AWARE: Audio Watermarking with Adversarial Resistance to Edits

TL;DR

AWARE introduces an adversarial optimization-based audio watermarking method that enhances robustness against edits without relying on traditional attack simulations, achieving high audio quality and low error rates.

Contribution

It proposes a novel adversarial training approach for audio watermarking that improves robustness and decoding reliability over existing methods.

Findings

Achieves high audio quality and speech intelligibility.

Maintains low bit error rate across various audio edits.

Often outperforms state-of-the-art learning-based watermarking systems.

Abstract

Prevailing practice in learning-based audio watermarking is to pursue robustness by expanding the set of simulated distortions during training. However, such surrogates are narrow and prone to overfitting. This paper presents AWARE (Audio Watermarking with Adversarial Resistance to Edits), an alternative approach that avoids reliance on attack-simulation stacks and handcrafted differentiable distortions. Embedding is obtained through adversarial optimization in the time-frequency domain under a level-proportional perceptual budget. Detection employs a time-order-agnostic detector with a Bitwise Readout Head (BRH) that aggregates temporal evidence into one score per watermark bit, enabling reliable watermark decoding even under desynchronization and temporal cuts. Empirically, AWARE attains high audio quality and speech intelligibility (PESQ/STOI) and consistently low BER across various audio edits, often surpassing representative state-of-the-art learning-based systems.