Research On Speech Information Hiding Method Based On Wavelet

Most information hiding methods always need repeatedly tests to determine the embedding amplitude. And these methods always use a constant embedding amplitude and the difference in different frames is not considered. The imperceptibility and robustness of these algorithms are very difficult to be considered at the same time. A novel information hiding method using adaptive selection of the embedding amplitude is proposed. This method uses the speech distortion measure to control the embedding amplitude of each frame in speech signals. It also uses the psychological maskers to modulate the secret information in each frame when the secret data are embedded. It has effectively controlled the transparency and robustness of the algorithm, so the imperceptibility and robustness of the algorithm can both have good effect. Experimental results show that the speech signal has not obviously quality decline after embedding secret information. The speech quality is good. This method can also resist kinds of possible attack, such as white noise, resampling, requantization and ADPCM compression.