A secure multilayer audio encryption scheme based on chain ring algebraic S-boxes, Rössler diffusion, and 4D hyperchaotic permutation

Abstract The secure transmission of digital audio requires encryption techniques that can protect audio information against different types of cryptanalytic attacks. Existing chaos-based audio encryption methods commonly use conventional $$\:GF\left({2}^{8}\right)$$ -based substitution boxes (S-boxes) and a single low-dimensional chaotic system for both diffusion and permutation, which restricts the achievable nonlinear confusion and the overall unpredictability of the encryption process. In this work, two $$\:8\times\:8$$ S-boxes are constructed using the finite commutative chain ring $$\:{R}_{13}={F}_{2}\left[x\right]/\langle {x}^{13}\rangle $$ , an $$\:8192$$ -element structure not previously used for dual S-box design. Bijective mappings from the ring structures to $$\:GF\left({2}^{8}\right)$$ , followed by projective linear fractional transformations, are used to obtain the proposed S-boxes. The S-boxes are incorporated into a three-stage audio encryption scheme consisting of Rössler-based XOR diffusion, 4D Lü hyperchaotic permutation, and dual S-box substitution, with diffusion and permutation drawn from two independent chaotic systems instead of one. The proposed method is evaluated using different audio signals through nonlinearity, entropy, correlation, NPCR, UACI, key sensitivity, NIST, and computational analyses. The constructed S-boxes reach a nonlinearity of $$\:112$$ with differential approximation probability as low as $$\:0.0148$$ . The encrypted audio achieves entropy values up to $$\:15.8999$$ bits, correlation values as low as $$\:-0.0017$$ , NPCR values up to $$\:99.6783\%$$ , UACI values close to $$\:33.4\%$$ , and full passage of the NIST SP 800 − 22 suite. Encryption of a $$\:28$$ -second audio file requires only $$\:\:0.2855\:$$ s. The results indicate that combining chain-ring-based S-boxes with chaotic diffusion and hyperchaotic permutation provides strong confusion and diffusion while maintaining practical computational efficiency for 16-bit audio encryption.