# Towards generalisable and calibrated synthetic speech detection with self-supervised representations

arXiv: https://arxiv.org/abs/2309.05384 · HTML: https://mlsanity.com/paper/2309.05384/

## Key facts

- Cited by 23 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 11 Sep 2023.

## Built on by (most cited first, 23 of 23)

- [The Codecfake Dataset and Countermeasures for the Universally Detection of Deepfake Audio](https://mlsanity.com/paper/2405.04880/) (2024)
- [Circumventing shortcuts in audio-visual deepfake detection datasets with unsupervised learning](https://mlsanity.com/paper/2412.00175/) (2024)
- [Nes2Net: A Lightweight Nested Architecture for Foundation Model Driven Speech Anti-spoofing](https://mlsanity.com/paper/2504.05657/) (2025)
- [DeCLIP: Decoding CLIP representations for deepfake localization](https://mlsanity.com/paper/2409.08849/) (2024)
- [Training-Free Deepfake Voice Recognition by Leveraging Large-Scale Pre-Trained Models](https://mlsanity.com/paper/2405.02179/) (2024)
- [Detect All-Type Deepfake Audio: Wavelet Prompt Tuning for Enhanced Auditory Perception](https://mlsanity.com/paper/2504.06753/) (2025)
- [Multilingual Source Tracing of Speech Deepfakes: A First Benchmark](https://mlsanity.com/paper/2508.04143/) (2025)
- [Replay Attacks Against Audio Deepfake Detection](https://mlsanity.com/paper/2505.14862/) (2025)
- [Do Compact SSL Backbones Matter for Audio Deepfake Detection? A Controlled Study with RAPTOR](https://mlsanity.com/paper/2603.06164/) (2026)
- [Improving Out-of-Domain Audio Deepfake Detection via Layer Selection and Fusion of SSL-Based Countermeasures](https://mlsanity.com/paper/2509.12003/) (2025)
- [AT-ADD: All-Type Audio Deepfake Detection Challenge Evaluation Plan](https://mlsanity.com/paper/2604.08184/) (2026)
- [DeePen: Penetration Testing for Audio Deepfake Detection](https://mlsanity.com/paper/2502.20427/) (2025)
- [A Practical Synthesis of Detecting AI-Generated Textual, Visual, and Audio Content](https://mlsanity.com/paper/2504.02898/) (2025)
- [Multilingual Dataset Integration Strategies for Robust Audio Deepfake Detection: A SAFE Challenge System](https://mlsanity.com/paper/2508.20983/) (2025)
- [A SUPERB-Style Benchmark of Self-Supervised Speech Models for Audio Deepfake Detection](https://mlsanity.com/paper/2603.01482/) (2026)
- [Echoes: A semantically-aligned music deepfake detection dataset](https://mlsanity.com/paper/2603.23667/) (2026)
- [Dual-Granularity Orthogonal Disentanglement for Generalizable Audio Deepfake Detection](https://mlsanity.com/paper/2606.16532/) (2026)
- [Detecting Audio Deepfakes on the Edge:Lightweight SSL-Based Detection in a Browser Plugin](https://mlsanity.com/paper/2606.30780/) (2026)
- [Audio-Anchored Fusion of Multi-Ratio DiT Reconstruction Residuals for Cross-Domain Audio Deepfake Detection](https://mlsanity.com/paper/2607.26472/) (2026)
- [CRAF: Cross-View Residual-Aware Fusion for Deepfake Speech Detection](https://mlsanity.com/paper/2609.13842/) (2026)
- [Spooftral: Can Voxtral Audio-Language Model Detect Speech Spoofing?](https://mlsanity.com/paper/2609.28713/) (2026)
- [What Survives the Codec Shift: Pooled No-Vocals Residuals for Speech Deepfake Detection](https://mlsanity.com/paper/2609.33375/) (2026)
- [GLAD: Global-Local Adaptive Detector for Robust Speech Deepfake Detection](https://mlsanity.com/paper/2609.35411/) (2026)

## Abstract (arXiv)

Generalisation -- the ability of a model to perform well on unseen data -- is crucial for building reliable deepfake detectors. However, recent studies have shown that the current audio deepfake models fall short of this desideratum. In this work we investigate the potential of pretrained self-supervised representations in building general and calibrated audio deepfake detection models. We show that large frozen representations coupled with a simple logistic regression classifier are extremely effective in achieving strong generalisation capabilities: compared to the RawNet2 model, this approach reduces the equal error rate from 30.9% to 8.8% on a benchmark of eight deepfake datasets, while learning less than 2k parameters. Moreover, the proposed method produces considerably more reliable predictions compared to previous approaches making it more suitable for realistic use.