Modify Speech Cryptosystem Based on Shuffling Overlapping Blocks Technique

A combination of high efficiency cryptosystem algorithms based on MOBS (Modified Overlapped Block Shuffling) and a HCS (Hybrid Chaotic System) are proposed to secure the transfer of an on-the-go audio signal (i.e. speech). As online voice-based communication is becoming increasingly more common, the need to protect these signals from an unauthorized third party becomes more significant. In natural speech, the values and positions for neighboring samples are highly correlated. The proposed algorithm breaks the correlation by using MOBS and encrypting with a HCS. First, the original speech signal is converted from a one dimensional into a two dimensional array. Then, it is divided into overlapped squared blocks followed by a subsequent permutation. Third, each permutated block is dynamically shuffled using Hénon and Arnold transformation. And finally a Hybrid Chaotic Map (Hénon and Arnold Cat map) is employed to generate a key matrix to encrypt the shuffled blocks, thus producing ciphered speech ready for transmission. Experimental results show that the cryptosystem can successfully encrypt/decrypt with symmetric keys. Experimental analyses like: Signal-to-Noise-Ratio (SNR), Segmental Signal-to-Noise-Ratio (segSNR), Correlation Coefficient Analysis (CCA), Log-Likelihood Ratio (LLR), Residual Deviation, key space and key-sensitivity reveal the effectiveness of the proposed technique for secure speech communication.