Secure speech coding using chaotic shift keying for encryption combined with error recovery
Messaouda Boumaraf,Fatiha Merazka
DOI: https://doi.org/10.1007/s11042-024-19560-7
IF: 2.577
2024-06-16
Multimedia Tools and Applications
Abstract:This paper presents a speech encryption algorithm utilizing chaotic shift keying alongside error recovery techniques for the G.722.2 codec. The algorithm employs chaotic data interleaving on speech inter-frames to replace continuous packet loss with isolated losses, thus restoring missing data packets. Codec parameters undergo encryption through the combination of multiple chaotic maps applied to intra sub-frames of speech, facilitating confusion and diffusion processes. Enhanced security is achieved through the utilization of multiple chaotic maps, including chaotic shift keying generators applied to codec sub-frame parameters. Specifically, the chaotic shift keying mechanism is employed using four chaotic maps: a Henon map for and , a logistic map for , a tent map for , and a cubic map for . The keys derived from these maps are utilized for chaotic data interleaving on both inter-frames and intra sub-frames of the speech file to shuffle frame parameters (confusion) and scramble sub-frame parameters (diffusion). Additionally, the proposed system enhances speech quality by mitigating degradation caused by packet loss through a waveform substitution-based recovery method, which replaces lost packets with previously received copies. A series of experiments, including waveform representation, spectrogram analysis, Enhanced Modified Bark Spectral Distortion (EMBSD) test, WideBand Perceptual Evaluation of Speech Quality (WB-PESQ) test, MUltiple Stimuli with Hidden Reference and Anchor (MUSHRA) score, Mean Opinion Score (MOS) score, correlation test, entropy test, Number of Samples Change Rate (NSCR) test, Unified Average Modified Intensity (UACI) test, key space analysis, and sensitivity test, are conducted to evaluate the security and speech quality of the proposed system. The results validate the system's security and confirm its ability to maintain good audio quality even under packet loss rates ranging from 5% to 20%.
computer science, information systems, theory & methods,engineering, electrical & electronic, software engineering