The Research And Implement Of Audio Digital Watermark

The audio digital watermark technology is one of the research focuses in recent years through embedding secret information into digital audio signal on purpose of copyright protection, information security and so on. It can be proved through detecting the watermark when owner's rights and interests are infringed.With a large amount of facture and issuance of the digital video product and music products,, the copyright protection of the audio data seems to be more and more important too. Through embedding watermark information into audio carrier, the function of copy restriction, use track and the pirate confirmation can be realized. In recent years, the research about the audio digital watermark technology is developed quickly. Especially the technology of audio digital watermark in the transform domain, make research have practicability because it can embed information to sensitive area of the carrier. This paper focus on the audio digital watermarking technique in order to protect audio file's copyright.The thesis looks back on the development history of the information hiding firstly. There after the thesis summarizes the characteristics of Human Auditory System (HAS).Then it indicates that people can make use of these characteristics to hide the secret data in a public audio. described Wavelet analysis and Cosine analysis, then focused on the principles of those techniques.According to the theory presented above, put forward a digital blind watermark algorithm in Wavelet transform and Cosine transform based on Hilbert curve. In order to improve the robust of watermarking algorithm, this algorithm introduce the discrete cosine transform into the discrete wavelet transform, fully uses their merit, The visually watermark is selected as watermark information, First scans the binary watermarking image through the Hilbert minute shape curve, converted it into a one-dimension sequence, the audio signal was divided into segments and through transforming the original audio signal (discrete cosine transform and discrete wavelet transform).Then the transformed watermark was embedded into the transform domain in the quantification process. The watermark can be extracted without the original audio signal. This helps it to use in the actual project. Experimental results demonstrate the algorithm is robust and imperceptible.