Spread spectrum audio information hiding algorithm based on auditory perception

To solve the problem that the auditory masking threshold curve calculated by common method has a large deviation, a new calculation method with comprehensive consideration of simultaneous and backward masking and independent calculation of masking threshold in the frequency domain point by point is proposed. The method can take full advantage of auditory masking characteristics and improve the capacity and robustness without losing good perceptual transparency. Furthermore, by adopting some measures including perceptual reshaping, in-phase modification and multi-frame superposition, a spread-spectrum-based audio information hiding algorithm resistant to desynchronization is presented. Simulation results show that the bit error rate of extracted messages is less than 1.6 × 10⁻², that is, the algorithm is robust against common signal interferences such as desynchronization, additive noise, re-sampling and filtering.