Abstract
In this paper we present a new approach to hide speech for secrecy applications. The adopted technique exploits the low-pass spectral properties of the speech magnitude spectrum to hide a speech signal in the low-amplitude high-frequency region of the cover speech signal. Simulation results on both English and Arabic speech reveal that the proposed approach is capable of hiding the secret message inside the cover speech to produce a stego speech that is indistinguishable from the original cover speech, while being able to recover the secret speech message without any noticeable degradations in quality.