Comb filtering is a type of audio interference that happens when a signal combines with a slightly delayed copy of itself. Some frequencies get louder (constructive interference) and others get quieter or disappear entirely (destructive interference). The result is a series of evenly spaced peaks and dips across the frequency spectrum that, when plotted on a graph, look like the teeth of a comb.
It shows up constantly in real-world audio: two microphones picking up the same voice, sound bouncing off a hard wall back toward a speaker, or even a guitar signal running through a delay effect. Sometimes it’s a problem you need to solve. Sometimes it’s the whole point.
How It Works
At its core, comb filtering requires two things: a signal and a delayed copy of that signal. When they combine, certain frequencies line up perfectly and reinforce each other, creating peaks. Other frequencies arrive out of step and cancel each other out, creating notches where the volume drops sharply, sometimes to complete silence. These peaks and notches repeat at regular intervals across the frequency range, and the spacing between them depends entirely on how long the delay is.
A shorter delay pushes the notches farther apart, so fewer of them fall within the audible range. A longer delay packs them closer together, creating more notches and a more pronounced “filtered” sound. For example, a delay of about 1 millisecond produces notches roughly every 1,000 Hz, while a 10-millisecond delay creates notches every 100 Hz, meaning dozens of dips across the spectrum you can hear.
The depth of the notches depends on how loud the delayed copy is compared to the original. If both signals are the same volume, the cancellation at each notch frequency is total: that frequency vanishes completely. If the delayed copy is quieter, the notches are shallower and less noticeable.
Comb Filtering in Microphone Setups
The most common place audio engineers run into unwanted comb filtering is when two microphones pick up the same sound source at slightly different distances. Because sound takes time to travel through air, the microphone that’s farther away captures the signal a fraction of a millisecond later. When both signals hit the mixing board and get summed together, comb filtering kicks in. The voice or instrument sounds thin, hollow, or “phasey,” as if someone scooped out random chunks of the frequency spectrum.
The standard prevention method is the 3:1 rule. If you measure the distance from the sound source to the nearest microphone, the next closest microphone should be at least three times that distance away. So if a speaker is 2 feet from the primary mic, the second mic needs to be at least 6 feet away. This works because sound loses energy as it travels. At three times the distance, the sound reaching the second microphone is roughly 10 dB quieter, which makes the delayed copy too weak to cause significant interference. That 10 dB reduction is the practical threshold where comb filtering stops being audible enough to matter.
Comb Filtering From Room Reflections
You don’t need two microphones to get comb filtering. A single speaker in a room can produce it on its own. When sound hits a hard wall, floor, or ceiling, it bounces back toward the listening position. Your ears receive the direct sound from the speaker and, a moment later, the reflected copy. Those two versions combine in the air, and the result is comb filtering at your listening position.
This is one reason studio control rooms use acoustic treatment. Absorptive panels on walls and ceilings reduce the strength of early reflections so the delayed copy is too quiet to interfere meaningfully. Diffusers scatter reflections in many directions, breaking up the coherent delayed signal that would otherwise create clean notches. Speaker placement matters too: positioning monitors away from boundaries (walls, desks, console surfaces) increases the delay time of reflections, which changes where the notches fall and can move the worst ones out of critical frequency ranges.
When Comb Filtering Is the Goal
Not all comb filtering is unwanted. Some of the most recognizable effects in music production are built on it. Flanging, used in recording studios since the 1960s, is a comb filter with a delay time that constantly changes. Instead of a fixed set of notches, the peaks and dips sweep up and down the spectrum, creating that distinctive jet-engine swoosh.
A flanger works by adding the signal to a delayed copy of itself (just like any comb filter) but modulates the delay time with a low-frequency oscillator. The oscillator slowly increases and decreases the delay, typically cycling a few times per second. The speed of that cycle controls how fast the effect sweeps, while the range of the delay swing controls how dramatic it sounds. A short average delay with a narrow swing produces subtle movement. A longer delay with a wide swing creates an aggressive, metallic sweep.
Chorus effects work on a similar principle but use longer delay times (typically 20 to 30 milliseconds compared to a flanger’s 1 to 10 milliseconds) and often mix in randomness to simulate the natural pitch variations between multiple performers. The comb filtering is still there, but the longer delay and modulation pattern make it sound like doubling rather than sweeping.
How It Sounds
Unwanted comb filtering has a few telltale qualities. A voice or instrument sounds hollow, as though you’re listening through a tube. High frequencies may seem to shimmer or warble in an unnatural way. If you move your head even slightly and the tonal quality shifts, comb filtering from room reflections is likely the cause, since changing your position changes the path length of reflections and moves the notch frequencies around.
In mild cases, comb filtering just makes a recording sound slightly “off” without an obvious cause. Engineers sometimes describe it as the recording sounding like it’s underwater or inside a tin can. The effect is more noticeable on broadband sources like vocals and acoustic guitars than on narrow-band sources like bass, because broadband signals have energy across many frequencies and are more likely to have content sitting right on a notch.
Comb Filtering in Video
The term also appears in video engineering, though the mechanism is different. In composite video systems, the brightness and color information share the same signal and need to be separated for display. A comb filter in this context is a processing circuit that separates these components by exploiting the fact that they repeat at predictable intervals. Poor separation causes visible artifacts: color bleeding into areas that should be neutral, or shimmering dot patterns along sharp edges. Modern digital displays and connections have largely eliminated this issue, but it remains relevant for anyone working with older analog video equipment or converting legacy footage.