A robust audio watermarking technique based on the perceptual evaluation of audio quality algorithm in the multiresolution domain

Abstract

Audio watermarking is a method that embeds inaudible information into digital audio data. This paper describes an audio watermarking scheme using the Least Significant Bit (LSB) in the multiresolution domain exploiting the perceptual evaluation of audio quality algorithm (PEAQ) and MP3 compression characteristics. To guarantee perfect inaudibility propriety, we adapt the PEAQ algorithm to select the adequate audio band used to hide the watermark in the LSB of the multiresolution coefficients. Additionally, we use some MP3 compression characteristics to increase robustness of the scheme and we duplicate the watermark bits in the whole audio signal blocks assuring then an important ratio property without affecting the inaudibility property. The described method provides good robustness results against compression and StirMark attacks. Besides, it allows blind retrieval of embedded watermark which does not need the original audio.