An Audio Steganography Scheme Robust to White Gaussian Noise Attack

Abstract

An innovative audio Steganography scheme based on optimize algorithm as Probabilistic Global Search Lausanne (PGSL) in modified discrete cosine transform domain (MDCT) is proposed. Inaudibility and robustness are two main important requirements for audio Steganography. The major contribution of the proposed system is to embed the secret data in the best MDCT coefficients against secret message detection from unauthorized users. Embed the secret data by employing original signal transformation with MDCT and best point searching process through PGSL Algorithm to achieve high inaudible and robustness. Results of the experimental test using clean utterances from Texas Instruments Massachusetts Institute of Technology (TIMIT) databases and music files show error-free data recovery from noise addition attack stego audio. Robustness of the system is evaluated with BER (bit error rate). Spectrogram and Informal listening test also showed that the distortions in the stego audio from the original cover speech and music files is imperceptible and inaudible even with high payload.