The Flow has not summarised this story yet — read it at Hugging Face Trending Papers.
Partial deepfake speech, where only limited segments of an utterance are synthesized or manipulated, poses a significant challenge to existing deepfake detection systems. As the proportion of spoofed...
The paper introduces a training‑free proactive defense for detecting partial deepfake speech by using self‑embedding steganography. It embeds a compressed version of the clean audio within itself, allowing post‑hoc extraction of reference content and enabling detection of spoofed segments via codec‑based restoration. Experiments on a benchmark dataset show that this method complements passive detectors and operates without any training, offering a robust, data‑efficient alternative for partial deepfake detection.
CRAW is a codec‑robust audio watermarking framework designed to embed imperceptible signals into synthetic speech, enabling provenance verification. It improves robustness against neural re‑synthesis, codecs, denoisers, and vocoders while preserving perceptual quality through distortion‑aware training, attention‑based pooling, perceptual masking, and error‑correcting codes. Experiments show CRAW outperforms existing post‑hoc watermarking methods in robustness without compromising audio quality.
arXiv:2603. 05310v3 Announce Type: replace-cross Abstract: While existing audio watermarking techniques have achieved strong robustness against traditional digital signal processing (DSP) attacks, they remain vulnerable to neural compression.
The paper introduces CRAW, a codec‑robust audio watermarking framework designed to embed imperceptible signals into synthetic speech. CRAW enhances robustness against neural re‑synthesis, codecs, denoisers, and vocoders while preserving high perceptual quality through distortion‑aware training, attention‑based pooling, perceptual masking, and error‑correcting codes. Experiments show CRAW outperforms existing post‑hoc watermarking methods in robustness without compromising audio quality.
The paper formalizes backward compatibility for zero‑bit neural audio watermarking and introduces a training framework that pairs a new detector with a sequence of frozen legacy embedders. Using a performance‑weighted sampling strategy, the detector maintains accuracy across all earlier versions, but this coupling can cause an embedder‑collapse effect that degrades the true negative rate against out‑of‑chain embedders. To mitigate this, an Angular Diversity Loss is added, ensuring each new embedder occupies a distinct output space region and restoring near‑perfect true negative rates without harming detection accuracy or audio quality.