Newshot Pinpointing Location of Videos from Background Airplane Sounds
This is fascinating:
The video describes a technique that can reveal where an otherwise unidentifiable recording was made by analyzing its audio for the faint sounds of passing aircraft. The technique is both fascinating and frightening. It works by looking at Doppler curves, which plot the way a plane’s sound changes as it passes overhead because of the Doppler effect. This is the same familiar Doppler effect that causes the frequency of a siren to rise as it approaches and then fall as it passes by and moves away. More generally, it’s the phenomenon whereby the frequency of a wave perceived by an observer changes when the observer and the source of the wave are moving in relation to one another. […] It’s sound that’s relevant here, though, and specifically the sound of aeroplanes. Early in his video, Colsto brings up a map of the flights that were in the air at the time the video he’s tracing was recorded. Much of this information is publicly available. Aircraft equipped with Automatic Dependent Surveillance–Broadcast (ADS-B) transmit data about their position, velocity, and altitude, which flight-tracking services collect and make available to the public. While not every plane is tracked, the system provides an enormous amount of information about aircraft movements.
The video describes a technique that can reveal where an otherwise unidentifiable recording was made by analyzing its audio for the faint sounds of passing aircraft. The technique is both fascinating and frightening.
It works by looking at Doppler curves, which plot the way a plane’s sound changes as it passes overhead because of the Doppler effect. This is the same familiar Doppler effect that causes the frequency of a siren to rise as it approaches and then fall as it passes by and moves away. More generally, it’s the phenomenon whereby the frequency of a wave perceived by an observer changes when the observer and the source of the wave are moving in relation to one another.
It’s sound that’s relevant here, though, and specifically the sound of aeroplanes. Early in his video, Colsto brings up a map of the flights that were in the air at the time the video he’s tracing was recorded. Much of this information is publicly available. Aircraft equipped with Automatic Dependent Surveillance–Broadcast (ADS-B) transmit data about their position, velocity, and altitude, which flight-tracking services collect and make available to the public. While not every plane is tracked, the system provides an enormous amount of information about aircraft movements.
Tags: air travel, geolocation, videos
Posted on November 6, 2026 at 9:57 AM • 19 Comments
Comments
Wicked Lad • November 6, 2026 12:16 PM
At the risk of stating the obvious, it seems this would also be useful in evaluating the likelihood that a video was taken where and when it is purported to have been taken.
Clive Robinson • November 6, 2026 1:03 PM
The idea is not exactly new as the concrete sound mirrors on the UK south coast from WWII will evidence.
But it goes back even further to attempts in WW1 to track submarines around the UK Coast, that has developed into a system to detect Russian subs as they go around various parts of Europe to get into the Atlantic.
But from another aspect there is the use of “mains hum” on recorded audio giving a very precise time indicator.
The problem with the idea is that you need a “starting location” as well as “time”. It’s the same problem that current “Global Navigation Satellite Systems”(GNSS) have,
https://www.advancednavigation.com/tech-articles/global-navigation-satellite-system-gnss-and-satellite-navigation-explained/
You have to solve a series of time / distance based “triangles” whilst also allowing for the propagation speed of the signals which is of course variable… Not just due to Doppler Shift but the characteristics of the communications channel.
The technique is generally known as “Trilateration”,
https://en.wikipedia.org/wiki/Trilateration
And starts with an assumption of a starting point location and time then iteratively work through untill you get a probable location match then keep refining untill you confirm and get the desired accuracy.
Interestingly you can do similar with VHF and UHF transmissions from broadcast stations or “Private Mobile Radio”(PMR) or “Ham/Amateur Radio”.
Basically the signals bounce of aircraft, and get crossmodulated by the aircraft physical movement with respect to the receiver.
As far as I’m aware this was first demonstrated in Britain by a couple of amateur radio engineers that gave rise to RADAR.
The credit for which was stolen by Robert Watson[1], who later hyphenated Watt to his family name.
The official story is two men were Robert Watson-Watt and Arnold Wilkins, drove into a field near the BBC transmitters at Daventry and set up a couple of antennas and using an oscilloscope detected reflections off of a Bomber flying by.
The reality is others had discovered the effect and talked about it and it got to the ears of Arnold Wilkins who was Watson’s assistant. But the only reason Robert Watson got involved was of his job measuring atmospheric effects we now call “propagation”[2] People in the Air Ministry were informed about the effect of bouncing signals off of aircraft and identifying their direction and they then asked for it to be investigated. Thus the job fell to Watson because he was on the public pay role in the nearest relating research role…
This much can be confirmed because of the flight of the bomber, was “officially arranged” before the supposed “experiment” that was in reality just a “demonstration” of what had been established experimentally by others some time before.
[1] During WWII my mother had the misfortune to work with Robert Watson-Watt… She really did not have a good word to say about him, and had quite a few stories about his appalling behaviours.
[2] Robert Watson does however have one claim to fame, he named the “ionosphere” whilst working for the Met Office in Slough South England that was the early location for the National Physics Laboratory. Where he was supposedly trying to track weather by lightening storm. Something he finally actually succeeded at doing, however it had problems due to the fact “the antenna could not be swung fast enough”. This was solved by the patented in 1919 work of British radio engineer Frank Adcock, who whilst an Army officer in the British Expeditionary Force”(BEF) in France invented what became named “The Adcock aerial / antenna array”. But it was in of it’s self still incapable of getting the direction fast enough. This was solved by acquiring an X/Y display Oscilloscope from the US. It was this system that was taken to the field in Daventry…
Confused • November 6, 2026 5:52 PM
if one were to wipe all metadata, which includes the time when video was shot, what would then be the indicator of time-frame correlating to any flights?
lurker • November 6, 2026 7:27 PM
“The technique is both fascinating and frightening.”
Yet another first world problem solved – or is it? To use this method, the video needs to have been shot close enough to an air route, and with background noise low enough to discern aircraft noise. It’s not high on my list of things to be frightened of.
Wharrrrrrgarbl • November 6, 2026 8:51 PM
@Confused, the video plays out a few scenarios. One is a long video (more than 900 seconds) with a known date of recording and basically no location data besides “in the continental USA”. Another knows when the video was recorded precisely but has only a single doppler trace to analyze. Both were able to locate the camera the video was taken from without referring to really any visual data. What correlates the flights is the doppler data that establishes a spatial relationship and the time gap between the doppler signals that constrains the second flight passing over the location.
Clive Robinson • November 6, 2026 10:57 PM
@ Confused, ALL,
You ask a quite sensible question of,
blockquote>if one were to wipe all metadata, which includes the time when video was shot, what would then be the indicator of time-frame correlating to any flights?”
@Wharrrrrrgarbl gives you an overview of what’s in the link.
But there are a lot of others, some of which are also useful as tools in their own right.
Any real world video/audio recording contains “interference / noise” signals that are “Man Made”(QRM) or “Natural”(QRN) in origin, they are an unavoidable fact of life.
Both appear in the audio and video sensor inputs as well as from the power supply to the recording equipment.
These can be “pulled out” of the other noise and from under the “wanted signals” in the video.
A couple of non-aircraft signals that will or can be “pulled out”,
An example of QRM is “mains hum” and this can give you “time” to within 1/800th or better of a second with correct processing and a course indication of location by “the grid”.
It was used long ago in the UK by Police to check that “audio recordings” were both continuous –ie not edited– and when they were made.
An example of QRN is sunlight and shadows. It’s usually best to get an astronomer, sailor or surveyor to explain the intricacies of how it works and to do it but even on cloudy days the sun can “make a sundial” with the shadows of objects. This can give both latitude and longitude quite well.
Another example of QRM which is less accurate but can be sufficient for both time and approximate location is vehicular and other noise that is fairly unique and runs on a time table. This is the one that is almost a trope in the way it gets used in detective movies and CSI TV shows.
But there is also QRN that is “animal noises”. We don’t notice it in cities or mainly urban environments but the passage of time is marked by creatures including humans. The most obvious is the now very rare “Cockerel Crow” but most farmyard/livestock makes noise of some kind based on time. Be it from the Sun or from those who feed them etc.
As days change with both the time of year and latitude animal noises can also give an approximate location.
There are other effectors but they may only be of use to certain people. The electrical supply grid in the US and other places is often put up on wooden poles with cables hanging in between. The thing is they deteriorate reasonably fast thus arcing and other electrical noise can be in both the audio and visual signals of a video recording. Such signals can pin down a time and location if the electrical infrastructure supplier are doing their job…
These are just some of the potential “forensic tells” not just in recordings but in those that have been “faked up”.
Privacy • November 7, 2026 8:43 AM
Weather plays a role in how sound travels, specifically temperature of air at different atmospheric altitudes. When I lived in a small town many miles from a major highway, sometimes the weather was such that at night at home I could hear tractor-trailer trucks on the highway as clearly as if I were standing by the side of the highway.
Regarding ADS-B, the public sites work well enough that I could see a plane passing overhead as I watched it on a map and view data of the flight origin and destination, airline, etc.
Privacy • November 7, 2026 9:15 AM
By the way, I suppose AI could make the technique described in the article much more accessible and accurate and automated.
Ferentarius • November 7, 2026 12:31 PM
This is both impressive and a little unsettling. Leveraging publicly available ADS-B data and the Doppler signature of passing aircraft to geolocate a video shows just how much open-source intelligence can reveal. It’s a great example of how creative analysis of widely accessible data can surface insights most people wouldn’t think possible. It’s also a reminder that what seems anonymous online often isn’t—ambient details can give away more than you expect.
Q • November 7, 2026 7:23 PM
I watched the video and was dismayed at the conclusion that the audio can’t be modified to make the technique fail.
But the options tried were only to attempt the filter out the Doppler sounds. No attempt was considered to add in lots of fake Doppler sounds.
If it can’t be filtered then just add more of it and confuse the classifier. SNR goes way down.
anon • November 8, 2026 3:15 AM
Will this be more or less accurate than using the high voltage hum from the power lines?
Clive Robinson • November 8, 2026 8:35 AM
With regards,
“If it can’t be filtered then just add more of it and confuse the classifier. SNR goes way down.”
Unfortunately lifting the noise floor with “fake signals” nolonger works for various reasons.
Primarily because the fake signals will always have some kind of distinguisher you won’t be aware of or can not account for.
At the very least a fake signal will not tie in with the ADS-B and acoustic signature of the environment (think multipath and selective absorption).
So if you record signals from somewhere else and add them in they will be measurably incorrect for genuine signals in the recording.
Clive Robinson • November 8, 2026 9:37 AM
You ask the question,
“Will this be more or less accurate than using the high voltage hum from the power lines?”
Like most things involving transducers the answer is “it depends”.
Basically transducers such as microphones and image sensors have limitations no matter how expensive they are.
Three important aspects of this are,
1, Bandwidth 2, Linearity 3, Phase/frequency
You would need to carefully study the characteristics of both signal types and what part of the signal you are not just measuring but actually using.
But, due to ADS-B being a requirement for nearly all aircraft[1] and in effect all records are publicly available. This makes getting reliable data fairly easy.
Whilst ordinary “mains hum” is effectively public to anyone on that grid at the time[2], as a general rule the historic data is not made openly public. But what is almost certainly not “public” is “fault data” such as arcing of switch gear and worn out infrastructure that has not been sufficiently maintained. Whilst the infrastructure suppliers will have records they will be “company confidential” “business records”. So whilst available to “authorities” not available to “the public”.
[1] Whilst ADS-B does many things it is primarily used to do “flight safety” activities hence the idea it is “mandatory”. But it requires considerable equipment and power to have it in an aircraft and not all aircraft can have it fitted (think “tail-draggers” that are more than twenty years old, unpowered aircraft like gliders, and experimental aircraft like microlights). Thus there is an exemption for “non-electric” aircraft, and also for aircraft in hostile / war / exclusion areas.
You can find out more about the rules for the US via the FAA website,
https://www.faa.gov/about/office_org/headquarters_offices/avs/offices/afx/afs/afs400/afs410/ads-b
[2] Using “mains hum” in the sense of “mains frequency” will mostly only give you a “time stamp” not a location. But… As the mains frequency is “load dependent” the closer you are to a step change in load the sooner you will see a change. That is the load change is a little like throwing a stone in a pond, it’s effect “ripples out”. In theory –I’m not aware of anyone testing it– such step change ripples can be used to provide “Direction Finding” thus a locality. However “mains hum” also includes signals generated by faults and breakdowns such as arcing. As “high frequency” signals they tend to remain “local” to portions of the grid as the distribution transformers have very low frequency bandwidths and don’t allow the signals to propagate. Thus give quite localised location data.
Privacy • November 10, 2026 7:16 AM
Clive, I used to sell fault recorders and protective relays to electric utilities and repair those. You are correct that arcing can be localised. Broken conductors can be localised. Mains frequency can be localised. Fault data was timestamped. The utilities used the timing to determine distance-to-fault so a repair crew could be dispatched to the fault location. The above equipment triggered on arcing, overcurrent, voltage, loss of load and frequency getting out of narrow limits. Inputs came from CVTs and CTs.
Winter • November 10, 2026 3:49 PM
Primarily because the fake signals will always have some kind of distinguisher you won’t be aware of or can not account for.
In general, what can be measured, can also be faked.
The way the bad guys get around DNA evidence is by emptying a full hoover bag of dust over the crime scene. Fingerprints can be copied and left at the scene of a crime. Blood and DNA can be removed with bleach, or added from storage.
If you even make a modicum of visiblebeffort, you will raise doubts about all signals found.
But the unaware are easily caught. Like after the first cave recorded video message of Al Qaida. Geologists identified the rocks in the cave walls and could pinpoint the location well enough for air bombardments.
Winter • November 10, 2026 3:59 PM
Primarily because the fake signals will always have some kind of distinguisher you won’t be aware of or can not account for.
In general, what can be measured, can also be faked.
The way the bad guys get around DNA evidence is by emptying a full hoover bag of dust over the crime scene. Fingerprints can be copied and left at the scene of a crime. Blood and DNA can be removed with bleach, or added from storage.
If you even make a modicum of visiblebeffort, you will raise doubts about all signals found.
But the unaware are easily caught. Like after the first cave recorded video message of Al Qaida. Geologists identified the rocks in the cave walls and could pinpoint the location well enough for air bombardments.
Clive Robinson • November 10, 2026 5:05 PM
You make the statement of,
“In general, what can be measured, can also be faked.”
Without qualification.
Yes you could make all manner of measurements but two questions arise,
1, Have you made all the necessary measurements required? 2, Are there the resources to fake all the results sufficiently?
All without creating new differentiating signals to trip up on…
Winter • November 10, 2026 5:55 PM
All without creating new differentiating signals to trip up on…
It’s an arms race. Technology can identify people on their voices alone. There are actual challenges organised to spoof automatic speaker recognition, and challenges to foil the spoofers.
In the end it is a matter of resources. And technology tends to make things cheaper over time.
What yesterday was exclusively the field of the TLAs, is today the weapon of organized crime, and tomorrow the tool of the script kiddies.
Btw, filtering out 50/60Hz hum and harmonics is not rocket science with modern tools.
See: ‘https://www.audio-editor.online/blog/how-to-remove-hum-from-audio
minor correction for Q • November 11, 2026 7:49 AM
Q: “I watched the video and was dismayed at the conclusion that the audio can’t be modified to make the technique fail.”
FWIW he was not saying that it’s impossible to modify the audio to make the technique fail.
Rather, he demonstrated why several natural/obvious methods that one would naively/intuitively expect to work won’t in fact work.
Clearly you CAN modify the audio to make the technique fail, e.g. by replacing the audio with some other audio (perhaps even re-recording your own speech from a different location later). 🙂
Subscribe to comments on this entry
Leave a comment Cancel reply
Blog moderation policy
Sidebar photo of Bruce Schneier by Joe MacInnis.
Powered by WordPress Hosted by Pressable