Voice quality in speech watermarking using spread spectrum technique

One of the important problems in communication is sending specific data privately. In addition to cryptography which was used many years ago, methods such as data hiding have been introduced in recent years. In data hiding, specific data is embedded in another signal, which is known as a host signal that can be in the form of video, image and audio. Because of narrow bandwidth limitation, speech signal is seldom used, despite its popularity in communication application such as military, bank phone and network security. A data hiding algorithm in speech signal is called speech watermarking. Watermarking will be more robust against attacks (i.e. watermark removal and impairment attacks) with uses the Spread Spectrum technique. Speech watermarking is implemented using simulation and then been evaluated based on the quality of voice (speech) by the Mean Opinion Score (MOS) from 20 respondents. The average MOS is 2.75 out of 5 showings between fair and poor perceptual audibility quality.