Research on digital audio watermarking algorithm based on lifting wavelet transform and chaotic keys

This paper proposes an audio watermarking algorithm that makes use of lifting wavelet transformation and chaotic encryption. We first perform the lifting wavelet transformation on the audio signal. By modifying the statistical mean value of the lifting wavelet parameters, we can acquire a synchronization code and the information necessary to embed the watermark. We can also adaptively ensure the best embedding strength by using masking thresholds. This guarantees that the embedded information cannot be perceived by human ears. In order to improve the security of the watermark, we make use of improved logistic chaotic mapping to generate a ciphering key sequence that can hide the messages. Results show that the method of this paper has better transparency and security than the traditional audio watermarking schemes based on the fourier transformation, the cosine transformation and the wavelet transformation. Additionally, it is more effectively robust against the attacks of various common signal processing.