The Repository Abyss Report: Into the Dark

Nobody Has Ever Heard the Bloop, Including You

August 26, 2026 · Runtime 16:40 · Watch on YouTube

The most famous mystery sound in the ocean is a recording running sixteen times too fast. Undo that, and the event lasted two minutes and forty-one seconds and sat between thirteen and twenty-six hertz - below the bottom of human hearing. Every version anyone has ever played is a translation.

Chapters

Transcript

Something made a noise 0:00

In the summer of nineteen ninety-seven, something in the South Pacific made a noise. Hydrophones more than five thousand kilometres apart heard the same event. It was catalogued, it was given a name, and the name got away from the data. For twenty years people said it was an animal. Bigger than a blue whale, they said. The loudest sound ever recorded.

You have never heard it 0:24

Here is the part that almost nobody says out loud. You have never heard the Bloop. Neither has anyone else. The recording everybody has played is running sixteen times too fast, and at its real speed the sound sits below the bottom of human hearing. This episode is about what the spectrogram actually shows, and about how long a mystery can outlive its own answer.

The array, and what it listens for 0:49

Start with the instrument, because the instrument is the reason there is a recording at all. The National Oceanic and Atmospheric Administration runs an acoustics programme out of its Pacific Marine Environmental Laboratory. Part of that programme is an array of autonomous hydrophones moored in the equatorial Pacific. They are not listening for monsters. They are listening for earthquakes on the sea floor, for volcanoes, and for whales.

Why one crack crosses an ocean 1:18

A hydrophone in the deep ocean can hear an extraordinary distance, and that is not because the instrument is good. It is because of the water. At around a kilometre down there is a layer where the speed of sound reaches a minimum, and low-frequency sound entering that layer is bent back into it from above and below. It stops spreading in three dimensions and starts travelling in two, like light in a fibre. That is how one crack in one block of ice reaches a microphone on the other side of an ocean.

How to read every picture here 1:49

One more piece of equipment, and it is not hardware. Everything anyone knows about this event comes from a spectrogram, so it is worth thirty seconds on what one is. Take the recording, chop it into short slices, and for each slice ask how much energy sits at each frequency. Lay the answers side by side. Time runs one way, frequency runs the other, and brightness is how much energy was there. A trained analyst does not listen to these files. They look at them. That is the whole discipline, and it is why the sped-up audio was never the evidence.

Summer 1997, on multiple sensors 2:27

So: summer of nineteen ninety-seven. The array records a broad-spectrum event. It is loud enough to appear on multiple sensors separated by more than five thousand kilometres, which is the detail that did most of the work in the story that followed. Nothing in the catalogue of known sounds matched it exactly, and the file went out into the world with an informal nickname attached.

The recording everyone has played 2:52

The nickname is Bloop, and the nickname is a description of a file, not of an event. Play the published recording and you hear a rising, watery swoop, over in about ten seconds. It does sound like something. That is exactly the problem. That file is a sped-up copy, made so that a human being can hear anything at all, and the sped-up copy is what got named.

Sixteen times too fast 3:28

Sixteen times. That is the number written on the page NOAA still publishes today. Recorded signal sped up sixteen times. Speeding a recording up moves both axes at once. Every second becomes a sixteenth of a second, and every frequency is multiplied by sixteen. Undo it and you get a different object entirely.

161 seconds, 13 to 26 hertz 3:52

I measured the published file to get the real numbers, and here they are. The clip is ten point three seconds long, so the event lasted about a hundred and sixty-one seconds. Two minutes and forty-one seconds. Not a swoop. A slow, sustained groan lasting almost three minutes. And the energy that reads as a two-hundred-and-fifty hertz whoop in the sped-up file sits, in the original, between about thirteen and twenty-six hertz.

The measurement, so it can be repeated 4:21

For the record, here is exactly how, because a claim like that should be reproducible. Download the file NOAA publishes. Slice it, run a Fourier transform on each slice, and find where the energy sits. Then undo the speed-up: multiply every time by sixteen and divide every frequency by sixteen. The first attempt got it wrong, and the way it got it wrong is worth knowing. Without restricting the search to a sensible band, the strongest thing in the file is the slow drift near zero hertz, so the trace comes out flat along the floor and looks like nothing happened. The signal was always there. The measurement was pointed at the wrong part of the spectrum.

Twenty hertz, and what is under it 5:06

Twenty hertz is roughly where human hearing stops at the bottom. Below that you can sometimes feel a sound in your chest, in a room with a big enough speaker, but you do not hear a pitch. Most of the Bloop is under that line. So the loudest sound ever recorded, the one that supposedly announced a creature larger than any animal alive, is a sound that no human ear has ever resolved. Every version you have heard is a translation.

Why the animal story was not stupid 5:34

And now the honest part, the part that makes the monster story understandable instead of stupid. Thirteen to twenty-six hertz is not a strange band to find a living thing in. It is the band the largest baleen whales call in. Blue whales and fin whales work down there, for the same reason ice does and for the same reason submarines do: low frequency travels. So the Bloop turned up in the right neighbourhood. What rules the animal out is not the pitch. It is the duration and the energy, sustained for almost three minutes, at a level no measured animal call approaches.

The chain that produced a monster 6:12

Now the story, presented as a story, because it is worth understanding why it worked. The claim was that the Bloop was biological. It had a rough, organic profile, it did not match any known machine, and it was far too powerful for any animal on the books. From there the arithmetic is easy and completely unearned: if it is an animal, and it is louder than a blue whale, then it is bigger than a blue whale.

The coincidence with the map 6:40

The story got a second engine from geography, and the geography came from NOAA itself. The old version of that same page gave a general location for the source: fifty degrees south, one hundred degrees west. That is open South Pacific, about seventeen hundred kilometres from the point in the ocean furthest from any land - and a writer of horror fiction had, decades earlier, put a sunken city in roughly that stretch of the map. The coincidence is genuinely striking, and it is a coincidence. It is also the reason this particular data file has fan art.

Far is not the same as loud 7:17

The phrase that did the most damage is the loudest sound ever recorded, and it fails for a boring technical reason. Loudness underwater is not measured on the same scale as loudness in air. The reference pressures are different, so the numbers are not comparable, and a figure that sounds apocalyptic converts to something ordinary. More importantly, detection at five thousand kilometres is not a measure of how loud a source is. It is a measure of how well that sound channel carries. Far is not the same as loud.

Two questions, in the right order 7:51

There is a cleaner way to test the animal hypothesis, and it does not require any speculation about undiscovered biology. Ask what a sound-producing organ has to do to put that much energy into the water for almost three minutes, and ask whether anything with a body has ever been observed doing it. Then ask the opposite question. What non-living process, in that ocean, routinely puts enormous energy into water at very low frequency, in long broad-band events?

Ice 8:22

Ice. The page today reads like this. The broad-spectrum sounds recorded in the summer of nineteen ninety-seven are consistent with icequakes, generated by large icebergs as they crack and fracture. Not a hypothesis offered in place of an answer. A match, against recordings of the same thing happening where someone could watch it.

The page did not always say ice 8:45

I said the page today, and I want to be exact about that, because the archived versions of it are their own small piece of evidence. The page did not always say ice. In the copy saved in two thousand and seven it is titled simply Bloop, it gives that South Pacific position, and it ends with one sentence: the origin of the sound is unknown. That is still the text in the copy saved in June two thousand and twelve. By January two thousand and thirteen the page has a new title - Icequakes, in brackets, Bloop - and the explanation you just heard. So NOAA carried an open question for fifteen years, and then closed it, and moved its own estimate of where the sound came from by thousands of kilometres, toward Antarctica.

The Scotia Sea recordings 9:32

This is the step that turns an explanation into evidence, and it is the step the legend never mentions. NOAA hydrophones deployed in the Scotia Sea, off the Antarctic peninsula, recorded numerous icequakes. Their spectrograms are very similar to the Bloop. That is the whole argument, and it is the right kind of argument: the unknown signal is matched against a known source recorded by the same kind of instrument.

An iceberg calving 10:00

Listen to what an iceberg actually does. This is a calving event, a large section breaking off a drifting berg, recorded by NOAA and published sped up three times. The signal is short and broad-band, running from about one hertz to four hundred and forty, and it is generated by cracks opening and running through ice.

Harmonic tremor 10:30

And this is what a berg does when it is in contact with something, either the sea floor or another iceberg. It is called harmonic tremor, and it is a stack of evenly spaced tones with the same spacing repeating upward through the spectrogram. Ice, in other words, has more than one voice, and each one leaves a shape you can recognise.

Slow Down, and what it is 11:00

There is a second famous mystery sound on the same NOAA page, and it has the same answer. It is called Slow Down, and the name is literal: the sound descends in frequency, slowly, for about seven minutes. In the published clip the band climbs to around fifteen hertz and then falls away to about four. That is an iceberg running aground. The berg is drifting, it touches the sea floor, and it grinds to a stop. What you are hearing is a mass of ice losing its speed.

Tracking iceberg A53a by sound 11:45

And here is the part that convinced me, because it is not an argument at all. It is an operation. In early two thousand and eight, NOAA used icequake recordings to track an iceberg. A fifty-three A, drifting near South Georgia island, disintegrating. They followed a block of ice breaking apart, acoustically, from thousands of kilometres away, using exactly the class of signal that the Bloop belongs to. You do not do that with a sound you do not understand.

Bransfield Strait to the Ross Sea 12:18

So where was the Bloop from? The arrival angles put the source somewhere between the Bransfield Strait and the Ross Sea, and possibly at Cape Adare, which NOAA describes as a well known source of cryogenic signals. That is a region, not a point, and the honesty of that answer is the reason to trust it. An array of moored hydrophones gives you a bearing. Bearings from far enough apart give you an area.

How a fixed microphone gives a bearing 12:44

How do you get a direction out of a microphone that cannot turn? You do not. You get it out of several of them. The same event arrives at different moorings at different moments, and the difference in arrival times, with the speed of sound in that layer known, gives a bearing from each. Cross the bearings and you have an area. It is the same geometry used to locate an earthquake, and it is why the array is an array and not a microphone.

What is still open 13:12

Which leaves one thing genuinely unresolved, and it is worth being precise about what it is. We do not know which iceberg. Nobody was watching that stretch of Antarctic coast on that day at that hour with a camera. To close it you would need an ice-shelf record with the date and the position, matched to the arrival times on the array. That is a filing problem, not a monster.

A fifteen-year head start 13:38

Now the uncomfortable question, and the archived pages make it sharper. For fifteen years the internet was right that NOAA did not know, because NOAA said so in writing. The monster had a fifteen-year head start, and the correction has now had more than a decade and has not caught up. Part of that is that a solved mystery stops being shareable. Part of it is more specific: the answer arrived as a paragraph and a spectrogram, and the story arrived as a sound file you could play at a friend.

The instrument, filmed at depth 14:09

It is worth seeing the instrument, because almost nobody ever does. This is a hydrophone going down, filmed on a NOAA dive off Puerto Rico in twenty eighteen. There is no laboratory. There is a pressure housing, a cable, and a ship. Everything in this episode, every number, every spectrogram, every claim about a berg five thousand kilometres away, comes out of something that looks like this, sitting in the dark, recording.

Julia, Train, Upsweep, Slow Down 14:39

The same programme has a catalogue of other named signals, and the names are all like this. Julia. Train. Upsweep. Slow Down. They are onomatopoeia, assigned by analysts to files, and every one of them has been claimed at some point as evidence of something enormous and hidden. The array was built to listen for earthquakes and volcanoes and whales, and it does that continuously, and the ice is a by-product it caught because ice is loud and the ocean carries.

What this actually forces 15:12

So what does this case actually force us to conclude? Not that the ocean is boring. The opposite. A block of frozen fresh water, cracking under its own weight in the dark, put enough energy into the southern ocean to be recorded on instruments a continent away. That is a real event, at a real scale, and we have the recording. The story replaced it with a fictional animal, and the trade was bad.

A very large piece of ice broke 15:40

The Bloop is not an open case. It is a closed case with a famous name. The recording is still there, still public, still sped up sixteen times because otherwise you cannot hear it. Somewhere between the Bransfield Strait and the Ross Sea, in the summer of nineteen ninety-seven, a very large piece of ice broke. The hydrophones were already listening. That is the whole story, and it is enough.

Next: a call at fifty-two hertz 16:07

Next time, a sound that is genuinely unexplained, and a case where being careful matters more, not less. In nineteen eighty-nine, American navy hydrophones picked up a whale call at fifty-two hertz, which is far above the frequency of any known species. It has been tracked for more than a decade, always alone, never answered. The story says it is the loneliest whale in the world. The data says something more interesting, and the difference is the episode.

Description and sources

In the summer of 1997 an array of autonomous hydrophones in the equatorial Pacific, run by NOAA's Pacific Marine Environmental Laboratory, recorded a broad-spectrum event picked up on sensors more than 5,000 kilometres apart. It got a nickname - the Bloop - and the nickname got away from the data: an animal larger than a blue whale, the loudest sound ever recorded. NOAA identified it as an icequake, a large iceberg cracking and fracturing, matched against recordings of the same thing from hydrophones in the Scotia Sea. The archived copies of that NOAA page are their own piece of evidence: it said "the origin of the sound is unknown" as late as June 2012, and was retitled "Icequakes (Bloop)" by January 2013. This episode measures the published audio directly: the speed-up is undone, the spectrogram is drawn from the measurement, and the same is done to three other NOAA recordings - an iceberg calving, harmonic tremor from a berg in contact with the sea floor, and Slow Down, which is a berg running aground. All four recordings play in the episode, unedited.

PRINT-READY, FROM THIS CHANNEL

Read This Dive, Vol. 01 — the method, and five dives read with it

https://therepository.gumroad.com/l/read-this-dive

PRIMARY SOURCES

NOAA PMEL Acoustics Program - Icequakes (Bloop). Recorded signal sped up 16 times; source region between the Bransfield Strait and the Ross Sea, possibly Cape Adare. https://www.pmel.noaa.gov/acoustics/sounds/bloop.html
NOAA PMEL Acoustics Program - Cryogenic (ice) spectrograms: calving, iceberg harmonic tremor, and Slow Down (iceberg grounding). https://www.pmel.noaa.gov/acoustics/specs_cryogenic.html
Internet Archive copies of the same NOAA page, read for this episode: 2007 and June 2012 versions ("the origin of the sound is unknown", source given as 50 S 100 W) and the January 2013 version (retitled "Icequakes (Bloop)").
Audio measured in this episode, all public domain: bloop.wav (sped up 16x), noise97139.wav / Slow Down (16x), A53JD33Calving_3x.wav (3x), HarmonicTremor2006 (3x).
Dive footage and frame grabs: NOAA Ocean Exploration, EX1811 dive 09, 9 November 2018, off Puerto Rico, hydrophone deployment. Public domain.
Antarctic ice imagery: NASA. Public domain.

ARCHIVE PLATES — public-domain documents, with author, date, archive and licence.

| NOAA Ocean Exploration | public-domain
| NOAA Ocean Exploration | public-domain
| NOAA Ocean Exploration | public-domain
JPL — 2013-11-15 | NASA / JPL | Public Domain (NASA)
NASA — 2001-11-14 | NASA | Public Domain (NASA)
GSFC — 2017-12-08 | NASA / GSFC | Public Domain (NASA)
JPL — 2001-01-03 | NASA / JPL | Public Domain (NASA)
| NOAA Ocean Exploration | public-domain
JPL — 2003-11-12 | NASA / JPL | Public Domain (NASA)

... and 5 more plates, each credited on screen in the video and listed in this project's manifest.

Abyss Report reads the deep ocean through the record that resolved it - instrument, date, position and the primary sources listed above. The evidence is the recording, not decoration.

About the material. All four sound recordings are NOAA PMEL Acoustics Program files, public domain, played unedited. Every spectrogram is drawn in code from a measurement of those same files, with the published speed-up undone - no spectrogram here is traced or redrawn from a picture. The dive footage and frame grabs are NOAA Ocean Exploration, EX1811 dive 09, 9 November 2018, credited on screen. The satellite imagery of Antarctic ice is NASA, public domain.

#TheBloop #NOAA #DeepSea #Acoustics #Antarctica