When you mix a cover song, the “high-pass filter” is often the first thing you touch. The name sounds difficult, but what it does is basically “reduce sounds lower than a frequency you specify”. This article explains in order what the filter does, why it is used on vocals, how to decide the frequency to cut, and what happens if you cut too much. It also covers how it differs from the low-cut switch on microphones and audio interfaces, as far as the makers’ materials go.
Disclosure: This article contains affiliate links. Prices are as of the date we checked them and can change.
Summary: a high-pass is a filter that “does not let through low end the voice does not need”
- A high-pass filter (HPF) and a low cut are the same thing. It reduces sounds lower than a frequency you specify.
- On vocals, it is used to reduce low rumbles such as footsteps, air conditioning and desk vibration, and to keep the voice from clashing with the low end of the backing track.
- There is no fixed number for the frequency to cut. FabFilter introduces a by-ear method: “find the point where the effect starts to be audible, then back off a little”.
- Cutting too much makes the voice thin. Check it by listening with the backing track, not in solo.
- You can fix an EQ high-pass as many times as you like later, but a switch on a microphone or interface is already in the sound at the moment you record.
What a high-pass filter is
What the name means
FabFilter’s explanation describes a high-pass filter as “a filter that lets high sounds pass as they are and removes sounds lower than the frequency you set (the cutoff frequency)”. Because it cuts low sounds, it is also called a “low-cut filter”, and sometimes abbreviated as HPF. The opposite, which cuts high sounds, is a low-pass filter (high cut).
The unit of frequency is Hz (hertz), and the smaller the number, the lower the sound. If the high-pass is set to 100 Hz, everything below 100 Hz is reduced.
How sharply it cuts is decided by the slope
A high-pass does not make sound vanish sharply at the set frequency. The farther you go from that frequency, the more strongly it is reduced. How steep this reduction is is called the “slope”, and it is expressed in dB per octave (dB/oct). The larger the number, the more sharply it cuts. The FabFilter Pro-Q 4 help explains that the slope can be changed finely in a range of 0 to 96 dB/oct.
Sources: FabFilter Learn, FabFilter Pro-Q 4 help
Why use a high-pass on vocals
FabFilter describes the high-pass filter as “a tool that helps remove non-musical sounds below an instrument’s range, and other sounds that bleed in”, and also writes that it is often the first tool you reach for when you start a mix. Applied to cover-song vocals, the following situations come to mind.
- Reduce low noise that got into the recording: footsteps, air conditioning, vibration through the desk or floor, contact sounds from the mic stand and so on. These tend to sit far below the range of the voice.
- Make room for the low end of the backing track: FabFilter explains that low frequency bands take up space when sounds overlap and greatly affect the balance of the mix. Tidying up the low components of the vocal may make it easier to hear the bass and kick.
- Tidy up the order of EQs: an EQ that cuts unneeded low end is generally placed before the compressor. The reason is explained in The order of plugins in a cover mix.
FabFilter describes the low range around 20 to 50 Hz as “a band you feel with your body rather than hear”, and also writes that in this band you cut more often than you boost. 50 to 125 Hz is the band where the fundamentals of bass and kick appear, and it too is a place where sounds easily overlap. The components that matter for the voice itself lie a little higher.
Source: FabFilter Learn
How to think about deciding the frequency
No fixed number such as “cut vocals at this many Hz” is given even in the makers’ materials, because the best position changes with the pitch of the voice, the microphone and the room. Instead, FabFilter introduces the following ways of thinking.
Method 1: decide by ear
- Create a high-pass (Low Cut) in your EQ and set the slope to 18 dB/oct (the value FabFilter gives as an example).
- Raise the frequency little by little from the low end, and find the point where you hear the low end start to be cut.
- Once you hear it, back off a little from there.
The goal is to reduce only the unneeded low end without changing the character of the sound. It helps to think of it as going up to “the point where the effect is audible” and stopping just before “the point where the voice changes”.
Method 2: think from the lowest note of the song
FabFilter explains that if you know the frequency of the lowest note an instrument produces, you can place the cutoff just below it. Some EQs, such as Pro-Q 3, show you how note names correspond to frequencies. For singing, “the lowest note you sing in this song” is the reference. However, a voice also has components above the fundamental (overtones), so placing it exactly at the lowest note can thin out the voice. It is safer to leave a little room below rather than putting it exactly on the note.
In either case
- Do not decide too much in solo: FabFilter explains that probing the sound in solo is instructive, but what finally matters is how it sounds in the mix.
- Compare on and off at the same volume.
- Treat the number as a starting point and decide by ear in the end.
Source: FabFilter Learn (Introduction to EQ)
Cautions about cutting too much
A high-pass is a filter that only “reduces”, so it looks safe, but if you raise it too much it also cuts into the body of the lower part of the voice. If you see signs like the following, lower the frequency or make the slope gentler and see how it goes.
- It sounds fine in solo, but when mixed with the backing track the voice sounds thin and distant.
- Only the low parts of the voice (low phrases in the song) suddenly lose their power.
- When you compare on and off, “on” sounds “smaller”.
The right position differs between singers with low voices and high voices. Even within one song, the impression can change between low and high parts. Searching with your own voice is less likely to go wrong than “using that person’s settings as they are”.
For the slope, try starting around 18 dB/oct as in FabFilter’s example, and adjust while comparing.
EQ high-pass versus the low-cut switch on gear
Besides EQ, some microphones and audio interfaces have a low cut as a switch on the unit itself. Here are the differences, as far as the makers’ official pages go.
| EQ high-pass (in the DAW) | Switch on a microphone or interface | |
|---|---|---|
| Where it is applied | After recording, inside the DAW | Before recording (at the input stage) |
| Frequency | Can be set freely (for example, FabFilter Pro-Q 4 lets you adjust frequency and slope) | Often fixed by the model |
| Redoing it | Can be changed or turned off at any time | It is already in the recording, so you cannot undo it later |
| Main purpose | Tidying the mix | Reducing unneeded low end at the time of recording |
Examples confirmed on official pages
- audio-technica AT2035 (microphone): the official page says it has an “80 Hz low-cut filter switch” that reduces noise such as air conditioning.
- SSL 2 MKII (audio interface): the official page says the microphone preamp has a “dedicated high-pass filter”.
Even with gear that has the switch, whether you use it is up to you. If you are unsure, the most recoverable approach is to record with the switch off and adjust later with the EQ high-pass. If your room clearly lets in low noise such as air conditioning, using the switch is also an option. Check whether each model has the switch and at what frequency on the maker’s specification page before you buy.
Sources: audio-technica official, Solid State Logic official
The steps in short
- Put an EQ on the vocal track and create one high-pass (Low Cut).
- Try a slope of around 18 dB/oct.
- Raise the frequency from the low end and find the point where the low end starts to be cut.
- Back off a little from there.
- With the backing track, compare on and off at the same volume.
- If the voice has become thin, lower the frequency.
The high-pass is groundwork to finish first, before the EQ work of tidying up muddiness and harshness. The rest is covered in How to EQ vocals.
Frequently asked questions
Q. Are a high-pass and a low cut different things?
They are different names for a filter with the same function. FabFilter also explains that it is called a “high-pass” because it lets high sounds through and a “low cut” because it cuts low sounds.
Q. Do I have to use one?
There is no rule. FabFilter also takes the position that EQ has no rules, only choices. If your recording has no low noise and the low end does not get muddy when mixed with the backing track, it is fine not to use one.
Q. If the high-pass is not enough, what else do I use?
For muddiness, other filters such as a bell are suited, and for sibilance, a de-esser is. How to use them is explained in How to EQ vocals and How to use a de-esser on vocals.
Q. Do free EQs have a high-pass too?
You can find free EQs that handle the basic adjustments. You can check their features in Free EQ plugins and EQ plugin comparison.
Q. Can I remove low-end noise after recording?
A high-pass can reduce components lower than the range of the voice. However, noise at the same pitch as the voice cannot be removed with a high-pass. It is more reliable to make the noise small when you record in the first place. Setting up your recording environment is explained in How to record vocals at home.
Comments