Robust Audio Watermarking Algorithm Against Re-recording Attack

In recent years,with the development and popularization of the Internet and multimedia technologies,people are generating information in digital form almost every moment in their daily lives.In addition,the development of high-speed networks has made it particularly fast and convenient for people to upload and download these digital information.At the same time,due to the convenience and high fidelity of the reproduction,editing and transmission of digital information,the creation,distribution,use and reuse of digital information has also become simple and convenient in the past ten years.However,while the explosive growth of this information has brought convenience to the entire society,people are also worried about the threat of illegal copying,distribution and use of digital information and the characteristics of the network.And with the popularity of smart phones,tablet computers and voice recorders,ordinary users can easily obtain copyrighted audio.After the copyrighted audio is dubbed,the watermark information may be attacked and cannot be completely extracted,thereby infringing the interests of the copyright owner.Therefore,the research on the copyright protection of digital multimedia data has become a research hotspot of multimedia information security.Based on the wide application of digital audio signal in radio,television,film and video companies,secret communication,and network security,as well as the actual demand for audio signal copyright protection and traceability,this article takes digital audio signals as the research object,and resists duplication and removal of digital audio signals.Synchronous attack technology was studied.The specific work is as follows:(1)In order to solve the shortcomings of the traditional digital audio watermarking technology under the re-recording attack,a source tracking audio watermarking algorithm is proposed to resist re-recording attacks.The definition of the frequency-domain coefficients logarithmic mean is given,and a robust audio watermarking algorithm based on the feature of frequency-domain coefficients logarithmic mean is constructed.The frame number and copyright information are embedded into the carrier signal together,and the audio frame watermarked by the frame number is synchronized by the extracting terminal,and the copyright information is further extracted to verify the owner of the audio content and trace the source.The algorithm has strong robustness to some signal processing operations,and at the same time has certain resistance to re-recording attacks.(2)On the basis of the definition of the frequency-domain coefficients logarithmic mean in the previous part,a method for separating the silent and non-silent segments of the digital voice signal based on this feature is given.The watermark information is embedded into the characteristics of the residual of frequency-domain coefficients logarithmic mean of two adjacent speech signals,and a robust speech watermarking algorithm against duplication and desynchronization attacks is proposed.The test results prove the inaudibility,robustness,and high system security of the algorithm.