Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2603.20686 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917356238798848 |
|---|---|
| author | Jung, Kyudan Kim, Jihwan Lee, Minwoo Kim, Soyoon Kim, Jeonghoon Choo, Jaegul Park, Cheonbok |
| author_facet | Jung, Kyudan Kim, Jihwan Lee, Minwoo Kim, Soyoon Kim, Jeonghoon Choo, Jaegul Park, Cheonbok |
| contents | Recent advancements in text-to-speech technologies enable generating high-fidelity synthetic speech nearly indistinguishable from real human voices. While recent studies show the efficacy of self-supervised learning-based speech encoders for deepfake detection, these models struggle to generalize across unseen speakers. Our quantitative analysis suggests these encoder representations are substantially influenced by speaker information, causing detectors to exploit speaker-specific correlations rather than artifact-related cues. We call this phenomenon speaker entanglement. To mitigate this reliance, we introduce SNAP, a speaker-nulling framework. We estimate a speaker subspace and apply orthogonal projection to suppress speaker-dependent components, isolating synthesis artifacts within the residual features. By reducing speaker entanglement, SNAP encourages detectors to focus on artifact-related patterns, leading to state-of-the-art performance. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_20686 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | SNAP: Speaker Nulling for Artifact Projection in Speech Deepfake Detection Jung, Kyudan Kim, Jihwan Lee, Minwoo Kim, Soyoon Kim, Jeonghoon Choo, Jaegul Park, Cheonbok Sound Artificial Intelligence Recent advancements in text-to-speech technologies enable generating high-fidelity synthetic speech nearly indistinguishable from real human voices. While recent studies show the efficacy of self-supervised learning-based speech encoders for deepfake detection, these models struggle to generalize across unseen speakers. Our quantitative analysis suggests these encoder representations are substantially influenced by speaker information, causing detectors to exploit speaker-specific correlations rather than artifact-related cues. We call this phenomenon speaker entanglement. To mitigate this reliance, we introduce SNAP, a speaker-nulling framework. We estimate a speaker subspace and apply orthogonal projection to suppress speaker-dependent components, isolating synthesis artifacts within the residual features. By reducing speaker entanglement, SNAP encourages detectors to focus on artifact-related patterns, leading to state-of-the-art performance. |
| title | SNAP: Speaker Nulling for Artifact Projection in Speech Deepfake Detection |
| topic | Sound Artificial Intelligence |
| url | https://arxiv.org/abs/2603.20686 |