Exposing and Mitigating Neural Codec Vulnerabilities in Audio Deepfake Detection

Authors: Abdullah, Awais Khan, Khalid Mahmood Malik

Published: 2026-10-05 18:27:37+00:00

AI Summary

This research exposes the vulnerability of current audio deepfake detection (ADD) systems to neural audio codec compression, noting that models trained on codec resynthesized data are particularly susceptible and often misclassify legitimate speech as fake. The authors introduce the ANC-Spoof dataset, derived from existing ADD benchmarks with seven neural codecs, and propose PCL-NET. PCL-NET, a fine-tuned XLS-R encoder with a pairwise consistency objective, significantly reduces detection error rates under neural codec compression while maintaining deepfake detection performance.

Abstract

Existing audio deepfake detection (ADD) datasets and detectors are primarily built for vocoder-based synthesis, evaluated against traditional post-hoc perturbations such as MP3/AAC compression or additive noise, applied independently of generation. However, recent speech synthesizers, particularly ALM-based systems, use neural audio codecs both for compression and as the resynthesis reconstructing waveforms from generated tokens, producing artifacts distinct from post-hoc compression. Neural codecs thus play a dual role: some are designed for pure compression under low-bandwidth communication, while others serve as resynthesis components. Despite this dual role, robustness to codec-based compression, unlike post-hoc compression, remains largely unexplored. We expose this gap, showing that state-of-the-art (SOTA) ADD models degrade drastically on codec-compressed speech; in particular, systems trained on Codec Resynthesized data as a proxy for codec-based generation prove most vulnerable, with legitimately compressed bona fide speech often misclassified as fake. To investigate this, we construct the Audio Neural Codec-Spoof dataset by applying seven neural codec algorithms to existing ADD benchmarks, isolating codec-induced resynthesis artifacts as a controlled proxy for codec-based generation. As baseline mitigation, we propose PCL-NET (Pairwise Consistency Learned Network), fine-tuning a pretrained XLS-R (300M) encoder with a pairwise consistency objective that minimizes the representation distance between an utterance's uncompressed and codec-compressed versions, disentangling codec artifacts from the real-versus-fake decision. As a result, PCL-NET reduces average EER under neural codec compression from 28.67% to 12.77%, while preserving competitive CoSG-based deepfake detection performance. We will also make the dataset publicly available on Hugging Face upon acceptance.