SilentCipher: Deep Audio Watermarking

TL;DR

SilentCipher is a novel deep learning-based audio watermarking model that combines psychoacoustic thresholding and pseudo-differentiable compression to achieve imperceptible, robust watermarks at high audio quality.

Contribution

It introduces the first psychoacoustic thresholding in deep audio watermarking, pseudo-differentiable compression layers, and eliminates perceptual losses for state-of-the-art performance.

Findings

Achieves imperceptible watermarks with high robustness.

Outperforms previous methods in audio watermarking benchmarks.

Enables encoding at standard audio sampling rates.

Abstract

In the realm of audio watermarking, it is challenging to simultaneously encode imperceptible messages while enhancing the message capacity and robustness. Although recent advancements in deep learning-based methods bolster the message capacity and robustness over traditional methods, the encoded messages introduce audible artefacts that restricts their usage in professional settings. In this study, we introduce three key innovations. Firstly, our work is the first deep learning-based model to integrate psychoacoustic model based thresholding to achieve imperceptible watermarks. Secondly, we introduce psuedo-differentiable compression layers, enhancing the robustness of our watermarking algorithm. Lastly, we introduce a method to eliminate the need for perceptual losses, enabling us to achieve SOTA in both robustness as well as imperceptible watermarking. Our contributions lead us to SilentCipher, a model enabling users to encode messages within audio signals sampled at 44.1kHz.