Research And Performance Test Of Digital Audio Watermarking Algorithm

With the rapid development of the Internet and information technology,various types of multimedia data become more and more convenient and fast in the process of production,storage,and distribution,which brings great convenience to people's life and work,but at the same time,there are also phenomena of infringement,piracy and malicious tampering of multimedia digital products,and in order to solve this problem,digital watermarking technology comes into being.Digital watermarking technology is an important branch in the field of information security,among which,digital audio watermarking technology aims to protect the copyright of digital audio products.An effective digital audio watermarking system needs to meet three requirements:imperceptibility,robustness,and capacity.To address the copyright protection of digital audio information,this paper proposes two digital audio watermarking algorithms that can have good robustness while maintaining certain imperceptibility and watermark embedding capacity.Firstly,we propose a discrete wavelet transform domain digital audio watermarking algorithm based on norm ratio,the proposed algorithm will first carry out discrete wavelet transform after framing the original audio carrier signal,construct the obtained low frequency coefficients into two vectors and calculate the vector norm,and finally realize the purpose of embedding watermark information by quantizing the vector norm.Secondly,we propose an adaptive digital audio watermarking algorithm based on the norm ratio of the lifting wavelet transform domain,which performs the lifting wavelet transform on the framed audio signal,the low-frequency coefficients are selected to construct two vectors and calculate their norm ratio,and embeds the watermark in the carrier audio signal dynamically by quantizing the norm ratio and taking the maximum value of the signal-to-noise ratio as the constraint.Extensive audio experimental simulation results show that the two digital audio watermarking algorithms proposed in this paper have good imperceptibility,meet the watermark embedding capacity requirements,and show good robustness against conventional signal processing attacks,including additive noise,resampling,requantification,low-pass filtering,and MP3 compression.In addition,in watermark preprocessing,the watermark information is converted into binary images,which improves the robustness of the watermarking system by increasing the information redundancy to some extent.The Arnold transform is used to scramble the watermark,and the watermark information is also encrypted by logistic chaotic sequences to further improve the security of the watermarking system.