Hybridization of speech information hiding and encryption for double-layer security in speech communication

Speech transmitted over communication systems must be highly secured since they are vulnerable to attacks. In this study, we use the singular spectrum analysis (SSA)-based information hiding method with the transformation method to boost security for speech communication. Arnold transformation is performed on watermark signals to provide secured watermarks. The secured watermarks are then embedded into the host signal using SSA-based information hiding to obtain a watermarked signal. The watermarked signal is encrypted before being sent through a communication channel. The experimental results show that effective encryption and information hiding are feasible. The key sensitivity and the discrepancy between the watermarked speech signal and its encrypted version measured using the correlation coefficient and the signal-to-noise ratio suggest that only the authorized person who has the encryption key will be able to access the speech's contents. A secured watermark signal's imperceptibility and a high bit error rate without a watermark key indicate hidden information is limited in access. This hybridized system provides speech security and selectively grants access to data at varying levels.