A blind audio watermarking algorithm based on FFT coefficients quantization

Abstract

A blind audio watermarking algorithm based on Fast Fourier Transform(FFT) coefficients quantization is proposed. In order to resist MP3 compression, the area of the entire host audio signal for embedding watermark is designated by adaptive adjustment. Then the designated area is separated to segments and every segment is transformed into frequency domain by FFT. The DC coefficient of every segment which is insensitive to the human auditory system(HAS) and has the biggest hearing margin is selected. Finally, the watermark is embedded according to quantization algorithm. The experimental results show that the watermarked audio signal is not only imperceptible, but also robust against various common signal processing attacks, such as additive noise, low-pass filtering, resampling, echo and MP3 compression.