An Algorithm For Synchronous Wavelet Audio Watermarking Based On Self-adapted Energy

The major content of the digital watermark focuses on the algorithms that insert the secret information into the host information. The host information may be the text, image, audio (speech) etc. Recently, with the development of information digital watermark, the audio (speech) information hiding technique is becoming a hot topic. It can be used to protecting copyright, hiding communication by embedding the secret information into the audio file.A number of basic theoretical problems in digital watermark are studied first. The general concepts and fundamental principles of digital watermark and secure communication are presented; possible directions for further research are pointed out also. The possibility of using speech as the host carry signal is analyzed. A self-synchronized watermark algorithm of wavelet domain is presented in which the energy is self-adapted. The short-term energy analysis of audio signals is proceeded, then according to the energy distribution to choose the location for watermark information embedding. After that, based on the logistic function of chaos model, BCH code of watermark is given, and then the frequency expanding of the watermark information are carried out to make it self-synchronized to clock and data concerned so to acquire better ability of anti-obstruction and anti-interception. Meanwhile, embedding watermark in the low-frequency components of wavelet domain can better improve the robustness of watermark.The experimental results show that the algorithms are blind.The audio signal is chosen as the carry signal. The experiment results show that there is no obvious difference between original speech and the speech with secret information. The secret information is robust. After enduring filtering, interpolation, decimation, and noise, compression, the watermark information can be well extracted.