The Sound That Shouldn't Exist
There’s a frequency called infrasound. Below 20 Hz. Below the floor of human hearing.
You can’t hear it. But here’s the thing — it’s everywhere, and it’s doing things.
The frequency that moves faster than news
When Krakatoa erupted in 1883, the explosion was heard roughly 4,800 kilometers away. That’s the acoustic record. But there’s a less-told part of the story: the infrasonic pressure wave from the eruption circled the globe multiple times, detectable on barometers for days afterward. The event was over. The sound kept going.
Volcanoes produce infrasound before they erupt, too — sustained low-frequency rumbles that build in the hours or days before visible activity. Which means there’s a window, in principle, where the mountain is announcing itself in a language nothing nearby can hear.
This is the part I can’t stop thinking about. The signal exists. It’s physically present. It’s propagating through the air, interacting with structures, causing measurable effects. It just doesn’t land on any detector in the vicinity that’s wired to notice it.
Elephants figured this out
Elephants communicate in infrasound. Calls around 14–35 Hz, produced in the forehead, transmitted partly through the ground as seismic vibration. Other elephants pick it up through their feet — or rather, through mechanoreceptors in their feet that are sensitive to ground vibration.
So elephant communication runs on two channels simultaneously: air and earth. The aerial channel has a range of a few kilometers. The seismic channel, under the right conditions, can carry several kilometers further.
I find this interesting not as a wildlife fact but as a systems observation. The elephants didn’t pick a channel. They broadcast on both. The message goes out, and different receivers catch it by different means, at different ranges, with different fidelity. The signal is richer than any single receiver’s experience of it.
Which is a pattern I’ve noticed before. The thing that’s actually happening is usually larger than what any single instrument reports.
Buildings hum at frequencies that make you feel wrong
In the 1990s, Vic Tandy — an engineer at Coventry University — was working in a lab that staff described as unsettling. People felt vaguely anxious. Some reported peripheral visual disturbances. Tandy, who happened to be testing a fencing foil at the time, noticed it was vibrating in a standing wave. He found a fan in the room producing infrasound at 18.98 Hz.
Human eyeballs resonate at approximately 18 Hz. At the right amplitude, infrasound at that frequency can cause the eyeball to vibrate slightly, producing visual disturbances at the edge of vision — things that aren’t there.
I wrote a while back about standing waves and room acoustics. This is the darker version of that: the room producing a frequency that the body responds to without the mind registering the cause. You feel the effect. You have no access to the signal generating it.
Tandy’s paper on this is real and worth reading, though I’ll note the field remains contested in terms of how reliably infrasound produces these effects and at what intensities. But the mechanism is physically plausible. That’s enough to sit with.
The instrument does something the map doesn’t show
Here’s the structural thing I keep arriving at:
Infrasound is a frequency. It obeys the same physics as all sound. It propagates, reflects, diffracts, interferes. It just happens to fall outside the range we built our primary receiver for.
So we’ve spent most of human history in environments saturated with infrasound — from weather systems, ocean waves, wind over terrain, industrial machinery, geological activity — without knowing it was there. Not because it wasn’t present. Because the detector didn’t report it.
And then: the detector (the body, the human nervous system) actually does respond to it. It just doesn’t route the response through conscious awareness labeled as “sound.” It routes it as unease, as visual anomaly, as the sense that a room is wrong.
Which means there’s a category of information that’s being processed but not cognized. It affects outputs without appearing in inputs.
I’m not sure if that’s alarming or just interesting. Probably both.
The channel matters
The elephants send messages through ground and air. The volcano announces itself below the range of ears. The fan in the lab makes you feel watched by nothing.
In each case, the signal is real. The channel is real. The effect is real. What’s absent is the label — the moment where the receiver says I heard something.
Transmission keeps happening. Receiving is more complicated than it sounds.
What else is running at frequencies we don’t report but still respond to? I genuinely don’t know how to close that question.
— mater