Mains Hum

The low drone underneath a bad recording is the sound of the electrical grid itself — fifty or sixty cycles per second, radiating from every wall, cable and power brick within reach of your microphone. Here's why the number depends on your continent, how it sneaks into a signal, and why it's one of the most completely fixable problems in audio.

What mains hum is

Everything plugged into a wall socket runs on alternating current. The voltage swings back and forth at the grid's frequency — 50 Hz across most of the world, 60 Hz in North America — and that swing sits squarely inside the audible range, near the bottom of a bass singer's register. Whenever some of that oscillation finds its way into an audio circuit, you hear it as hum: a steady, pitched drone that never pauses for breath.

That steadiness is the good news. Unlike hiss, which is random energy smeared across the whole spectrum and needs statistical treatment like spectral noise reduction, hum is periodic and locked to a frequency you can name in advance. You know exactly where it lives, which means it can be removed surgically rather than approximately.

Why Europe hums at 50 and America at 60

The split is an accident of 1890s industrial standardization. When alternating current won out over Edison's DC — the era of the "war of the currents" — each manufacturer picked its own frequency, and early AC systems ran at a bewildering assortment of them. The frequency had to be a compromise: high enough that arc and incandescent lamps didn't visibly flicker, low enough to suit the transformers, generators and motors of the day.

In the United States, Westinghouse — which had licensed Tesla's polyphase AC patents — converged on 60 Hz, and its commercial success made that the de facto American standard. In Germany, AEG settled on 50 Hz, and as its equipment and engineering practice spread across the continent, Europe followed. Neither figure is technically superior in any way that matters; they were simply the choices of the two firms whose gear everyone bought. Japan, memorably, bought from both — German generators in the east, American ones in the west — and still operates a split grid today, with Tokyo on 50 Hz and Osaka on 60. Once synchronous motors and clocks were counting grid cycles to keep time, the numbers froze for good. For your microphone, the takeaway is simple: your hum's fundamental frequency is set by the country you plug in to, and it is always one of exactly two values.

How hum gets into a signal

Mains hum doesn't arrive by one road. The common routes are worth recognizing, because each one sounds subtly different:

• Ground loops — two devices in the same signal chain grounded at different points. A small voltage difference between those grounds drives a mains-frequency current along the cable shield, and the audio circuit reads it as signal. This is the classic strong, stubborn hum that appears the moment you connect one particular cable.
• Induction — transformers, power cables and wall-wart supplies radiate a magnetic field at the mains frequency, which induces a matching current in any nearby audio wiring. It gets louder as cables get closer to the source and vanishes when you move them.
• Power-supply ripple — a cheap or failing supply lets residue of the AC waveform leak into the DC that powers the audio circuitry. Because most supplies rectify both halves of the cycle, this ripple typically lands at twice the mains frequency — 100 or 120 Hz.
• Dimmers and switching loads — a phase-cut dimmer works by chopping the AC waveform partway through every cycle. Those sharp edges are rich in high harmonics, which radiate into nearby wiring as an aggressive buzz that changes character as you move the dimmer.

The harmonic ladder: hum versus buzz

A mathematically pure 50 or 60 Hz tone would be a soft, dark hum — barely reproducible on laptop speakers at all. Real-world hum is never pure. Rectifiers, dimmers and saturating transformers distort the waveform, and any distortion of a periodic signal generates harmonics: energy at integer multiples of the fundamental. A 50 Hz hum brings friends at 100, 150, 200, 250 and 300 Hz; a 60 Hz hum climbs 120, 180, 240, 300, 360.

This ladder is why engineers distinguish hum from buzz. The fundamental and first harmonic are the hum — low, round, easy to ignore. The higher rungs land in the lower midrange, where the ear is far more sensitive, and read as buzz — edgy, insistent, impossible to ignore. In a dimmer-polluted room the upper harmonics often carry more energy than the fundamental itself, which is why filtering only the fundamental can leave the annoyance almost intact. Point a spectrum analyzer at a humming signal and you'll see the whole picket fence at once: evenly spaced spikes marching up from 50 or 60 Hz. Any fix has to climb the ladder, not just cut the bottom rung.

Notch filters, Q, and the high-pass below the voice

The surgical tool for a known, fixed frequency is the notch filter: a deep, narrow cut at one frequency that leaves everything on either side untouched. Its sharpness is described by Q — center frequency divided by bandwidth — so a high Q means a narrower notch. Because hum sits at exactly predictable frequencies, the notches can be made very narrow indeed, and narrowness is what makes them safe. A voice is a broadband, constantly moving signal; the energy it places at any single frequency is momentary and small, so a notch a fraction of an octave wide removes a continuous tone while costing speech almost nothing. Contrast that with a broad EQ cut in the same region, which would noticeably thin the voice — the difference between a scalpel and a shovel, and the same narrow-cut logic that a parametric EQ applies more broadly.

The second tool is the high-pass filter. Spoken fundamentals rarely reach below about 85 Hz even for deep male voices, so everything under roughly 70 Hz is not voice: it's rumble, desk thumps, traffic and the mains fundamental region. Rolling all of it off costs nothing audible and removes the hum's foundation along with plenty of other low-frequency junk. The combination is the standard recipe: a high-pass to clear the floor, then notches to pick off the harmonics that survive above it. And because hum never stops, a noise gate can't solve it — the gate opens for speech and the hum walks in with it. Filters, not gates, are the answer.

In DeskBroadcast

DeskBroadcast's hum filter implements exactly that recipe. A 70 Hz high-pass clears everything below the voice, and six narrow notch filters sit on the mains fundamental and its harmonics — ×1 through ×6, so 50 through 300 Hz or 60 through 360 Hz depending on your grid. You make one choice: 50 Hz or 60 Hz, matching the country you're plugged in to. Each notch is 0.25 octaves wide — narrow enough that your voice passes through untouched while the steady tones at those exact frequencies are pulled out. It covers both the low hum and the midrange buzz of the harmonic ladder in one switch, and it runs ahead of the rest of the mic chain, so downstream processing like noise reduction works on a signal that's already free of the grid.