Abstract: The voluminous amounts of data that are transmitted through wireless network are majorly constituted by speech signal. In the prevailing environment where in speech data that are critical by its nature in terms of privacy must be protected. This paper would focus on suggesting a novel algorithm based on audio steganography to secure voice and speech related data during transmission in mobile devices. Audio steganography is a method of hiding the secret data in the cover medium by which only the sender and the intended receiver are only able to realize the existence of a secret data. The process of steganography is carried out in cepstral domain and the key is constructed using the Mel-frequency cepstral coefficients. The imperceptibility in hearing is exploited in a way where the data are embedded in low power levels to make the detection more complicated. The generic auto selection of the key members helps in making the critical data's existence highly non vulnerable in the cover medium. Thus the algorithm stands at high tolerance on interception.
Keywords: Mobile Communication, Audio Steganography, Cepstral Analysis, Mel-Frequency Cepstral Coefficients