Digital Audio Watermarking Based On Wavelet Analysis Technology Research

With rapid development of the Internet and digitalize of media resources, digital information piracy arise, the multimedia information security problem has become a very important and challenging research topic. Digital watermarking technique is a potentially effective method, which to protect the copyright of digital products. It has become a research hotspot in information security field. There are important theoretical and practical significance in researching the key technology of digital watermarking.Digital watermarking mainly focus on image and video watermarking nowadays. Embedding watermark in audio signal is difficult, as human auditory system’s sensitivity and relevant aspects of this technology are not very well, we choose digital audio watermarking as the research subject. We select to embed watermark in the wavelet domain after learning and analyzing wavelet analysis theory.Firstly, we design a adaptive quantization embed audio watermarking algorithm in wavelet domain, as the advantages based on wavelet transform in signal processing, the audio signal use the quantify methods to get the visual image watermark embedded in the low frequency coefficient after wavelet transform. Using encryption technology to improve security, it means to adaptive select the quantization step in wavelet low frequency coefficient. It’s a blind watermarking algorithm, when extract watermark without original audio signal. Then we improve the algorithm, design it to embed watermark into the low frequency coefficients with different length of quantization interval, as masking characteristics of human auditory system, we also quantify embed watermark in wavelet high frequency coefficient, this dual watermark embedding method is also a blind one, and it also has a certain robustness.Experiment simulate the above algorithm and the improved one, including the watermark embedding, extraction and extraction after various attacks, these algorithms respectively have transparency, security and robustness watermark detection, through six different types of audio samples, we use experiment objective quantitative evaluation its performance, including NC, SNR, MSE and PSNR. Simulation results show that these watermark systems have good robustness after variety of attacks, wavelet transform get watermark’s imperceptible, and the two algorithms are both feasible digital audio watermarking methods.