Secure Loss less Speech Signal Encryption using SNMS, Logistic Chaotic map and Random Sequences

Abstract

This paper suggests a lossless way to encrypt and decrypt speech using an algorithm based on a chaotic map. The primary goal is to preserve audio signal confidentiality while preserving bit-level accuracy during decryption. To maintain full audio resolution, the original audio streams are first transformed to a 16-bit integer representation. A Suresh Nakum and Mehul Shah's (SNMS) map and logistic map is used to construct a random chaotic sequence. This sequence is used for two layers of encryption, permutation of audio samples and XOR-based diffusion with 16-bit keys formed from chaotic values. The beginning parameters, which operate as secret keys, are used to construct the exact same chaotic sequence during decryption, and the encryption operations are reversed in the exact order they were performed. The suggested method achieves flawless reconstruction with no quality loss by guaranteeing that the original and decrypted audio signals are bit-exact. The efficacy and resilience of the approach are demonstrated by experimental validation.  The encrypted speech signal does not reveal any meaningful information when analyzed through its histogram, frequency spectrum, and spectrogram also the largest difference between the original and decrypted samples is zero. For applications like voice communications, military audio or biometric systems, this method is promising.