Two-Dictionary Approach of Audio Watermarking with Image Embedded

An audio watermarking technique using spikegram is introduced. Here spikegram is a representation of an audio signal. Two techniques – FFT (Fast Fourier Transform) and Angular Spectrum are used to produce the spikegram. An efficient embedding technique known as two-dictionary approach is used which considers signs and phases of gammatones in the audio signal to increase the effectiveness. In addition to spikegram an image with key is embedded to further maximize the security. Proposed technique’s efficiency is improved and is shown via several simulation results executed in MATLAB.  Several parameters are measured such as MSE (Mean Square Error), PSNR (Peak Signal to Noise Ratio) for both the input audio signal and the Watermarked signal to mathematically study the variations in the signals. The embedded signal and image are retrieved from Watermarked signal so as to prove the Copyrights.