A novel approach for audio steganography by processing of amplitudes and signs of secret audio separately

Abstract

Now a days, cases of theft of important data both by employees of the organization and outside hackers are increasing day-by-day. So, new methods for information hiding and secret communication are need of today. Steganography is an option for it. Embedding a secret message into other meaningful message (cover media) without disturbing the features of the cover media is known as steganography. A novel approach for audio steganography is proposed in this paper. Here, secret message and cover media both are digital audio. Proposed approach is robust with respect to both LSB removal and re-sampling attacks. This approach adds extra layer of security because a transformation function is applied on amplitude bits of secret audio before embedding. This approach is more resistive towards white Gaussian noise addition (WGN) during transmission of stego file. The proposed approach is also suitable for embedding secret audio during real time audio communication because processing time is low while embedding capacity is high. Embedding capacity of the proposed approach is same as of conventional LSB approach because in both approaches one bit of secret is being inserted in each sample of cover audio. Standard parameters: Perceptual Evaluation of Speech Quality (PESQ) and Mean Opinion Score (MOS) are used for measuring the imperceptibility between cover audio & stego audio. For the proposed approach, PESQ and MOS are found as 4.47 and 5 that are very close to their respective highest values 4.5 and 5 when there is no attack.