A robust audio watermarking system for acoustic channels

Digital audio watermarking is one of the many technologies that evolved in response to the recent fast expansion in the area of digital information transmission and multimedia. In fact, the possibility to robustly hide information in an audio signal has many applications, such as protection of intellectual property, broadcast monitoring, and metadata transmission. This article presents an innovative method for watermark embedding in the time-frequency domain using a sparse multicarrier approach. The proposed system involves a watermarked audio signal that is played back by a loudspeaker and picked up by a microphone mounted on a mobile device. The system is built to be used in an enclosed environment such as a room and is therefore designed to be robust against various impairments such as reverberation, microphone self-noise, and movement. The system robustness is confirmed by performance evaluations, whereas its inaudibility is verified by subjective listening tests