Toward Robust Audio Spoofing Detection: A Detailed Comparison of Traditional and Learned Features
Open accessThe DOI matches its Crossref record; 9 of the 9 references it cites were found in a registry or an index.
Abstract
Automatic speaker verification, such as every other biometric system, is vulnerable to spoofing attacks. Using only a few minutes of recorded voice from a genuine client of a speaker verification system, attackers can develop a variety of spoofing attacks that might trick such systems. Detecting these attacks using the audio cues present in the recordings is an important challenge. Most existing spoofing detection systems depend on knowing the used spoofing technique. With this research, we aim at overcoming this limitation, by examining robust audio features, both traditional and those learned through an autoencoder, which is generalizable to different types of replay spoofing. Furthermore, we provide a detailed account of all the steps necessary in setting up the state-of-the-art audio feature detection, preprocessing, and postprocessing, such that the (non-audio expert) machine learning researcher can implement such systems. Finally, we evaluate the performance of our robust replay spoofing detection system with a wide variety and different combinations of both extracted and machine-learned audio features on the “out in the wild” ASVspoof 2017 dataset. This dataset contains a variety of new replay spoofing configurations. Since our focus is on examining which features will ensure robustness, we base our system on a traditional Gaussian mixture model-universal background model (GMM-UBM). We then systematically investigate the relative contribution of each feature set. The fused models based on both the known audio features and the machine learned features, respectively, have a comparable performance with an equal error rate (EER) of 12. The final best performing model, which obtains an EER of 10.8, is a hybrid system that contains both known and machine-learned features and is trained on an augmented dataset, thus revealing the importance of incorporating both types of features when developing a robust spoofing prediction model.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Spoofing and countermeasures for speaker verification: A survey
Wu, Z., Evans, N., Kinnunen, T., Yamagishi, J., Alegre, F., & Li, H. (2015) · Speech Communication · doi:10.1016/j.specom.2014.10.005 · Cited by 601
The ASVspoof 2017 challenge: assessing the limits of replay spoofing attack detection
Kinnunen, T., Sahidullah, M., Delgado, H., Todisco, M., Evans, N., Yamagishi, J., & Lee, K. A. (2017) · Interspeech 2017 · doi:10.21437/interspeech.2017-1111 · Cited by 535
A comparison of features for synthetic speech detection
Sahidullah, M., Kinnunen, T., & Hanilçi, C. (2015) · Interspeech 2015 · doi:10.21437/interspeech.2015-472 · Cited by 382 · cited 5×
“The procedure for calculating LFCCs is similar to that of RFCCs [41], however, it uses triangular filters instead of rectangular filters [19], [42].”
BackgroundMethodInfluentialAudio replay attack detection with deep learning frameworks
Lavrentyeva, G., Novoselov, S., Malykh, E., Kozlov, A., Kudashev, O., & Shchemelinin, V. (2017) · Interspeech 2017 · doi:10.21437/interspeech.2017-360 · Cited by 292
An Introduction to Application-Independent Evaluation of Speaker Recognition Systems
van Leeuwen, D. A., & Brümmer, N. (2007) · Lecture Notes in Computer Science · doi:10.1007/978-3-540-74200-5_19 · Cited by 179 · cited 1×
“We included a logistic regression fusion procedure [35], [36] to make a final prediction.”
MethodThe reddots data collection for speaker recognition
Lee, K. A., Larcher, A., Wang, G., Kenny, P., Brümmer, N., Leeuwen, D. V., Aronowitz, H., Kockmann, M., Vaquero, C., Ma, B., et al. (2015) · Interspeech 2015 · doi:10.21437/interspeech.2015-95 · Cited by 136
ASVspoof 2021: Automatic Speaker Verification Spoofing and Countermeasures Challenge Evaluation Plan
Delgado, H., Evans, N., Kinnunen, T., Lee, K. A., Liu, X., Nautsch, A., Patino, J., Sahidullah, M., Todisco, M., Wang, X., et al. (2021) · arXiv · doi:10.48550/arxiv.2109.00535 · Cited by 129
DNN Filter Bank Cepstral Coefficients for Spoofing Detection
Yu, H., Tan, Z.-H., Zhang, Y., Ma, Z., & Guo, J. (2017) · IEEE Access · doi:10.1109/access.2017.2687041 · Cited by 67
Alegre, F., Vipperla, R., & Evans, N. (2012) · Interspeech 2012 · doi:10.21437/interspeech.2012-462 · Cited by 43
Research works printed but not linked: 60
- rabiner, 1993, Fundamentals of speech recognition, 14
- muda, 2010, Voice recognition algorithms using MEL frequency cepstral coefficient (MFCC) and dynamic time warping (DTW) techniques, arXiv 1003 4083
- 10.1016/S0167-6393(98)00033-8 · DOI
- strand, 2004, Cepstral mean and variance normalization in the model domain, Proc COST278 ISCA Tutorial Res Workshop (ITRW) Robustness Issues Conversational Interact, 38
- hubens, 2017, Deep Inside Autoencoders
- ng, 2011, Sparse Autoencoder, 1
- 10.1109/MSP.2012.2205597 · DOI · Paper page
- 10.1080/00450618.2012.733025 · DOI · Paper page
- sadjadi, 2013, MSR identity toolbox V1. 0: A MATLAB toolbox for speaker-recognition research, Speech Lang Process Tech Comm Newsl, 1, 1
- 10.1109/ICASSP.2017.7953187 · DOI · Paper page
- 10.1016/0167-6393(95)00017-I · doi:10.1016/0167-6393(95)00017-i
- 10.1109/89.928915 · DOI · Paper page
- martin, 1997, The DET curve in assessment of detection task performance, Proc EUROSPEECH, 1895, 10.21437/Eurospeech.1997-504 · DOI · Paper page
- 10.1109/ICASSP.1979.1170788 · DOI · Paper page
- gonçalves, 2017, On the generalization of fused systems in voice presentation attack detection, Proc Int Conf Biometrics Special Interest Group (BIOSIG), 1
- brooks, 2013, The Voicebox Toolkit
- 10.21437/Interspeech.2017-1246 · DOI
- 10.1007/978-3-319-92627-8_16 · DOI
- 10.21437/Interspeech.2017-456 · DOI
- galston, 2004, The Internet in Public Life
- levy, 1984, Hackers Heroes of the Computer Revolution, 14
- 10.21437/Interspeech.2016-1280 · DOI · Paper page
- kinnunen, 2018, The 2nd automatic speaker verification spoofing and countermeasures challenge (asvspoof 2017) database version 2
- 10.1006/dspr.1999.0361 · DOI · Paper page
- 10.1002/j.1538-7305.1987.tb00198.x · DOI
- 10.1109/CCECE.1995.526613 · DOI
- matrouf, 2007, A straightforward and efficient implementation of the factor analysis model for speaker verification, Proc 8th Annu Conf Int Speech Commun Assoc, 1242
- 10.1109/PROC.1975.9792 · DOI · Paper page
- 2017, DSP System Toolbox Transforms and Spectral analysis Linear prediction
- 10.1016/S0893-6080(05)80056-5 · DOI · Paper page
- 10.1109/ICMLA.2015.199 · DOI
- nair, 2014, Comparison between MEL-frequency and complex cepstral coefficients for forensic voice comparison using a likelihood ratio framework, Proc World Congr Eng Comput Sci, 1
- bogert, 1963, The quefrency alanysis of time series for echoes; Cepstrum, pseudo-autocovariance, cross-cepstrum and saphe cracking, Proc Symp Time Ser Anal, 209
- rabiner, 2011, Theory and Applications of Digital Speech Processing, 64
- 10.1109/TAU.1968.1161965 · DOI · Paper page
- 10.1109/ICASSP.1998.675340 · DOI · Paper page
- kua, 2010, Investigation of spectral centroid magnitude and frequency for speaker recognition, Proc Odyssey, 7
- 10.1109/ICASSP.2012.6288895 · DOI · Paper page
- 10.21437/Odyssey.2016-41 · DOI · Paper page
- kons, 2013, Voice transformation-based spoofing of text-dependent speaker verification systems, Proc INTERSPEECH, 945, 10.21437/Interspeech.2013-292 · DOI
- 10.1121/1.4830853 · DOI
- 10.1007/978-3-642-19530-3_25 · DOI
- de leon, 2011, Detection of synthetic speech for the problem of imposture, Proc IEEE Int Conf Acoust Speech Signal Process (ICASSP), 4844
- wu, 2012, Detecting converted speech and natural speech for anti-spoofing attack in speaker recognition, Proc Annu Conf Int Speech Commun Assoc, 1700
- lau, 2004, Vulnerability of speaker verification to voice mimicking, Proc Int Symp Intell Multimedia Video Speech Process, 145
- hautamäki, 2013, I-vectors meet imitators: On vulnerability ofspeaker verification systems against voice mimicry, Proc INTERSPEECH, 930, 10.21437/Interspeech.2013-289 · DOI · Paper page
- 10.1007/978-3-319-24177-7_30 · DOI
- 10.1109/BTAS.2015.7358783 · DOI · Paper page
- de leon, 2010, Evaluation of the vulnerability of speaker verification to synthetic speech, IEEE Odyssey Speaker and Language Recognition Workshop
- 10.1109/APSIPA.2014.7041636 · DOI · Paper page
- wu, 2014, ASVspoof 2015: The first automatic verification spoofing and countermeasures challenge evaluation plan, Proc IEEE Signal Process Soc Speech Lang Tech Committee Newslett
- 10.1109/TASL.2012.2201472 · DOI · Paper page
- 10.1007/s40595-016-0071-3 · DOI
- 10.1007/s00521-018-3933-z · DOI
- 10.1016/j.csl.2017.01.001 · DOI · Paper page
- 10.1109/LSP.2010.2100380 · DOI · Paper page
- 10.1109/BTAS.2013.6712706 · DOI · Paper page
- hasan, 2013, CRSS systems for 2012 NIST speaker recognition evaluation, Proc IEEE Int Conf Acoust Speech Signal Process (ICASSP), 6783
- 10.1016/j.asoc.2016.12.024 · DOI · Paper page
- chakroborty, 2007, Improved closed set text-independent speaker identification by combining mfcc with evidence from flipped filter banks, Int J Signal Process, 4, 114
Cited by
Deepfake generation and detection: Case study and challenges
Patel, Y., Tanwar, S., Gupta, R., Bhattacharya, P., Davidson, I. E., Nyameko, R., Aluvala, S., & Vimal, V. (2023) · IEEE Access · doi:10.1109/access.2023.3342107 · Cited by 186
“In [92], the authors proposed a hybrid approach in detecting spoofed audio.”
BackgroundA deep learning framework for audio deepfake detection
Khochare, J., Joshi, C., Yenarkar, B., Suratkar, S., & Kazi, F. (2022) · Arabian Journal for Science and Engineering · doi:10.1007/s13369-021-06297-w · Cited by 135
Khanjani, Z., Watson, G., & Janeja, V. P. (2023) · Frontiers in Big Data · doi:10.3389/fdata.2022.1001063 · Cited by 121
A Survey on Speech Deepfake Detection
Li, M., Ahmadiadli, Y., & Zhang, X.-P. (2025) · ACM Computing Surveys · doi:10.1145/3714458 · Cited by 101
“…[22], FastAudio [41], SincNet [202] Deep embed-dings ResNet [140], X-vector [16], auto-encoder [9], Bi-LSTM [65], U-net [17] Pre-trainedembeddings Utilizing SSL models or other DL models pre-trained by external large datasets to extract latent representations of the raw speech waveform. wav2vec2.”
A Comprehensive Review of DeepFake Detection Using Advanced Machine Learning and Fusion Methods
Gupta, G., Raja, K., Gupta, M., Jan, T., Whiteside, S. T., & Prasad, M. (2023) · Electronics · doi:10.3390/electronics13010095 · Cited by 99
Metadata
- Authors
- Balamurali, B. T., Lin, K. E., Lui, S., Chen, J.-M., & Herremans, D.
- Venue
- IEEE Access
- Published
- 2019
- Volume
- 7
- Pages
- 84229-84241
- ISSN
- 2169-3536 (ISSN-L)
- OpenAlex
- W2950733131
- Cited by
- 75
- References
- 9
- FWCI
- 3.80
- Type
- journal article
- Language
- en
- Open access
- gold
- Funders
- National Institute of Education (AFD 05/15 SL)
- Singapore University of Technology and Design (SRG ISTD 2017 129)
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.
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
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.