Reversible and Robust Audio Watermarking Based on Spread Spectrum and Amplitude Expansion

Abstract

Recently, a technique that uses quantization index modulation and amplitude expansion has been proposed for reversible and robust audio watermarking. However, when applied to modified stego audio, the technique is not reversible. Here, a novel technique that is both reversible and robust is proposed for hiding data in audio. It is perfectly reversible for unmodified stego signals and is semi-reversible for perceptually coded stego signals. A robust payload is embedded by direct-sequence spread-spectrum modulation, with the sequence determined from the amplitude expansion in time and frequency of integer modified discrete cosine transform (MDCT) coefficients. Simultaneously, a reversible payload is embedded into the apertures in the amplitude histogram that result from amplitude expansion of the integer MDCT coefficients. The robustness and size of the reversible payload were evaluated by simulation for 20 music pieces. The reversible-payload capacity was approximately 4.8 kilobits per second. Encoding the stego signals into MP3, tandem MP3, and MPEG4AAC, and applying an additive single delay to the stego signals, revealed a maximum bit error rate of less than 6.5 % with a robust payload of 7.2-bits per second. Measurement of the objective quality of the stego audio and audio recovered from the modified stego signal was done by using software based on the perceptual evaluation of audio quality algorithm. The results showed that the mean objective difference grade (ODG) was better than ‘perceptible, but not annoying’ for the stego audio. Semi-recovery from the perceptually coded stego signals was realized in terms of small differences in ODG between the recovered and coded signals.