Digital audio watermarking method capable of resisting re-recording attack

Abstract

The invention discloses a digital audio watermarking method capable of resisting re-recording attack. The method comprises the following steps of: 1, watermark synchronization reference point extraction: dividing a carrier audio signal into audio frames with equal length, extracting the 12-dimension semitone characteristics of each audio frame, and thus obtaining watermark synchronization reference points; 2, watermark embedding: coding watermark information to obtain a binary watermark bit coding sequence, dividing a carrier audio into a plurality of watermark synchronization units, dividing each watermark synchronization unit into a plurality of audio sample sections, and embedding a watermark into a frequency spectrum of a low-frequency domain of each audio sample section through binary coding; and 3, watermark extraction: during the watermark extraction, dividing each watermark synchronization unit into the corresponding audio sample sections according to synchronization reference points extracted by parameters for the watermark embedding, respectively calculating a mean value of the frequency spectrum in a low-frequency spectrum interval of each audio sample section, and extracting the audio section interval watermark bit of each mean value. According to the method, the embedding position of the watermark can be accurately determined, and the re-recording attack can be resisted. The method is high in robustness.

Description

technical field [0001] The invention relates to a digital audio watermarking method, in particular to a digital audio watermarking method which can resist re-recording attacks. Background technique [0002] Digital audio can be copied infinitely without distortion, and will be tampered with or replaced during transmission, so the invisible digital audio watermarking technology came into being. [0003] A successful invisible watermarking method must meet the two basic technical requirements of imperceptibility and robustness. For audio watermarking, imperceptibility is auditory transparency, and embedding watermarks in audio signals should not cause distortion that can be perceived by the human ear. Robustness means that after the audio signal is attacked and processed without destroying the sound quality, the digital watermark can still be reliably extracted and correctly identified. [0004] Digital audio watermarking methods are generally divided into two categories: ti...

AI Technical Summary

Problems solved by technology

This method can resist attacks such as color noise, MPEG encoding, multiple watermarking and resampling, but it is still not robust to "play-re-recording" attacks.

[0005] To sum up, the audio watermarking techniques mentioned above only consider watermark extraction from digital audio files, and they are robust to common attacks, but cannot resist "Play-re-record" attack

This is because when the watermarked digital audio is played by the hardware device and then recorded and digitized again, its waveform and spectrum will undergo major changes, and its attack strength is much greater than that of conventional signal processing, thus destroying the watermark information

Moreover, the time base and sampling rate in the re-digitization cannot be consistent with the original audio, so the synchronization of the watermark information cannot be realized by the usual method, and the watermark cannot be extracted from the re-recorded copy.

Images