Earshot

Sonic investigations for communities affected by corporate, state, and environmental injustice

What we do

Expand

Sound Authentication

Earshot has developed audio analysis techniques to verify the authenticity and integrity of video or audio files acquired by organizations, journalists, activists, and community groups. Our forensic audio analysis uses advanced computer-based pattern recognition tools to verify the audiovisual material under scrutiny. Our work has been useful to human rights investigators in assessing whether recordings from the public or whose origin has not been verified have been unedited or manipulated.

Case study: Debunking video of alledged nurse at al-Shifa hospital in Gaza

Play video
The truth behind the al-Shifa Hospital nurse video | Fact Check
The truth behind the al-Shifa Hospital nurse video | Fact Check
Earshot carried out sound authentication for Al Jazeera's investigative unit to verify the authenticity of footage published by the IDF about an alleged nurse claiming that the Al-Shifa hospital in Gaza was controlled by Hamas.

Case study: Electromagnetic verification 

Sound Authentication electromagnatic-verification.png

The Organised Crime and Corruption Reporting Project (OCCRP) approached us with concerns regarding the authenticity of leaked wiretapped recordings containing evidence of illicit activities involving government officials and their relatives. These concerns stemmed from suspicions that authorities might be disseminating counterfeit documents to undermine journalists. In this critical juncture, our audio verification services proved invaluable.

We used various techniques, including the examination of electromagnetic interference patterns. Electromagnetic interference creates distinctive rhythms of pulses and tones recurring at precise intervals. Any interruption of discontinuity in these patterns could indicate tampering. Our analysis unveiled dual low-frequency pulses audible in the right channel of the stereo signal, lasting for 100 milliseconds, at regular 2.13-second intervals (as depicted in the figure above). The persistent and uniform pattern strongly suggested that no alterations had been made to the audio signal during the period when this interference was detectable.

Two other sources of constant electromagnetic interference were mapped and when we combined the three sources of interference, we found that they covered the whole period of the recording and could therefore provide OCCRP with the verification that this recording is original and has not been altered.

Featured investigations

Expand

Press

Expand

Partners

Expand