NeuMark-Native: Robust Text-to-Speech-Native Watermarking Through Full Utilization of Neural Audio Codec Latent Space

2026-10-04

AI Summary

This study addresses the vulnerability and insufficient robustness of conventional post-hoc speech watermarking by proposing a TTS-native watermarking framework. The method pioneers the utilization of the complete latent space of neural audio codecs for multi-layer information embedding, completing the marking process prior to waveform decoding. Furthermore, the pre-trained model remains frozen while only the watermarking modules are jointly optimized. Experimental results demonstrate that the proposed framework maintains high detection rates under various digital signal processing and re-synthesis attacks, achieving full-pipeline traceability. Meanwhile, the naturalness and quality of the synthesized speech remain comparable to the original level.

Abstract

Speech watermarking offers proactive traceability for synthetic speech, yet most existing models operate only after text-to-speech (TTS) synthesis by adding a watermark perturbation to the generated waveform. This post-hoc design leaves watermarking as an external step that can be omitted or bypassed and restricts the watermark to a shallow waveform representation. We propose NeuMark-Native, a TTS-native watermarking framework for neural codec-based synthesis. It embeds payload information into every generated codec-latent layer before waveform decoding, improving watermark persistence under downstream digital signal processing (DSP) and neural codec resynthesis. NeuMark-Native keeps the pretrained TTS model and the neural codec frozen, while optimizing only the watermark modules on generated codec tokens. Experiments on two corpora under 11 DSP attacks and 9 neural-codec attacks demonstrate robust watermark detection while preserving naturalness, intelligibility, and speech quality close to synthetic speech.

Problem

• Speech watermarking
• Text-to-Speech
• Neural audio codec
• Robustness
• Latent space

Innovation

• Text-to-Speech Watermarking
• Neural Audio Codec
• Latent Space
• TTS-native
• Robustness