that is usually the shell and the room. |
| Hi-hat | 300–400 Hz | Anything lower is bleed from the other drums. |
| Lead vocal | 80–100 Hz | Removes proximity effect and handling noise without touching the note. |
| Backing vocals | 100–150 Hz | Often thinner and duller than the lead, so the corner can be higher. |
| Acoustic guitar | 70–90 Hz | Removes the thump of the body hitting the player's chest. |
| Bass guitar | 30–40 Hz | Only removes rumble. Cutting higher steals the note. |
| Full mix bus | 20–30 Hz | Removes DC offset and subsonic junk that is not music. |
These are starting points, not rules. A 55 Hz kick and a 40 Hz kick want different corners. The number that matters is the lowest frequency you actually intend to keep.
The Mistake That Makes Them Sound Thin
Nearly every complaint about high-pass filters comes from the same error: setting the corner too high, in the belief that a higher setting removes more.
It does not remove more. It removes less, and it removes the part you wanted. A vocal with its lowest note at 80 Hz, filtered at 200 Hz, loses most of that note. The fundamental is the note. Everything above it is the tone of the note. Cut the fundamental and you have not made the vocal cleaner, you have made it quieter and thinner, and you have probably reached for a compressor to compensate.
A better method: set the filter low, where it does nothing audible, then raise it in small steps while listening. Stop at the point where the unwanted energy is gone. If the sound thins, you have gone too far. That is the test, and it works in every case.
Slope: 12 dB or 24 dB
A 12 dB per octave slope is gentle. It suits removing low-frequency noise that is not particularly loud, and it is the safer default because it is less likely to affect the sound you want.
A 24 dB per octave slope is for problems that a gentle slope cannot fix: a resonant frequency in a room, a DC offset, or a subsonic rumble that is loud enough to be pushing your mix bus around. It also works closer to the music before it becomes audible, which lets you set a higher corner without thinning the source.
Beyond 24 dB per octave you are into brick-wall territory. It is occasionally the right answer for a problematic noise, but it is also how a filter starts doing strange things to phase, and it is not something to reach for by default.
What You Will Hear
Two things, and they are easy to confuse.
The first is less low end, which is the point. If the low end was carrying something you wanted, that something is now gone.
The second is more apparent clarity, and this is the one people do not expect. Removing a rumble takes energy out of the band that the ear was using to judge the size and weight of the sound. The source sounds tighter and more defined, not thinner. A mix that has been high-passed across the board usually sounds bigger, because the low end is no longer being smeared across everything.
This is also why high-passing every channel is not a substitute for arranging the mix. If two instruments are fighting in the same octave, cutting both of them does not separate them. It just makes them quieter. Space comes from arrangement and from carving with EQ, not from filtering.
Order of Operations
Where the filter sits in the chain matters more than most people expect.
- Before compression, usually. A compressor reacts to everything it is fed. If rumble is getting into the compressor, the compressor is ducking the whole signal because of it, and you are fighting a problem you created.
- After distortion, if you are using it. If a source is saturated, a high-pass before the saturator removes low-end energy that would otherwise have driven the distortion into the top end and changed the character.
- On the mix bus last, and gently. A final high-pass at 20 to 30 Hz with a 12 dB slope removes DC offset and keeps subsonic content out of the master bus. Do not filter the bus much harder than that, or you will lose the low end of the whole track because of a problem in one channel.
How to Check Your Work
The analyser will show you the result. Load a source, look at the far left of the spectrum, and find the lowest frequency with meaningful energy. That is your corner frequency. Set the filter just below it and nothing else about the sound should change.
If the energy runs all the way down to 20 Hz, the source has a problem: a DC offset, a handling noise, or a resonant room tone. A filter will remove it, but it is worth knowing that it was there.
Questions
What does a high-pass filter do?
It removes everything below a chosen frequency and passes everything above it. On audio it is a steep cut of the low end, usually set with a corner frequency and a slope of 12 or 24 dB per octave.
Where should I set a high-pass filter?
Low enough to remove the problem and high enough to keep the fundamental. For most vocals 80 to 100 Hz is right, for bass 30 to 40 Hz, for acoustic guitar around 80 Hz, and for a full mix bus 20 to 30 Hz.
Will a high-pass filter make my mix thin?
Only if it is set too high. A filter that removes the part of an instrument you actually hear will thin it out. Start low, raise it in small steps, and stop when the unwanted energy is gone rather than when the sound starts to thin.
Is 12 dB per octave enough?
For rumble and stage noise, yes. For a steep problem such as a resonant sub-bass bleed or a proximity effect, 24 dB per octave is better, because it removes the offending band without reaching up into the audible low end.