top of page
fake audio.png

Fake Audio Detection

Many times before the audio is grabbed by police, it is tampered by the suspected person so that his identity will not be disclosed. The culprit may concatenate number of audios and in the process some dc shift, phase shift or pause may get introduced in the audio. All these types of tampering is detected by the system.

The system also detects if it is a real recorded audio or a deep fake means a synthesized audio using some text to speech engine.

salient features of the system:

  1. Can detect pause, dc shift, phase shift etc

  2. Detects ENF or background noise variation in case of concatenation of to audios recorded on two different mobile devices.

  3. Can detect original sampling frequency and metadata for the audio

  4. Can detect if the audio is real recorded or synthesized

  5. Has automatic report generation facility

bottom of page