Easy, Interpretable, Effective: openSMILE for voice deepfake detection
This work addresses the problem of voice authenticity for security applications by providing an effective and interpretable method, though it is incremental as it applies existing features to a new dataset.
The paper tackles voice deepfake detection by showing that simple, interpretable features from openSMILE can identify attacks in the ASVspoof5 dataset with high accuracy, achieving up to 0.8% Equal Error Rate (EER) for specific attacks and an overall EER of 15.7 ± 6.0%.
In this paper, we demonstrate that attacks in the latest ASVspoof5 dataset -- a de facto standard in the field of voice authenticity and deepfake detection -- can be identified with surprising accuracy using a small subset of very simplistic features. These are derived from the openSMILE library, and are scalar-valued, easy to compute, and human interpretable. For example, attack A10`s unvoiced segments have a mean length of 0.09 +- 0.02, while bona fide instances have a mean length of 0.18 +- 0.07. Using this feature alone, a threshold classifier achieves an Equal Error Rate (EER) of 10.3% for attack A10. Similarly, across all attacks, we achieve up to 0.8% EER, with an overall EER of 15.7 +- 6.0%. We explore the generalization capabilities of these features and find that some of them transfer effectively between attacks, primarily when the attacks originate from similar Text-to-Speech (TTS) architectures. This finding may indicate that voice anti-spoofing is, in part, a problem of identifying and remembering signatures or fingerprints of individual TTS systems. This allows to better understand anti-spoofing models and their challenges in real-world application.