A Robust Audio Watermarking Technique for Uncompressed Audio Samples Using SWT-SVD Enhanced with Arnold Transform and Fuzzy Logic

The widespread availability of digital multimedia creates substantial weaknesses in intellectual property security, including concerns about piracy and unlawful usage. Traditional audio watermarking approaches sometimes suffer from decreased imperceptibility and robustness, particularly when subjected to complicated attacks such as cropping and pitch-invariant transformation. This paper proposes a robust audio watermarking scheme that integrates stationary wavelet transforms (SWT), singular value decomposition (SVD), Arnold Transform, and Fuzzy Logic. The approach employs Sugeno Type-2 Fuzzy Systems for adaptive watermark embedding and Type-1 for accurate extraction. Results demonstrate that this approach enhances robustness against conventional and advanced signal-processing attacks while preserving excellent audio quality. Metrics such as Signal-to-Noise Ratio (SNR), Objective Difference Grade (ODG), Subjective Difference Grade (SDG), Mean Opinion Score (MOS), and Bit Error Rate (BER) validate its effectiveness. This research advances multimedia security, offering a secure, non-intrusive, and efficient solution for intellectual property protection in audio distribution.