High scrambling degree in audio through imitation of an unintelligible signal

Abstract

A reversible scheme of audio scrambling based on the imitation of Gaussian unintelligible signals is presented in this paper. It is supported by the similarities it shares with Gaussian unintelligible signals in terms of the Probability Density Function and the entropy. It is feasible for an audio signal to imitate the behavior of a Gaussian noise signal, and then the residual intelligibility is zero. Our proposed scheme, termed ASGI (Audio Scrambling by Gaussian signal Imitation), is tested with four different music genres, and the experimental tests reveal that the scrambled audio signals look like Gaussian noise signals and have high scrambling degrees. Additionally, our scheme preserves the advantages of imitation-based scrambling schemes.