How it works
The complete audio watermarking scenario passes through three successive stages.
Data transmission
The audio signal carrying the hidden info is transmitted through a communication channel. This channel is an abstraction for all the degradations and attacks the watermark may suffer before extraction.
Possible attacks are: physical transmission through a noisy audio channel, digital to analog conversions, on air recording, resampling, recoding, lossy compression, cut and paste manipulations, etc.
Ideally, the watermark should survive to all those attacks.
Watermark extraction
The audio signal is processed at destination by a forensic tool that rebuilds and extracts the hidden information embedded in the first stage. Just the authorized user can extract the information.
If segments of the audio have been manually deleted, added or moved, the manipulation is detected.
AudioWatermark features
911 Emergency & Contact Centers
Embedding of the user ID retrieving a call recording, for keeping track of the source in case of leaking to media.
Detecting manipulations in sensitive call recordings.
Integrity check of legally relevant voice recordings
Detecting cut & paste manipulations in recordings of meetings, trials or contracts completed by phone.
Copyright control for movies & music
Embedding of the customer ID into the downloaded content from TV, movies and music streaming platforms.
Secret communications
Hiding information into phone calls or radio transmissions.
