2. Wave Interaction & Interference · Concept 4 of 10
Comb Filtering
A hollow, metallic, filtered tone you get when a sound and a slightly delayed copy of it mix together.
Direct sound plus a delayed copy add at peaks (+6 dB) and cancel at notches, drawing evenly spaced comb teeth set by f = 1/(2t).
What it is
A hollow, metallic, filtered tone caused by a sound mixing with a slightly delayed copy of itself.
Key facts
Speed of sound in air = 343 m/s at 20 degrees C (rises ~0.6 m/s per degree C)
Sound travels ~1 metre in roughly 2.9 ms (1 ms ≈ 34.3 cm of path)
First-notch formula: f = 1 / (2 x t), where f = notch frequency in Hz and t = delay in seconds
Notches sit at f = 1/(2t), 3/(2t), 5/(2t)... (odd multiples); peaks at 1/t, 2/t, 3/t... (whole multiples)
A 1 ms delay puts the first notch at 500 Hz, with a notch every 1000 Hz after; 0.5 ms = 1 kHz; 5 ms = 100 Hz
Two identical signals in phase ADD = +6 dB (double voltage); fully out of phase (180 deg) CANCEL toward minus infinity
Dips and peaks repeat evenly on a LINEAR Hz axis, which is what draws the comb shape; teeth get denser as delay grows
Extra path that causes the delay = delay x 343 m/s, so 1 ms = 34.3 cm of extra distance
3:1 rule: keep mics 3x further apart than each is from its source, putting spill ~9-10 dB down to weaken the comb
Worst case is two copies at EQUAL level (full nulls); unequal levels give shallower, more forgiving dips
How it works
Source makes a sound, and a second copy arrives a tiny bit late.
The delay shifts the copy's phase differently at every frequency.
Where the two line up (in phase) they add and you get a peak.
Where they oppose (180 degrees out) they cancel and you get a notch.
This add/cancel pattern repeats up the spectrum, drawing comb teeth.
Result: a fixed pattern of peaks and dips, heard as a hollow phasey tone.
Real examples
Two vocal mics close together both picking up one singer.
A floor/overhead mic plus the close mic on a snare drum.
A speaker plus its reflection bouncing off a nearby wall or table.
Lav or headset mic near a hard podium reflecting sound back.
Two speakers on the same signal overlapping at the audience seat.
How it helps in live sound
Obey the 3:1 rule: mics 3x further from each other than from the source.
Mic ONE source with ONE mic where you can; fewer copies = no comb.
If you must double-mic, use a phase/polarity flip and time-align (delay) the second mic.
Use the desk delay to nudge a mic by ms until the hollow tone fills out.
Keep mics off reflective surfaces; tape foam or angle away from podiums/tables.
Listen for that hollow flange-y vowel on voices, then mute one mic to confirm the comb.
Everyday analogy
Like two people clapping a hair apart in time so close it is not an echo, just a weird ringing colour where some notes vanish and others jump out.
Watch out
Myth: more mics on a source = louder and fuller. Truth: a second close mic adds a delayed copy and combs the sound thinner and phasey, so often quieter and worse.
Fun fact
A delay of just 1 millisecond already carves a notch at 500 Hz, right in the vocal body, which is why two close vocal mics can make a singer sound hollow without anyone touching an EQ.
Key takeaways
Comb filtering = sound + slightly delayed copy of itself.
Some pitches add (peaks), others cancel (notches), evenly spaced in Hz.