A novel DWT based Audio Watermarking adopting Fibonacci Numbers

Abstract

Audio watermarking is the process of embedding audio data in signal. Security is the main aim behind audio watermarking. Audio watermarking is the recent scope due to its reliability and high security depends on the data embedding technique as well as data extraction technique. In this paper we discussed about both embedding the data as well as extracting the data from watermark. The new in this technique is its type of embedding and extracting data in bit extract manner by changing the FFT spectrum magnitudes. The main theme in this implementation is FFT spectrum divided into small frames and we applied Fibonacci numbers to some selected number of FFT samples. By using Fibonacci series we can change the frequency samples. The advantages of using this system is it will provide the data which is having maximum change upto 61% and also the average error is less than 25%.This is very robust and transparent technique of watermarking. Experimental results shows that the proposed audio watermarking with Fibonacci series is having high capacity as well as it is having significant perceptual distortion and also shows its robustness against some audio signal processing attacks such as filtering, echo, added noise and MPEG compression (mp3). Finally we also proved the fidelity is also better for proposed work.