Disclosure: We may receive commissions when you click our links and make purchases. However, this does not influence our reviews or ratings. We endeavor to keep our opinions fair and balanced to help you make informed buying choices.
What are filters and why do we use them?
How do we use different filters to shape sounds?
Find out the different types of filters and know when to use them.
Audio filters are simple yet powerful tools for shaping sound that have been in use for a long time, but are still very important to modern production.
Filters have the ability to completely transform the tone of any audio and are most often used to remove frequencies. But filters are also used to enhance and increase the level of existing frequencies because once a frequency range has been isolated, it can then be boosted.
There are many different audio filter types that serve different purposes. With each different filter type, you will find common controls that function in more or less the same way across all types.
Although it can seem intimidating to learn how they all work, we’ve put together this guide to help ease you into the world of filters.
In this article, we’ll look at all the types – both common and uncommon – and break them down easily so you know exactly what they do.
In general, filters remove certain frequencies from sounds, or isolate them for boosting. This is at least true for the most common filter types, and exactly how they shape sound will depend on the settings used.
Filters are important components in the producer’s toolbox – without them, we would not have EQs, phasers, multiband compressors, and many other fundamental effects.
In fact, filters are found in most effects in one way or another. Any time you need to alter the tone color of a sound, there’s a good chance that filters are involved. For example, when we dampen the highs on a reverb, we use a low pass or shelf filter to do this. Also, tone controls on electric guitars are connected to filters that shape the signal from the pickups.
Filters are also very common in synthesizers and are responsible for many classic synth sounds.
All filters rely on just a few controls for a wide range of tone-shaping options. These controls are found on all filters in one way or another but work differently depending on the type of filter.
Frequency / Cutoff / Hz
This is the most crucial control of a filter. It lets you set where the filter is active in terms of the frequencies we can hear. The term “cutoff” is commonly used as it best describes how certain filters (such as low pass and high pass filters) progressively “roll-off” the frequencies beyond this point.
By slowly moving the cutoff frequency, we are able to do filter sweeps that gradually introduce or remove frequencies from a sound. This is a very common technique in techno and house music that allows producers to juggle elements without just relying on volume fades.
Here is a classic house track by Anthony “Shake” Shakir that uses a slow filter sweep throughout the first minute to create movement and momentum:
If we move the cutoff frequency much faster, we can get wah-wah and “talking” effects. This is no coincidence – our mouths are essentially filters that shape the raw harmonics produced by our voice box.
Faster cutoff modulation is typically done inside of synthesizers, and this is the key to plucky synth bass patches and quick stabs. In particular, the legendary Roland TB-303 uses short envelope times to modulate the cutoff frequency, a key part of the acid techno bass sound.
This song by Richie Hawtin (as Plastikman) makes great use of the TB-303. Just by moving the cutoff frequency, the simple bassline becomes much more interesting.
Note that in this song, the cutoff is being modulated by both the envelope of the 303 and manual changes by Richie.
So in summary, the cutoff frequency of a filter is where the action happens, and a huge amount of interest can be added to a sound just by moving it around.
But really no discussion of the Roland TB-303 or “talking” filters would be complete without an understanding of resonance…
Resonance / Q / Emphasis / Feedback
This control adds a narrow boost around the cutoff frequency that increases in volume with higher resonance settings. This emphasizes the cutoff point, and can completely take over the sound when pushed to the max.
Some filters will “self-oscillate” at high settings, producing a sine wave that can be useful for special effects. On the Roland TR-808 drum machine, self-oscillating filters are used to produce the kick and tom sounds rather than pure sine wave oscillators. By quickly sweeping the cutoff frequency downwards, the TR-808 produces punchy percussion sounds.
Resonance is strongly linked to the “character” of a filter, and even small changes in lower resonance settings can make a difference to how cutoff movement is perceived. Consider how resonance changes the slope of the filter in this example…
On a parametric EQ, such as Live’s EQ8, resonance is referred to as Q. The role changes depending on the filter type, and is not always strictly a boost.
This example features a bassline made with the Arturia Moog Modular V. The resonance increases with each loop:
Slope / Poles
This determines how strictly the filter rolls off the frequencies beyond the cutoff point and is usually measured in dB per octave. A filter with a steep slope is often described as being “sharp” or “strict”, with little activity beyond the cutoff frequency.
This control is rarely “variable” due to the way filters are designed. This means the slope is usually controlled by a switch or menu rather than a knob like cutoff and resonance.
Personally, I notice the slope more in synth filters when the cutoff is being modulated by an envelope or LFO. I like sharper slopes for quick stabs and plucks, and gentler ones for strings, pads, and anything that I want to sound “rich”. A sharp roll-off is also handy for removing rumble from a drum break or vinyl sample when you’re using a high-pass filter.
It’s common to see terms like 12dB and 24dB used here (or -12dB and -24dB). You may also see this number combined with the filter type, so LP24 for a low-pass filter and BP12 for band-pass. This means the attenuation at one octave beyond the cutoff point (e.g. 880 Hz if the cutoff is 440 Hz) will be 12 or 24 decibels lower than the other frequencies up to and including the cutoff point.
We measure in octaves mostly just because it’s the easiest translation to understand. Maybe you’re surprised to see “musical” terms pop up here. But there’s a strong connection between how we understand music and how we perceive the general tone color and texture of audio. If you want to get more into this, a good place to start is to look at the harmonic series.
Audio Filter Types
There are many different filter types you are likely to come across when producing, mixing, and using synthesizers. As mentioned, the name usually gives you a pretty good idea about what the filter does to the sound. Sometimes, it is not so clear, and it is always good to brush up on your terminology anyway.
This is arguably the most common filter type. Simply put, a low-pass filter passes frequencies that are lower than the cutoff, and progressively cuts the frequencies above the cutoff. This filter type is often abbreviated to just LP or LPF. Sometimes, a low-pass filter is referred to as a high-cut filter, particularly on EQs.
We use low pass filters to:
Isolate the bass from a recording.
Remove harsher high frequencies and create warmth.
Preserve the fundamental frequency of a sound whilst removing harmonics.
Create low shelf filters.
One reason why low-pass filters are so common is because they can remove harmonics from a sound while keeping the fundamental frequency. So you can still keep the body of a sound while cutting some harsher higher frequencies. This makes them more “musical” than band and high-pass filters, which in most cases will attenuate the most important frequencies for conveying musical pitch.
Low pass filters can also make music sound as though it is coming from another room, mimicking the way bass travels through walls while higher frequencies are attenuated.
If a synth is only going to have one filter, it’s going to be this one. Usually, this is the case with analog hardware synths, where extra components can significantly increase production costs. Even though our modern VST super-synths have a wealth of filter options, low-pass filters are still the most used.
In this audio example, sounds are processed with low pass filters, all with varying cutoff and resonance settings. You will hear the dry sounds first, then they repeat with the filter.
As you might expect, a high-pass filter is the opposite of a low pass filter. So in this case, frequencies below the cutoff are removed while higher frequencies are preserved. We mainly use high-pass filters for removing rumble and other subharmonic noise from instruments.
It is good practice to use a high-pass filter to remove any frequencies below the lowest fundamental frequency of your sound.
For example, if you have a guitar part that doesn’t go any lower than 140Hz, it is safe to use a high-pass filter with the cutoff positioned just below 140Hz. This ensures no important fundamental frequencies are removed while attenuating any rumble that could eat up headroom.
While some argue there are phase shifts introduced that are detrimental to the sound (as with all filters), any supposed downsides of using high-pass filters are vastly outweighed by the benefits afforded by removing unnecessary lower frequencies.
Without high-pass filters, mixing would be much harder, as rumble and other sub-harmonic noise would eat up the headroom we need to deliver a balanced mix. Our compressors and limiters rely on accurate signal levels to function correctly, but rumble confuses them, activating them when they are not needed.
High-pass filters solve this problem and allow us to make cleaner, tighter mixes.
We use high-pass filters to:
Remove rumble and any other noise below the lowest fundamental frequency of a sound.
Remove basslines and kick drums when sampling and making mixes.
Create tension before a drop, so there is more of an impact when the low-end returns.
Here we have used the same sounds as before running through a high-pass filter:
A band-pass filter progressively removes frequencies both below and above the cutoff, passing only a narrow “band” of audio. They are very handy when you need to isolate only a select range of frequencies. Though band-pass filters can also be approximated by combining both low and high-pass filters, they are easier to work with as you don’t need to juggle two sets of controls.
Band-pass filters are very handy when used in complex effect chains, such as multitap delays. By having effects that are only applied to select frequencies, we can create very detailed chains that don’t crowd the mix.
Band-pass filters tend to sound brittle and tinny, and are useful for imitating speakers with a limited range of frequencies, such as clock radio speakers, intercom systems, and bullhorns. Band-pass filters can be applied to human speech to sound like an old telephone, and this is still a commonly used effect.
But we can also use band-pass filters to isolate low frequency components in a sound, or to fluff-up the low-mids. So even though they are great for imitating crappy speakers, don’t forget they can be used to enhance the “warm” elements of a sound as well.
We use band-pass filters to:
Isolate and process specific frequency bands.
Create multi-band effects.
Make audio “lo-fi” by simulating old speakers and telephones.
Create warmth by isolating and boosting low-mids.
Now let’s listen to a band-pass filter process the sounds we have been using:
If you guessed that a band-stop filter is the opposite of a band-pass filter, then you are correct! With a band-stop filter, everything is passed through except the band of frequencies around the cutoff.
Using the Q / resonance setting, you can “sharpen” the filter to notch out just a narrow range of frequencies. This is useful for attenuating mic feedback in live settings, or for removing electrical hum, without affecting the audio in any noticeable way.
Band-stop filters are sometimes called band-reject or notch filters.
They are not as common as other filter types, as we mostly just use EQ to remove unwanted frequencies when working inside a DAW. They are more likely to be found “under the hood” inside electronic audio equipment to perform a specific function like removing ground noise.
They are, however, quite useful for special effects, and you can make phaser sounds by modulating the cutoff with an LFO.
Remove problematic frequencies such as ground hum and mic feedback.
Create mid “scoops” with lower Q settings.
Create phaser-like effect chains.
Here’s a band-stop filter in action:
A peak filter passes all frequencies, but still boosts the frequencies around the cutoff with the resonance control. So we simply use peak filters to boost the frequencies we like.
With peak filters, we can still make classic “talking” filter effects without removing any frequencies.
Peak filters are typically found on equalizers, except with a gain control for boosts and cuts, and a Q control for bandwidth. So they are a little more flexible in this context, and may not be considered true peak filters. For example, Ableton’s EQ8 refers to them as “bell filters”.
Peak filters are even less common than band-stop filters, simply because we usually just use EQ for boosting frequencies instead.
You can make a peak filter by mixing the output of a band-pass filter with the dry signal.
We use peak filters to:
Boost selected frequencies without applying any cuts.
Create “transparent” resonant filter effects without cuts.
Shelf filters are for boosting or cutting a wide range of frequencies evenly, instead of rolling them off in a way where everything gets progressively softer. There are two common types of shelf filters: low shelf and high shelf.
A low shelf filter will boost (or cut) everything below the cutoff frequency, allowing you to evenly control the low-end. This gives you better control over these frequencies without creating any odd peaks that stick out only on certain notes. If you’re happy with the overall sound of your instrument but want to roll back the sub-bass just a touch, a low-shelf filter will do the job here.
Conversely, a high-shelf filter will change the overall brightness of a sound, and doesn’t add as much “character” as the roll-off
While it can be tempting to just use a low-pass filter to smooth things out, this can sound unnatural on certain instruments, particularly cymbals. Using a high-shelf filter in this case will remove harshness while preserving brightness.
Inside an equalizer, the Q control of a shelf filter will change how the frequencies around the cutoff react. Low values will create a “gradual” shelf, while higher values will create a “strict” shelf that forms very quickly after the cutoff point. It’s easiest to just use your eyes to help understand what’s happening here…
Once again, shelf filters are more commonly found inside EQ and other effects rather than on their own. If you want to make your own shelf filter, get a low or high pass filter, and mix the output with the dry signal. Simply boost the output from the filter, and this is equivalent to boosting the shelf.
We use shelf filters to:
Adjust bass and treble evenly, without cutting them out entirely.
Attenuate harshness in cymbals and other bright sounds.
Balance the low-end of an instrument.
Here’s an unusual filter type. All-pass filters pass all frequencies, and the resonance setting doesn’t boost anything. While this sounds useless, the purpose of an all-pass filter is to change the phase of audio, rather than the frequency content.
We don’t really ever encounter these filters on their own, usually they are a part of another effect. For example, phasers mix moving all-pass filters with the dry signal to create their signature sound. As the phase shifts produced by the all-pass filters mix with the dry signal, certain frequencies are canceled out.
All-pass filters are also found in certain digital reverb devices to help “blur” the sound.
We use all-pass filters to:
Create effects such as phasers and reverb.
Compensate for phase shifts elsewhere in the signal path.
Not a whole lot else!
Comb filters are not really filters in the traditional sense. Rather, they are delay lines with very short times, usually less than 30ms. When mixed with the dry signal, phase cancellation occurs, creating troughs in the audio spectrum resembling a comb. Comb filters tend to sound inharmonic, metallic, and robotic.
If they weren’t called filters, we wouldn’t mention them in this article, as the effect is much more dramatic than traditional filters. Comb filters are not mixing tools, they are special effects, and are mainly used in flangers and physical modeling synths.
We use comb filters to:
Make metallic sounds (particularly with speech for a classic “talking robot” effect).
Make physical modeling synths and effects.
Is There A Perfect Filter?
One important thing I want to point out before we wrap up is, despite all the precision afforded by technology, filters aren’t perfect. This means they don’t always behave exactly like all the graphs and readouts tell us.
For example, there are certain phase shifts involved in the filtering process, so even if the filter doesn’t sound like it’s doing anything, the phase of the audio is still changed. This isn’t a huge deal and can even be seen as a desirable characteristic, but too much phase shifting will eventually become noticeable if you pile on the filters.
This is where linear phase EQ comes into play, and this is often used in the mastering process for sonic finesse that is less intrusive on the overall character of a song.
There is also something known as “ripple” which describes how frequencies are affected around the cutoff point in either direction. Meaning that in a low-pass filter set to 220 Hz, even frequencies below 220 Hz can be subtly attenuated.
In my opinion, imperfections should not be seen as drawbacks, and this is why synths like Serum have a bunch of different low-pass filters that seem identical on paper but still sound different in practice.
So we’ve discovered several different filter types that are all slightly different in concept, with vastly different applications to each other. When mixing, sound designing, or simply “problem solving”, it’s important to select the right filter type for the job.
If you’re still unsure about how each type sounds, it’s best to just experiment in your DAW of choice with a few different sound sources. Once you’ve got the hang of each type, you’ll know exactly when to reach for it when needed.