A Study On Dual Digital Audio Watermarking Algorithm Based On Lifting Wavelet

The rapid development of network and multimedia technology provides a very largespace for the storage and transmission of data and information, it is not only provideconveniently but also brings some negative effects for the human. For example: the tort of thework, distorting and other problems and so on. Especially in recent years, more and moreaudio informations can be edited and copied etc. This problems can lead to much heavily lostfor the owner of the copyright. To sum up, as an effective solution, the digital watermarkingtechnology has received increasing attention and concern from scholars at home and abroad.This paper summarizes the technology and principle which is closely related to the work ofdigital audio watermark, introduce the typical audio watermarking algorithm, reviews theattack methods of replication and evaluation criteria, and gives the main content of theresearch.Firstly, this paper presents an algorithm of dual digital audio watermarking based onLWT is proposed in this paper. This algorithm is used in the original audio liftingwatermarking embedded two watermark, a watermarking adaptive to choose embeddingstrength. This original audio watermarking embedding of low frequency domain to protectaudio copyright. Another watermarking adaptive to choose embedded position, the originalaudio watermarking embedding of high frequency domain so as to play a dual protection. Atthe same time, before the embedded watermark image, scrambling and the chaos encryptionin order to improve the safety of the algorithm. Experiment proves, the algorithm has betterwatermark. The experiment results show that the watermark is imperceptible. After all kindsof attacks, watermarking system still has strong robustness.Secondly, an algorithm of double track of dual digital audio watermarking based onLWT is proposed in this paper. First, this algorithm selects a double track audio, extracts theleft and right channel, and decomposes the left and right channel on the way of LWT3. Thispaper selects the difference of the left and right channel of low frequency on LWT3as thefirst watermarking. The image is regarded as the second watermarking. Then this paperchooses the embedded position and strength adaptively into the left and right channel. Thispaper can realize the dual watermarking embedding and extracting. The results ofexperimental show that this algorithm has strong robustness for all kinds of attacks such asnoise and low pass.