A Novel Audio Scrambling Algorithm in Variable Dimension Space

Conventional audio scrambling methods use one-dimensional linear congruence mapping in real time. The combination of encrypting parameters of those methods is too less which decreases the security, and the one-dimensional linear congruence mapping is easy to be attacked. To address these problems, a novel audio scrambling algorithm in variable dimension space is presented in this paper. The proposed method increases the dimension of sample points coordinates and uses arbitrary matrices to scramble audio, which increases the difficulty being attacked. Experiments show that the proposed algorithm is validity in audio scrambling and easy to implement. It is robust in enduring common attacks including noise addition, filtering, resample and so on. The proposed algorithm can be used in audio information hiding, preprocessing and post-processing of audio watermarking.