| name | snap-speaker-nulling-deepfake-detection |
| title | SNAP: Speaker Nulling for Artifact Projection in Speech Deepfake Detection |
| version | 0.0.3 |
| engine | skillxiv-v0.0.3-claude-opus-4.6 |
| license | MIT |
| url | https://arxiv.org/abs/2603.20686 |
| keywords | ["Deepfake Detection","Speaker Embeddings","Subspace Projection","Audio Forensics","Domain Generalization"] |
| description | Detect speech deepfakes by decomposing features into speaker, artifact, and context subspaces—nulling speaker information via PCA projection to isolate artifact patterns. Train minimal logistic regression classifier on orthogonally-projected representations, achieving 56% error reduction while maintaining cross-speaker and cross-TTS generalization. |
SNAP: Speaker Nulling for Deepfake Detection
Core Insight: Speaker Entanglement Problem
Audio deepfake detectors trained on self-supervised learning (SSL) representations exploit speaker-identity patterns instead of learning genuine synthesis artifacts. When models like WavLM encode speaker information strongly, detectors learn speaker-specific correlations rather than the acoustic cues that distinguish real from synthetic speech. This causes catastrophic failure when encountering new speakers or unseen TTS models.
Why Non-Obvious: Conventional wisdom assumes SSL representations separate "what is being said" from "who is speaking." In practice, speaker-identity encoding is so dominant that artifact-detection layers become speaker classifiers instead.
Minimal Recipe: Subspace Projection
Decompose feature space as ℋ = 𝒮 + 𝒜 + 𝒞 where:
- 𝒮 = speaker-dependent subspace
- 𝒜 = artifact-related subspace
- 𝒞 = context (utterance, channel) subspace
Three-Stage Pipeline:
- Feature Extraction: Concatenate WavLM-Large layers 8 and 22, apply mean pooling, L2-normalize
- Speaker Subspace Identification: Compute PCA on speaker centroids to find the speaker-dependent subspace
- Nulling via Orthogonal Projection: Project features orthogonally to suppress speaker subspace while preserving artifact and context information
- Classification: Train logistic regression (only 2,049 parameters) on nullified features
Results & Cross-Domain Robustness
ASVspoof 2019 LA (benchmark): 0.35% EER—56% improvement over WavLM-ECAPA baseline
In-the-Wild Dataset: 15.39% EER vs. 22.22% for WavLM baseline
Generalization to Unseen TTS:
- CosyVoice2 (not in training set): Maintains strong performance
- F5-TTS (new diffusion-based model): Effective detection despite architectural novelty
Key Finding: Nulling speaker subspace removes speaker-specific decision boundaries, forcing the classifier to rely on genuine synthesis artifacts that generalize across speakers and TTS models.
Deployment Considerations
- Enrollment Phase: Collect 3-5 clean utterances from each reference speaker to establish speaker subspace
- PCA Dimensionality: Determine via scree plot or cross-validation; typically 20-50 dimensions
- Scaling: Method is lightweight—PCA + logistic regression runs on CPU in real-time
- Generalization: Works best when training set includes diverse speakers; speaker diversity in training improves nulling robustness
- Failure Mode: Performance degrades if speaker subspace overlaps with artifact subspace (rare but possible for certain TTS models); mitigate via larger enrollment sets
Practical Impact
- Simplicity: Minimal parameters (logistic regression) reduce overfitting to seen speakers/TTS models
- Interpretability: Orthogonal projection provides clear subspace decomposition; decisions are traceable
- Deployment: Fast inference, no need for large models; suitable for edge deployment