Secure Speech Encryption System Using Segments for Speech Synthesis

A number of analog speech encryption methods have been proposed. At present, since most speech encryption systems use digital encryption, the demand for analog speech encryption is diminishing. However, the risk of speech tapping by decryption attack has increased as a result of the increasing availability of immense computing power and worldwide computer networks. In this paper, we propose a new secure analog speech encryption method that uses speech segments in a large-scale speech corpus as encrypted "gtransmission segments," instead of transmitting the synthesized speech itself. As such, the proposed method is robust with respect to decryption attacks because of the complexity of solving a permutation problem. Some experiments were carried out to evaluate the proposed encryption system, and the results show that it can transmit encrypted speech with good intelligibility, at a signal-to-noise ratio of 3 dB.