Publication · October 11, 2026
Asymmetric Phase Coding Audio Watermarking Accepted at NeurIPS 2026
A research project sponsored and led by Phi Lab Foundation, Asymmetric Phase Coding Audio Watermarking, has been accepted at the 40th Conference on Neural Information Processing Systems (NeurIPS 2026).
The paper introduces Asymmetric Phase Coding (APC), a training-free cryptographic signing layer for audio. An Ed25519 signature is hidden in the phase and magnitude of the signal, and anyone holding the public key can verify it, with no model and no shared secret.
Key results
- Cryptographic verification rates of 97.5% to 98.3% across eight attack conditions on 1,000 LibriSpeech clips, including cropping, low-pass filtering, resampling, and MP3 and OGG compression.
- Mean PESQ of 3.02 and tens of milliseconds of CPU latency.
- Interoperable open-source implementations in JavaScript, Rust and Python, released under the BSD 2-Clause License.
Authors
Guang Yang, Fengchen Liu, Amir Ghasemian, Zhong Wang, Ninareh Mehrabi, Homa Hosseinmardi.
Links
- Research page: Asymmetric Phase Coding Audio Watermarking
- Interactive project page: gy19a.github.io
- Paper: NeurIPS 2026 poster