YouTuber Reveals How Sounds Humans Cannot Hear Can Expose a Video's Exact Location

A YouTuber has demonstrated how inaudible aircraft sounds hidden in a video's background can help reveal where it was recorded, even when there are few visible clues. By analysing subtle changes in sound frequency and comparing them with aircraft tracking records, creator colsto says he located a security camera in the continental United States to within a few hundred feet.

The investigation began with a blurry recording showing little more than trees and a riverbank. The only information available was that it had been filmed on 8 September somewhere in the continental US. The creator identified two aircraft sounds in the background, separated by 920 seconds, and used their flight paths to narrow down possible locations.

He then compared the predicted sound patterns with the original recording, identified a leading location and checked it against Google Street View. The tree shapes and a utility pole visible in the online imagery matched the footage, allowing him to identify the security camera.

How Aircraft Sounds Can Reveal a Recording's Location

The method relies on the Doppler effect, which changes the frequency of a sound as its source moves towards or away from a listener. Although the aircraft sounds could not be heard during normal playback, their patterns appeared on a spectrogram, which displays frequencies over time.

The creator used Automatic Dependent Surveillance-Broadcast (ADS-B) data to examine aircraft positions, speeds, altitudes and flight paths across the continental United States. He said collecting and processing the information involved about 50GB of flight data and more than 5,000 lines of code.

His calculations also accounted for the time sound takes to travel from an aircraft to the ground. Because aircraft altitude is measured above sea level, he incorporated terrain elevation data to calculate the three-dimensional distance between each plane and possible recording locations.

The first filtering stage identified about 70,000 candidate points. By examining the 920-second gap between the two aircraft sounds and ruling out locations where another flight should have been audible, he reduced the number to approximately 15,000.

He then modelled the Doppler patterns each aircraft would produce from different locations and compared them with the spectrogram. The highest-scoring candidate emerged ahead of the alternatives, giving him a much smaller area to investigate.

Separate Test Also Identifies the Location

The creator carried out another test using a separate one-minute recording with a known time. He applied the same general approach to a different aircraft signal and compared the results with flight data.

The location that scored highest in this test was the same one identified through the original two-aircraft analysis. The creator presented this as a separate check using different flight information.

What the Demonstration Means for Online Privacy

The video also examined whether common audio-processing methods could remove the signals. The creator tested high-pass filtering, background noise and telephone compression, finding that the aircraft patterns remained detectable in his demonstrations.

He said cutting sections from a recording or applying additional noise cancellation could make the method harder to use. However, he cautioned against treating the technique as an immediate threat, saying the chances of someone using it to locate a person were slim and that easier methods could be available.

The demonstration highlights how background audio may contain location information that viewers cannot hear. It does not establish that every video can be traced, as the method depends on suitable aircraft signals and usable flight data.