Multimedia Security in Audio Signals

The security of Digital media has been varying continuously due to advanced malware attacks. Multimedia security has become one of the major concerns. Earlier works primarily relied on encryption techniques to provide multimedia security, but those works have become obsolete and may be challenged by imposters as technology grows. The proposed work applied watermarking technique in the digital audio signal in which unique data is inserted in one-dimensional data so that it must not affect the significant information of the audio signal. The hybrid-decomposition scheme has been introduced over the audio data to extract features in terms of energy bands. The hiding of the watermark (image data) occurs in a low significant energy band that is stable and contains less information. The watermark extraction method is also introduced in which the hidden watermark data has been extracted back to its original form without leaving any distortion in audio data. The hybrid audio decomposition technique is a scheme in which dual decomposition of host audio signal has been applied containing fast Fourier transform (FFT) and Cordic QR scheme. The disintegration of the audio file will finally result in Q and R matrices under the Cordic QR scheme. Pseudo-random sequence method has been applied to find the correct location from the energy band of the R matrix to embed the watermark data. Before embedding, the watermark first undergoes dual encryption using cyclic coding and Arnold's cat map scheme. The reverse procedure has been applied for the watermark extraction and decryption process. The average PSNR value is 37.8 dB for watermarked audio.