Answers are grounded in this paper and cite their sources.
Chaotic speech encryption provides a flexible approach for protecting speech data, but its practical performance depends strongly on the dynamical characteristics of the underlying chaotic system and on the association between the generated keystream and the input signal. This study proposes a three-dimensional quadratic sinusoidal robust chaotic map (3D-QSRC) and applies it to a dual-channel speech encryption framework with statistical-feature-dependent initial states. The proposed map combines quadratic nonlinear feedback, trigonometric cross-variable coupling, and modulo-based state bounding. Its dynamical behavior is evaluated using Lyapunov exponents, bifurcation analysis, the 0–1 test, correlation dimension, Kolmogorov entropy, sensitivity analysis, and NIST SP 800-22 statistical tests. In the speech-encryption framework, the mean, extrema, and sample length of each channel are mapped to channel-specific initial states. The left and right branches further employ Logistic and Cubic auxiliary mappings, respectively, followed by sorting-based scrambling and sample-wise modulo diffusion. Experimental results on dual-channel speech samples show that the average plaintext–ciphertext correlation coefficients are -0.001543 and 0.001287 for the left and right channels, respectively, while the encrypted-speech entropy approaches the theoretical maximum under the adopted quantization setting. The results indicate that the proposed framework effectively reduces the directly observable statistical relationship between plaintext and ciphertext while maintaining a computationally simple dual-channel processing structure.
No takes yet. Share an insight, caveat, or question.
Yang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: