Abstract

Inserting image in audio file is the purpose of this paper. Music listener is not able to detect the difference between original audio file and watermarked one. Songwriter can protect his artwork against illegal copy and theft of works. Discrete wavelet transform, singular value decomposition are the mathematic approach used. They are popular on these domains but the way to operate is different for each algorithm. Frequency transformation is needed to ensure that the proposed algorithm is robust/fragile across different attack then we operate only on non-sensitive data. This criterion is not enough, so we add a module formatting the matrix to another matrix before applying singular value decomposition. Watermark image must be a binary and its extraction doesn’t require original information. It means that the watermarking algorithm that we describe here is a robust/fragile and blind audio watermarking.