The DOI matches its Crossref record; 41 of the 41 references it cites were found in a registry or an index; none of the 11 sources its claims cite is checked for support yet.
Does Audio Deepfake Detection Generalize?
Open accessAbstract
Current text-to-speech algorithms produce realistic fakes of human voices, making deepfake detection a much-needed area of research. While researchers have presented various techniques for detecting audio spoofs, it is often unclear exactly why these architectures are successful: Preprocessing steps, hyperparameter settings, and the degree of fine-tuning are not consistent across related work. Which factors contribute to success, and which are accidental? In this work, we address this problem: We systematize audio spoofing detection by re-implementing and uniformly evaluating architectures from related work. We identify overarching features for successful audio deepfake detection, such as using cqtspec or logspec features instead of melspec features, which improves performance by 37% EER on average, all other factors constant. Additionally, we evaluate generalization capabilities: We collect and publish a new dataset consisting of 37.9 hours of found audio recordings of celebrities and politicians, of which 17.2 hours are deepfakes. We find that related work performs poorly on such real-world data (performance degradation of up to one thousand percent). This may suggest that the community has tailored its solutions too closely to the prevailing ASVSpoof benchmark and that deepfakes are much harder to detect outside the lab than previously thought.
Overview
CoChat Summary
TL;DR
This paper re-implements and uniformly evaluates twelve audio deepfake detection architectures, finding that feature choice and input length strongly affect performance. It also introduces a new 37.9-hour “in-the-wild” dataset and shows that models trained on ASVspoof 2019 generalize poorly to real-world deepfakes, with EER degradation up to 1000%.
Key findings
- Raw-waveform models outperform spectrogram-based models: up to 1.2% EER on ASVspoof and 33.9% EER on the “in-the-wild” dataset (RawGAT-ST and RawNet2), versus up to 6.3% EER on ASVspoof and 37.4% on “in-the-wild” for spectrogram-based models (LCNN and MesoNet).
- Replacing melspec with cqtspec or logspec improves performance by 37% EER on average, all other factors constant.
- Full-length audio input outperforms fixed 4-second input: average EER on ASVspoof drops from 19.89% to 9.85% when full-length input is used.
- On the new “in-the-wild” dataset, EER values deteriorate by about 200 to 1000 percent compared with ASVspoof 2019, and models often perform no better than random guessing; training RawNet2 with 4s input on all ASVspoof 2019 splits yields 33.1 ± 0.2% EER, no improvement over training with only train and dev data.
Methods
- The authors re-implemented twelve architectures from related work and evaluated them under a common standard, training on the ASVspoof 2019 Logical Access train and dev splits and testing on the ASVspoof 2019 eval split and a new “in-the-wild” dataset.
- The new dataset contains 37.9 hours of audio: 17.2 hours of deepfakes and 20.7 hours of authentic material from 58 English-speaking celebrities and politicians; fake clips were created by segmenting 219 publicly available video and audio files that explicitly advertise audio deepfakes, and authentic clips were manually collected from public sources. Clips average 4.3 seconds and were downsampled to 16 kHz.
- Models using preprocessed features were evaluated on 513-dimensional cqtspec, logspec, and melspec features; other models used raw waveforms. Experiments compared fixed 4-second inputs with variable-length inputs, were run three times with random initialization, and report averaged EER and t-DCF with standard deviation; t-DCF is reported only for ASVspoof.
Limitations
- Stated in the paper: the authors do not fine-tune hyperparameters, which likely explains why their ASVspoof implementations perform about 2–4% EER and 0.1 t-DCF worse than related work. t-DCF is not reported for the “in-the-wild” dataset because false alarm and miss costs are available only for ASVspoof. The new dataset is intended as evaluation data.
- Stated in the paper: all ASVspoof splits are fundamentally based on the same VCTK dataset, so adding the eval split does not improve real-world generalization.
- Evident from the text: the “in-the-wild” dataset is limited to English-speaking celebrities and politicians, uses fake clips from publicly advertised deepfakes, relies on manual collection of authentic counterparts, and consists of short clips averaging 4.3 seconds downsampled to 16 kHz; evaluation covers ASVspoof 2019 Logical Access and this new dataset only.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
0 Valid · 0 Not valid · 8 Not checked · 35 Own results and interpretations
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We collect and publish a new dataset consisting of 37.9 hours of found audio recordings of celebrities and politicians, of which 17.2 hours are deepfakes.”
- Also stated in: § 3. Datasets · plain
“It consists of 37.9 hours of audio clips that are either fake (17.2 hours) or real (20.7 hours).”
§ Does Audio Deepfake Detection Generalize? > Abstract
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Also stated in: § 3. Datasets · plain
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We find that related work performs poorly on such real-world data (performance degradation of up to one thousand percent).”
§ Does Audio Deepfake Detection Generalize? > Abstract
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We identify overarching features for successful audio deepfake detection, such as using cqtspec or logspec features instead of melspec features, which improves performance by 37% EER on average, all other factors constant.”
§ Does Audio Deepfake Detection Generalize? > Abstract
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We see that there is a large performance gap between the ASVSpoof 2019 evaluation data and our proposed 'in-the-wild' dataset.”
- Also stated in: § 6. Conclusion · plain
“Performance drops sharply, and some models degenerate to random guessing.”
§ Does Audio Deepfake Detection Generalize? > 5.3. Evaluation on 'in-the-wild' data
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Also stated in: § 6. Conclusion · plain
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“For n = 58 celebrities and politicians, we collected both bona-fide and spoofed audio (represented by blue and red boxes per speaker).”
§ Does Audio Deepfake Detection Generalize? > 6. Conclusion
plainn = 58- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
22 more claims
Create a free account to see all 43 claims and every source verdict.
- The quote only partly states this claim. 2
- It may state two separate claims as one. 9
- Probably not a substantial claim. 8
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“These constraints hinder the models’ ability to generalise to real-world deepfake audio.”
From Audio Deepfake Detection: What Has Been Achieved and What Lies Ahead (2025) · cites this paper as [64]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“AISEC “In-the-Wild” [43] is a synthetic speech dataset designed to replicate real-world conditions. It includes 38 h of audio clips (17 h fake, 21 h real) featuring English-speaking celebrities and politicians.”
From Deepfake Media Forensics: Status and Future Challenges (2025) · cites this paper as [43]
The cited source says what the claim attributes to it.
In the citing paper
“In 2022, Müller et al. [64] pointed out that most audio deepfake detection models are trained and evaluated on the ASVspoof dataset, which includes limited spoofing algorithms and is based on the VCTK corpus recorded in controlled studio environments.”
From Audio Deepfake Detection: What Has Been Achieved and What Lies Ahead (2025) · cites this paper as [64]
The cited source says what the claim attributes to it.
In the citing paper
“a significant body of research [64,96,97,98,99] has demonstrated that the performance of detection models largely depends on the choice of input features.”
From Audio Deepfake Detection: What Has Been Achieved and What Lies Ahead (2025) · cites this paper as [64]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Does Audio Deepfake Detection Generalize? > 1. Introduction
“We collect 17.2 hours of high-quality audio deepfakes and 20.7 hours of of authentic material from 58 politicians and celebrities.”
§ Does Audio Deepfake Detection Generalize? > 6. Conclusion
“In total, we collected 20.8 hours of bona-fide and 17.2 hours of spoofed audio.”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Does Audio Deepfake Detection Generalize? > 6. Conclusion
“For n = 58 celebrities and politicians, we collected both bona-fide and spoofed audio (represented by blue and red boxes per speaker).In total, we collected 20.8 hours of bona-fide and 17.2 hours of spoofed audio.”
§ Does Audio Deepfake Detection Generalize? > 6. Conclusion
“On average, there are 23 minutes of bona-fide and 18 minutes of spoofed audio per speaker.”
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., & Polosukhin, I. (2017) · arXiv · doi:10.48550/arxiv.1706.03762 · Cited by 120,404
Adam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946
Exploiting Generative AI to Scale up Intelligent Tutoring Systems
Jakubův, J., Chvalovský, K., Goertzel, Z., Kaliszyk, C., Olšák, M., Piotrowski, B., Schulz, S., Suda, M., & Urban, J. (2023) · DROPS (Schloss Dagstuhl – Leibniz Center for Informatics) · doi:10.4230/lipics.itp.2023.19 · Cited by 76,658
Going deeper with convolutions
Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., & Rabinovich, A. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298594 · Cited by 48,562
SciPy 1.0: fundamental algorithms for scientific computing in Python
Virtanen, P., Gommers, R., Oliphant, T. E., Haberland, M., Reddy, T., Cournapeau, D., Burovski, E., Peterson, P., Weckesser, W., Bright, J., et al. (2020) · Nature Methods · doi:10.1038/s41592-019-0686-2 · Cited by 38,430
SciPy 1.0: fundamental algorithms for scientific computing in Python
Virtanen, P., Gommers, R., Oliphant, T. E., Haberland, M., Reddy, T., Cournapeau, D., Burovski, E., Peterson, P., Weckesser, W., Bright, J., et al. (2019) · Apollo - University of Cambridge Repository · doi:10.17863/cam.64134 · Cited by 10,460
WaveNet: a generative model for raw audio
Oord, A. V. D., Dieleman, S., Zen, H., Simonyan, K., Vinyals, O., Graves, A., Kalchbrenner, N., Senior, A., & Kavukcuoglu, K. (2016) · arXiv · doi:10.48550/arxiv.1609.03499 · Cited by 4,911
librosa: Audio and Music Signal Analysis in Python
McFee, B., Raffel, C., Liang, D., Ellis, D., McVicar, M., Battenberg, E., & Nieto, O. (2015) · Proceedings of the Python in Science Conference · doi:10.25080/majora-7b98e3ed-003 · Cited by 3,033
MesoNet: a Compact Facial Video Forgery Detection Network
Afchar, D., Nozick, V., Yamagishi, J., & Echizen, I. (2018) · 2018 IEEE International Workshop on Information Forensics and Security (WIFS) · doi:10.1109/wifs.2018.8630761 · Cited by 1,872
Research works printed but not linked: 1
- A. Nautsch, X. Wang, N. Evans, T. H. Kinnunen, V. Vestman, M. Todisco, H. Delgado, M. Sahidullah, J. Yamagishi, and K. A. Lee, "ASVspoof 2019: Spoofing Countermeasures for the De- tection of Synthesized, Converted and Replayed Speech," vol. 3, no. 2, pp. 252-265.
Cited by
Deepfake: definitions, performance metrics and standards, datasets, and a meta-review
Altuncu, E., Franqueira, V. N. L., & Li, S. (2024) · Frontiers in Big Data · doi:10.3389/fdata.2024.1400024 · Cited by 134
Deepfake Media Forensics: Status and Future Challenges
Amerini, I., Barni, M., Battiato, S., Bestagini, P., Boato, G., Bruni, V., Caldelli, R., De Natale, F., De Nicola, R., Guarnera, L., et al. (2025) · Journal of Imaging · doi:10.3390/jimaging11030073 · Cited by 113
A Survey on Speech Deepfake Detection
Li, M., Ahmadiadli, Y., & Zhang, X.-P. (2025) · ACM Computing Surveys · doi:10.1145/3714458 · Cited by 101
AI-Synthesized Voice Detection Using Neural Vocoder Artifacts
Sun, C., Jia, S., Hou, S., & Lyu, S. (2023) · 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw59228.2023.00097 · Cited by 81
MLAAD: The Multi-Language Audio Anti-Spoofing Dataset
Müller, N. M., Kawa, P., Choong, W. H., Casanova, E., Gölge, E., Müller, T., Syga, P., Sperl, P., & Böttinger, K. (2024) · 2024 International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn60899.2024.10650962 · Cited by 73
Self-citation
Metadata
- Authors
- Müller, N., Czempin, P., Diekmann, F., Froghyar, A., & Böttinger, K.
- Affiliations
- Fraunhofer Institute for Applied and Integrated Security (DE)
- Technical University of Munich (DE)
- Corresponding author
- Müller, Nicolas
- Venue
- Interspeech 2022
- Publisher
- ISCA
- Published
- 2022-09-18
- Pages
- 2783-2787
- ISSN
- 2308-457X (ISSN-L)
- OpenAlex
- W4297841787
- Cited by
- 258
- References
- 29
- Citations by year
- 2026: 82 · 2025: 93 · 2024: 60 · 2023: 19 · 2022: 4
- FWCI
- 24.31
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- Interspeech (USB) (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- Fraunhofer-Publica (Fraunhofer-Gesellschaft) (submittedVersion)
- arXiv (Cornell University)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Speech Recognition and Synthesis
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Digital Media Forensic Detection
- Physical Sciences › Computer Science › Signal Processing › Music and Audio Processing
- Keywords
- audio deepfake detection, generalization, synthetic speech, audio forensics
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
A language model drew the claims from the paper on 2026-10-29, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-29 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Does Audio Deepfake Detection Generalize? https://app.cochat.ai/papers/10.21437/interspeech.2022-108
- The paper
- Müller, N., Czempin, P., Diekmann, F., Froghyar, A., & Böttinger, K. (2022). Does Audio Deepfake Detection Generalize? Interspeech 2022. https://doi.org/10.21437/interspeech.2022-108
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.