Open-Set Source Tracing of Audio Deepfake Systems

2025/07/09 by Nicholas Klein, Klein, Nicholas, Hemlata Tak +3 · 3 citations
Computer Science · #Audio and Speech Processing (eess.AS) #Digital Media Forensic Detection #FOS: Computer and information sciences #FOS: Electrical engineering #Music and Audio Processing #Sound (cs.SD) #Speech and Audio Processing #electronic engineering #information engineering

paper · pdf · doi:10.48550/arxiv.2507.06470

openalex publication_date 2025/07/09 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Existing research on source tracing of audio deepfake systems has focused primarily on the closed-set scenario, while studies that evaluate open-set performance are limited to a small number of unseen systems. Due to the large number of emerging audio deepfake systems, robust open-set source tracing is critical. We leverage the protocol of the Interspeech 2025 special session on source tracing to evaluate methods for improving open-set source tracing performance. We introduce a novel adaptation to the energy score for out-of-distribution (OOD) detection, softmax energy (SME). We find that replacing the typical temperature-scaled energy score with SME provides a relative average improvement of 31% in the standard FPR95 (false positive rate at true positive rate of 95%) measure. We further explore SME-guided training as well as copy synthesis, codec, and reverberation augmentations, yielding an FPR95 of 8.3%.

Citations

• ASVspoof 5: Design, Collection and Validation of Resources for Spoofing, Deepfake, and Adversarial Attack Detection Using Crowdsourced Speech
• Towards Explainable Spoofed Speech Attribution and Detection:a Probabilistic Approach for Characterizing Speech Synthesizer Components
• Neural Codec Source Tracing: Toward Comprehensive Attribution in Open-Set Condition
• Investigating Prosodic Signatures via Speech Pre-Trained Models for Audio Deepfake Source Attribution
• ASVspoof 5: Crowdsourced Speech Data, Deepfakes, and Adversarial Attacks at Scale
• Generalized Source Tracing: Detecting Novel Audio Deepfake Algorithm with Real Emphasis and Fake Dispersion Strategy
• MLAAD: The Multi-Language Audio Anti-Spoofing Dataset
• Distinguishing Neural Speech Synthesis Models Through Fingerprints in Speech Waveforms
• ADD 2023: the Second Audio Deepfake Detection Challenge
• Source Tracing: Detecting Voice Spoofing
• Spoofed training data for speech spoofing countermeasure can be efficiently created using neural vocoders
• HiFi-GAN: Generative Adversarial Networks for Efficient and High Fidelity Speech Synthesis
• Energy-based Out-of-distribution Detection
• Neural source-filter waveform models for statistical parametric speech synthesis
• MCE 2018: The 1st Multi-target Speaker Detection and Identification\n Challenge Evaluation
• WaveGlow: A Flow-based Generative Network for Speech Synthesis
• Deep Anomaly Detection with Outlier Exposure
• A Simple Unified Framework for Detecting Out-of-Distribution Samples and Adversarial Attacks
• Open Category Detection with PAC Guarantees
• CosFace: Large Margin Cosine Loss for Deep Face Recognition
• Enhancing The Reliability of Out-of-distribution Image Detection in Neural Networks
• Deep Residual Learning for Image Recognition

Cited by

• FakeMark: Deepfake Speech Attribution With Watermarked Artifacts
• Forensic Similarity for Speech Deepfakes
• Advancing Zero-Shot Open-Set Speech Deepfake Source Tracing

Related

• Neural Codec Source Tracing: Toward Comprehensive Attribution in Open-Set Condition
• Generalized Source Tracing: Detecting Novel Audio Deepfake Algorithm with Real Emphasis and Fake Dispersion Strategy
• Unveiling Audio Deepfake Origins: A Deep Metric learning And Conformer Network Approach With Ensemble Fusion
• TADA: Training-free Attribution and Out-of-Domain Detection of Audio Deepfakes
• Audio Deepfake Verification