Audio Watermarking Based On MPEG Psychoacoustic Model

With the rapid development of the field of digital media,issues of copyright protection are paid more and more attention to.Digital watermarking has been proposed as a means to identify the owner or distributor of the digital data.Watermarking is the process of encoding hidden copyright information in digital data by making small modifications to the samples.Most schemes utilize the fact that digital media contain perceptually insignificant components which may be replaced or modified to embed copyright protection.However, the techniques do not directly exploit frequency masking.Thus,the watermarking is not guaranteed inaudible.In this paper,we propose a novel audio watermarking scheme which exploits the human auditory system to guarantee that the embedded watermark is imperceptible.The algorithm description is as follows:The audio signal is divided into segments.For each segment,a masking threshold is calculated according to the psychoacoustic model in MPEG-1,Layer I.And a pseudo-random noise sequence is achieved via direct spread-spectrum system method.To enhance of the performance of error correction, convolution coding and interleaving is used.The noise sequence is also divided into segments,which is colored by the filter approximating the masking threshold.The original signal is watermarked by adding the colored noise.The project described hereafter contains MATLAB simulation.The watermark is perceptual transparent,which was confirmed throughout listening tests.Furthermore,we will show in the sequel that the watermark is extremely robust to some signal processing operations such as re-sampling,modification and adding white noise.Our next work will focus on a satisfactory tradeoff between a robust watermark and a nature audio.