Real-time, embedded speech-in-speech hiding has not been widely researched. Nevertheless, it could be useful, among other cases, in secure mobile telephony. In this paper, we propose a new scheme of data hiding which takes advantage of the masking property of the Human Auditory System (HAS) to hide a secret (speech) signal into a host (speech) signal. The embedding process is carried out into the wavelet coefficients of the speech signals. The main point of the proposed scheme is that the embedding process is suitable for real-time processing, and the secret’s coefficients are relocated by an adaptive key, instead of a pseudo-noise sequence of some approaches. The latency of the embedding module makes this approach useful for real-time speech communication because the total delay added by the proposed system is low compared to the highest delay allowed for a high quality speech transmission.

Peer Reviewed