Perceptual audio watermarking driven by human auditory system