A mysterious external source produces the Taos Hum

Enough people heard it that Congress got involved

Give the strongest version of this argument its due, because the people making it are not cranks. By 1993, enough Taos-area residents were describing the same basic experience (hearing a low, mechanical-sounding drone that others around them could not) that they petitioned for outside help, and the request was taken seriously at a level few unexplained local phenomena ever reach. A formal investigation was commissioned, combining University of New Mexico researchers with scientists connected to Sandia National Laboratories, Los Alamos National Laboratory, and Phillips Air Force Laboratory, a team with genuine expertise in acoustics, seismology, and signals detection. That alone tells you the reports were not dismissed as mass delusion from the outset.

The team did not just take residents' word for it. Investigators surveyed 1,440 residents and worked directly with self-identified “hearers,” extrapolating that roughly 2% of the local population, about 161 people, could reliably perceive the hum. That is a small minority, but it is not one or two isolated eccentrics; it is a measurable, replicable slice of a community, consistent enough that researchers could bring hearers into testing and have them match the sound to specific frequencies, generally in the 30 to 80 Hz range, with a slow pulsing or modulation.

There is also a reasonable steelman for why an external source might elude detection even with good equipment: low-frequency sound behaves unlike higher-pitched noise. It travels further, diffracts around obstacles, resists localization by the human ear, and can be radiated by very ordinary infrastructure (transformers, compressors, pipelines, distant industrial equipment) in ways that are geometrically hard to pin down from a fixed sensor network. The comparison case of Windsor, Ontario is instructive here, and not dismissible. Residents there began reporting a very similar low, throbbing hum in late 2011, and by 2013 the Canadian federal government had funded a formal investigation of its own: a University of Windsor team led by engineering professor Colin Novak deployed noise-monitoring and beamforming equipment across the Windsor-Essex region, and its report, released in May 2014, concluded that the hum's acoustic signature pointed to blast-furnace operations at a U.S. Steel plant on Zug Island, an industrial island in River Rouge, Michigan, across the Detroit River from Windsor. That conclusion held up in the most convincing way an unresolved Hum case ever has: in 2020, U.S. Steel idled its Zug Island blast furnaces, and the Windsor Hum, after nearly a decade of complaints, went quiet along with them. If one well-studied Hum case had a real, external, industrial source that a formal investigation identified and a real-world shutdown later corroborated, it is not unreasonable to have expected Taos might too, and to keep looking rather than settle for “equipment found nothing.”

What the instruments actually recorded

The trouble for an external-source explanation is not that no one looked. It is that people looked hard, with the right tools, during a period when hearers were actively reporting the sound, and came back empty. The 1993 team deployed low-frequency microphone arrays alongside seismic and electromagnetic monitoring equipment around Taos, instrumentation sensitive enough to register signals far weaker than a hum loud enough for a person to notice and be bothered by. Their findings, released as the Taos Hum Investigation: Informal Report on August 23, 1993, and later formalized in a peer-reviewed account by team members Joe Mullins and James Kelly, were direct: there was no consistent acoustic signal, seismic event, or electromagnetic anomaly that matched what hearers described. Whatever the hearers were perceiving, it was not arriving as an ordinary, air-conducted or ground-conducted sound wave that a well-instrumented sensor network could register.

The pattern among hearers themselves also cuts against a single shared external wave. When researchers asked individual hearers to match the sound they perceived to a specific tone, the matches were inconsistent: different people, including two standing side by side, frequently matched the “same” hum to different frequencies within that 30 to 80 Hz band. A single real sound wave, propagating through the air, produces the same acoustic signal for every ear in its path; it does not produce a different frequency for each listener. That variability points toward each hearer perceiving something generated or amplified within their own auditory system, rather than one external tone reaching every hearer's ear identically.

That is exactly the direction later peer-reviewed research has pushed. A 2016 study in the International Tinnitus Journal analyzing questionnaire data from Hum-hearers proposed that, for most, the sound represents a rare form of tinnitus rooted in the inner ear rather than an external signal. A more rigorous 2026 study published in PLOS ONE tested this directly: researchers measured low-frequency hearing thresholds and searched for spontaneous otoacoustic emissions (faint sounds the healthy inner ear produces on its own) in people who reliably perceive a Hum-like sound. They found no unusual low-frequency hearing sensitivity and no low-frequency otoacoustic emissions that could explain the reports, and concluded that for most participants with normal hearing thresholds, a rare, low-frequency variant of subjective tinnitus (sound generated by the auditory system itself, without any outside source) was the most likely explanation. Combined with plausible contributions from ordinary, low-level environmental low-frequency noise (traffic, HVAC systems, distant industrial equipment) that sits at the edge of audibility for people with unusually acute low-frequency perception, and the well-documented psychological tendency to lock onto and amplify a sound once primed to expect it, the mundane explanations account for the pattern of findings considerably better than an undetected external source does. Seventy-plus years of combined acoustic and physiological research into Hum-like phenomena has never produced a verified recording of the Taos hum itself, nor identified any concealed emitter, industrial or otherwise, near Taos.

The Windsor comparison, too, is less tidy than its eventual outcome makes it look. The same 2013–2014 investigation that led the University of Windsor team to Zug Island ran alongside a parallel study by University of Western Ontario physicists, who deployed portable infrasound arrays across the region between February and April 2013. Their bearings pointed well south of Zug Island, and their published conclusion, formalized in a 2018 peer-reviewed paper, was direct: the data did not support the hypothesis that the hum's source was Zug Island, pointing instead to quarry activity or industrial ventilation systems as equally plausible alternatives. Neither Canadian team was ever granted access to Zug Island itself, a private U.S. industrial site, so even the widely reported 2020 outcome (furnaces idled, decade-old complaints stopped) is a real-world correlation, not a confirmed, on-site-measured mechanism. It took the better part of a decade and a fortuitous industrial shutdown for Windsor's hum to look solved, and even now the resolution rests on timing rather than a verified signal traced to its source. Taos, more than thirty years on, has had no comparable shutdown to test its own hum against.

Why the hearers, and the belief, persist

None of this means the hearers are imagining a problem or exaggerating for attention. Whatever its origin, a persistent, uncontrollable low drone that intrudes on sleep and concentration is a genuinely difficult thing to live with, and the people who report it in Taos have, by every account, described it consistently, soberly, and often at real personal cost: strained sleep, frustration, a sense of being disbelieved by neighbors who hear nothing at all. That last part matters enormously to why an external-source belief persists: when your own perception is real and disruptive but invisible to nearly everyone around you, and even to sensitive scientific instruments, the two available conclusions are “something real is out there that we haven't found yet” or “the problem originates inside my own body.” The first is, understandably, the easier one to sit with.

An unresolved federal investigation compounds that pull. When a serious team of scientists studies something and reports back we could not identify a cause rather than there is nothing here, that ambiguity reads, fairly enough, as an open question rather than a closed one, and open questions invite theories, including that officials connected to nearby national laboratories and an Air Force facility might know more than they have said. Taos sits near real, historically secretive federal science infrastructure, which gives that suspicion a plausible-sounding neighborhood to live in even though no document, leak, or declassification has ever tied the hum to any program there.

The broader, worldwide pattern of Hum reports reinforces belief in an external cause, too. It is genuinely true that not every Hum case has turned out to be internally generated: Windsor, Ontario's hum, after years of dispute between competing research teams, ultimately tracked closely with the operating schedule of a specific steel plant's blast furnaces: when they were idled in 2020, the hum went quiet too, a real, external, mechanical cause the record now supports even without a direct, on-site measurement confirming it. Knowing that at least one Hum was real, external, and industrial makes it entirely reasonable for a Taos hearer, or an interested observer, to expect the same must be true here, just not yet found. The mundane truth for most Taos hearers may be less satisfying: a rare perceptual condition inside their own ear, with no external culprit to uncover, no matter how long the search continues.