Secret Sharing-Based Reversible Data Hiding for Enhanced Audio Data Security across Multiple Genres

The rapid development of digital technology has facilitated data exchange and communication, while simultaneously increasing security threats such as data theft and manipulation. As personal data is highly confidential, effective protection mechanisms are required in the digital era. Audio steganography hides secret messages (payload) within audio signals; however, many existing approaches rely on a single stego-audio output, which can lead to information leakage during storage or transmission if the file is intercepted. This vulnerability allows an attacker to more easily reconstruct the steganographic scheme from a single output. To address this limitation, this study proposes a secret-sharing-based audio steganography method in which the payload is divided into multiple parts prior to embedding, thereby providing an additional protection mechanism. Nevertheless, the use of secret sharing may degrade stego-audio quality during embedding, as indicated by lower Peak Signal-to-Noise Ratio (PSNR) values. To mitigate this issue, a linear interpolation technique is incorporated to optimize the quality of the stego-audio. The proposed method focuses on three main aspects: improving protection by embedding messages in multiple parts, maintaining embedding capacity without introducing additional stego outputs, and improving audio quality through interpolation-based optimization. Experimental results show that the proposed approach improves stego-audio quality by approximately 6.95% in PSNR compared with several previous studies, while maintaining relatively high PSNR under the evaluated experimental conditions. Statistical evaluation using Normalized Correlation (NC) and entropy measurements indicates consistent payload reconstruction with limited statistical variation after embedding. Overall, the method contributes to an audio steganography scheme with a balanced trade-off between security, capacity, and audio quality.