Abstract

Digital audio scrambling is a process used in audio security applications. Scrambling of audio files breaks the correlation between adjacent samples in order to convert the original audio to an unintelligible format. Scrambling is used to protect the audio against wiretapping and illegal surveillance, in addition to being a step in security algorithms, such as watermarking and encryption algorithms. Cellular automata are models that are discrete in nature and depend on simple and local rules to achieve an interesting overall behavior. Two-dimensional cellular automata were previously proposed as a key generation mechanism to scramble audio files. The mechanism was built upon be researchers in the multimedia security field. This paper explores the use of one-dimensional cellular automata in audio scrambling, which simplifies the process as deploying two-dimensional cellular automata requires changing the dimension of the audio file and the one-dimensional cellular automata does not, additionally, elementary one-dimensional cellular automata requires less parameters to configure. The scrambling degree is used to evaluate the model effectiveness in breaking the correlation of adjacent samples. In the experiments, different parameters are taken into account including the cellular automata class, the iterations needed and the method used to calculate the cells at the boundary. Experiments show that the one-dimensional cellular automata are capable of scrambling the audio file without any dimensional change and the chaotic rules tested give the highest scrambling degree.

This work is licensed under a Creative Commons Attribution 4.0 License.