Research On Synchronous Blind Audio Watermarking Technology Based On Wavelet Package Transform And Psychoacoustic Model

With the rapid development of Internet and digital audio coding technology, digital audio signals become much easier to transmit and obtain, so people's lives become rich and colorful. But the pirate problem came with it becomes more terrible, the copyright protection has become more and more important. The digital audio watermarking technique is an effective implement for the copyright protection of audio signal. It embeds some identified information as watermark into original audio signal imperceptibly, and verifies the property and integrity from detecting or abstracting watermark. This paper mostly combines with human auditory masking and synchronous coding technique, and makes some research on robust digital audio watermarking technique.The paper firstly analyzes the basal theory and arithmetic of digital audio watermarking techniques, and also gives a general summary of current researching condition. Then combining with human auditory masking of HAS (Human Auditory System), the paper improves the imperceptibility of traditional digital audio watermarking techniques which are based on wavelet analysis, and proposes a new audio blind watermarking algorithm based on wavelet packet transform. Finally, the paper proposes an audio blind watermarking algorithm with self-synchronization, which could resist familiar synchronous attacks in time-domain of the audio and further improve the robustness of watermark.The main working content as follows:1．The paper introduces and analyzes existing digital audio watermarking techniques, and expounds of researching background and meaning. Furthermore, it analyzes and compares the typical robust watermarking algorithms based on instances at present, and analyzes their characteristics. These works lay a foundation for my further job.2．The theories of wavelet analysis and human auditory masking effect are introduced and analyzed, and their applications in the digital audio watermarking techniques are also deeply investigated, then our paper propose an improved audio watermarking algorithm using auditory masking threshold to determine the embedding positions of watermarks. This algorithm firstly analyses the character of power spectrum in wavelet packet domain for audio masking threshold, then locates the imperceptible signals with audio masking threshold, finally chooses some signals with bigger power to imbed watermarks. This algorithm offset the disadvantages of traditional algorithms in selecting embedding positions of watermark, such as poor definition and too subjectivism. This algorithm obviously improves the imperceptibility of watermark with the precondition of ensuring the robustness..3．Because the weak robustness of traditional audio blind watermarking to resist synchronous attacks, the paper proposes an audio watermarking algorithm based on m-sequence to against synchronous attacks. This algorithm uses m-sequence as synchronous code to embed into frequency-domain of audio signal, and embeds the sequence no with watermark into signal. In order to locate and rebuild abstracted watermark, this algorithm detects synchronous codes and watermarking no at the same time. This algorithm strengthens the robustness of watermark to against synchronous attacks such as sample cropping and copying in time-domain.