A lightweight fragile audio watermarking method using nested hashes for self-authentication and tamper-proof

Watermarking methods have traditionally emerged from the imperative to guarantee the authenticity of digital documents. As we navigate the current landscape of advancing voice imitation and forgery techniques, safeguarding forensic evidence signals, historic speeches, and music signals has become an urgent challenge. A notable concern with existing methods is the presence of blind spots—areas where tampering goes undetected—significantly impacting detection sensitivity and rates. To address these challenges, this research introduces a fragile audio watermarking method designed to identify even the slightest modifications in audio signals. The approach relies on embedded nested SHA-256 hashes. Performance evaluation demonstrates that any inversion of a bit in the watermarked signals is promptly detected on the blind side of detection. Notably, blind spots are minimized, leading to a remarkable 100% detection rate. Furthermore, comparative results highlight the lightweight nature of the watermark, ensuring high signal imperceptibility. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2024.