A Generalized Spread Spectrum-Based Audio Watermarking Method with Adaptive Synchronization

چکیده مقاله

This paper introduces a novel spread spectrum-based audio watermarking method designed to enhance security, robustness, and embedding capacity while addressing the critical issue of synchronization. The proposed scheme utilizes Hamming coding, convolutional encoding, and a generalized pseudo-noise (PN) sequence to spread the watermark energy widely across the host audio spectrum, making it statistically and perceptually undetectable. For watermark recovery, the method employs Viterbi decoding alongside an adaptive filter for precise synchronization, enabling blind detection without requiring the original audio signal during extraction.

A key innovation of the approach is the strategic embedding of the watermark into specific high-energy regions of the host signal, identified using a psychoacoustic model based on the ISO/MPEG standard. By targeting areas with the highest auditory masking threshold, the method significantly increases the permissible watermark energy and embedding capacity without introducing audible distortions. This selective embedding, combined with adaptive synchronization, reduces computational complexity compared to full-spectrum spreading, enabling faster processing while maintaining high transparency.

The experimental results demonstrate the method's exceptional robustness against a wide range of common audio attacks. Under conditions such as additive white Gaussian noise, low-pass filtering (with a cutoff as low as 3.2 kHz), resampling, re-quantization, echo addition, and MP3 compression, the system achieved a bit error rate (BER) of nearly zero for many attacks, particularly when using a modulation center frequency between 1000 Hz and 2000 Hz. The use of the generalized PN sequence proved superior to conventional PN sequences, substantially lowering the BER across all tested scenarios.

In conclusion, the proposed audio watermarking system successfully balances imperceptibility, high capacity, and strong robustness. Its integration of adaptive synchronization, psychoacoustic masking, and advanced spread spectrum techniques presents a significant advancement over conventional methods. The scheme is particularly effective for securing digital audio in applications requiring copyright protection, authentication, and covert communication, offering a reliable and efficient solution for modern multimedia security challenges.