A comparison of features for synthetic speech detection
Open accessThe DOI matches its Crossref record; 36 of the 36 references it cites were found in a registry or an index.
Abstract
The performance of biometric systems based on automatic speaker recognition technology is severely degraded due to spoofing attacks with synthetic speech generated using different voice conversion (VC) and speech synthesis (SS) techniques. Various countermeasures are proposed to detect this type of at-tack, and in this context, choosing an appropriate feature extrac-tion technique for capturing relevant information from speech is an important issue. This paper presents a concise experi-mental review of different features for synthetic speech detec-tion task. A wide variety of features considered in this study include previously investigated features as well as some other potentially useful features for characterizing real and synthetic speech. The experiments are conducted on recently released ASVspoof 2015 corpus containing speech data from a large number of VC and SS technique. Comparative results using two different classifiers indicate that features representing spectral information in high-frequency region, dynamic information of speech, and detailed information related to subband characteris-tics are considerably more useful in detecting synthetic speech. Index Terms: anti-spoofing, ASVspoof 2015, feature extrac-tion, countermeasures
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
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?
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 [96]
In the citing paper
“other machine learning models, such as the Support Vector Machine and Logistic Regression, were widely studied in earlier research [96,134,135,136]. With the advancement of feature extraction techniques, these models continue to play a significant role in current studies [137,138,139].”
From Audio Deepfake Detection: What Has Been Achieved and What Lies Ahead (2025) · cites this paper as [96]
Citations
Cites
Multiresolution gray-scale and rotation invariant texture classification with local binary patterns
Ojala, T., Pietikainen, M., & Maenpaa, T. (2002) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2002.1017623 · Cited by 15,398 · cited 1×
“The authors of [8], in turn, proposed to use one-class approach us-ing local binary pattern [9] of linear frequency cepstral coef-ficients (LFCCs) followed by support vector machine (SVM) for voice conversion, speech synthesis and artificial signal detection [8].”
MethodFundamentals of speech recognition
Rabiner, L. R., & Juang, B.-H. F. (1993) · Cited by 7,680
Davis, S., & Mermelstein, P. (1980) · IEEE Transactions on Acoustics, Speech, and Signal Processing · doi:10.1109/tassp.1980.1163420 · Cited by 5,465 · cited 1×
“In MFCC, the filters are placed in mel scale, having denser spacing in t he low-frequency region [21].”
MethodPerceptual linear predictive (PLP) analysis of speech
Hermansky, H. (1990) · The Journal of the Acoustical Society of America · doi:10.1121/1.399423 · Cited by 2,536 · cited 1×
“Another all-pole representation of speech called perceptual linear prediction cepstral coefficients (PLPCC) is also computed by first performing a series of perceptual processing prior to LP analysis [24].”
MethodCepstral analysis technique for automatic speaker verification
Furui, S. (1981) · IEEE Transactions on Acoustics, Speech, and Signal Processing · doi:10.1109/tassp.1981.1163530 · Cited by 1,173
Continuous probabilistic transform for voice conversion
Stylianou, Y., Cappe, O., & Moulines, E. (1998) · IEEE Transactions on Speech and Audio Processing · doi:10.1109/89.661472 · Cited by 931 · cited 1×
“Synthetic speech signal created using different voice conv ersion (VC) and speech synthesis (SS) techniques can be used to poof biometric systems based on automatic speaker recognition t echnology [1, 2, 3, 4].”
MethodConstruction and evaluation of a robust multifeature speech/music discriminator
Scheirer, E., & Slaney, M. (2002) · 1997 IEEE International Conference on Acoustics, Speech, and Signal Processing · doi:10.1109/icassp.1997.596192 · Cited by 873 · cited 1×
“Spectral flux based feature: Spectral flux measures the frame-by-frame change in the power spectrum [25].”
BackgroundCOVAREP — A collaborative voice analysis repository for speech technologies
Degottex, G., Kane, J., Drugman, T., Raitio, T., & Scherer, S. (2014) · 2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp.2014.6853739 · Cited by 695 · cited 1×
“2, First 20 coefficients are retained after DCT [5] Short-term APGDF (60) LP Order=20 — phase features CosPhase (60) First 20 coefficients are retained after DCT [5] RPS (60) RPS computed with COVAREP tool [14], 20 filters in mel filter ba nk [6, 7]”
MethodSpoofing 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
Self-citationThe cepstrum: A guide to processing
Childers, D. G., Skinner, D. P., & Kemerait, R. C. (1977) · Proceedings of the IEEE · doi:10.1109/proc.1977.10747 · Cited by 600 · cited 1×
“Cepstrum: Cepstral coefficients (Cep) are computed from the power spectrum by applying discrete cosine transform (DCT) [17].”
Method
Cited by
Items 301–383 of 383
Data Augmentation with Signal Companding for Detection of Logical Access AttacksPreprint
Das, R. K., Yang, J., & Li, H. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2102.06332 · Cited by 2
Ravanbakhsh, M., Setayeshi, S., Pedram, M. M., & Mirzaei, A. (2020) · Advances in Cognitive Science · doi:10.30699/icss.22.2.71 · Cited by 1
Robust Spoofed Speech Detection with Denoised I-vectors
Dişken, G. (2022) · GAZI UNIVERSITY JOURNAL OF SCIENCE · doi:10.35378/gujs.1062788 · Cited by 1
Significance of Quadrature and In-Phase Components for Synthetic Spoofed Speech Detection
Gupta, P., Chodingala, P. K., & Patil, H. A. (2022) · Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) · doi:10.23919/apsipaasc55919.2022.9980241 · Cited by 2
Multi-Sinkhorn Teacher Knowledge Aggregation Framework for Adaptive Audio Anti-Spoofing
Zhang, R., Wei, J., Lu, X., Zhang, L., Jin, D., Lu, W., & Xu, J. (2025) · IEEE Transactions on Audio Speech and Language Processing · doi:10.1109/taslpro.2025.3606191 · Cited by 4
“Early approachesfocusedonhand-craftedfeatureextractors[51],[52], whilelaterworkexploredlearnablefeaturesintegratedwith neuralnetworks,suchasSincNet[53].”
Huang, Y., Zhu, C., Yang, L. T., Ruan, Y., Deng, X., Liu, S., & Hou, R. (2025) · IEEE Transactions on Consumer Electronics · doi:10.1109/tce.2025.3610181 · Cited by 1
“Among them, Spec [22] and its variants, such as MFCC [23], LFCC [24] and linear predictive cepstral coefficients (LPCC) [25], transform the speech signals from time domain to frequency domain, which can well capture the differences of energy distribution in their various frequency components.”
Wu, B., Qian, Q., Ran, L., & Wang, H. (2025) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn64981.2025.11228680 · Cited by 1
“The front-end typically utilizes specific acoustic features, such as Constant-Q Cepstral Coefficients (CQCC), Linear Frequency Cepstral Coefficients (LFCC) and Constant-Q Transform (CQT) [12], which capture spoofing artifacts introduced by speech synthesis and conversion processes.”
Bekiryazıcı, Ş., Hanilçi, C., & Ozcan, N. (2025) · Computer Speech & Language · doi:10.1016/j.csl.2025.101906 · Cited by 4
Self-citationSpeaker Privacy and Security in the Big Data Era: Protection and Defense Against Deepfake
Chen, L., Lee, K. A., Ling, Z.-H., Wang, X., Das, R. K., Toda, T., & Li, H. (2025) · Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) · doi:10.1109/apsipaasc65261.2025.11249028 · Cited by 2
“Feature engineering has advanced from purely using digital signal processing (DSP) algorithms [44] (e.g., linear-frequency cepstrum coefficients [45]) to the hybrid of DSP and deep learning methods (e.g., trainable filters integrated in a convolution network [46]).”
Deepfake Voice Detection for Underrepresented Languages: A Romanian Case Study
Robert-Alin, B., Ioana-Daniela, I., Mihu, C., Robert-Andrei, M., Leon, O. M., & Sorana-Ioana, R. (2025) · International Journal of Advanced Statistics and IT&C for Economics and Life Sciences · doi:10.2478/ijasitels-2025-0003 · Cited by 1
“The implementation is adapted from an open-source project [5].”
Robust Detection of Partially Spoofed Audio Using Semantic-Aware Inconsistency Learning
Cao, J., Tian, H., Tian, P., Li, H., & Wang, J. (2025) · IEEE Transactions on Audio Speech and Language Processing · doi:10.1109/taslpro.2025.3617241 · Cited by 3
Multi-Speaker Conversational Audio Deepfake: Taxonomy, Dataset and Pilot Study
Ahmed, A., Janeja, V. P., & Purushotham, S. (2025) · IEEE ... International Conference on Data Mining workshops · doi:10.1109/icdmw69685.2025.00163 · Cited by 3
“Linear Frequency Ceptral Coefficients (LFCC) [38] is the input of the Light Convolutional Neural Networks (LCNN) [39] as our baseline model.”
Ge, Z., Xu, X., Guo, H., Wang, T., & Yang, Z. (2026) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2026.3698578 · Cited by 1
“…mainly relied on manually crafted acoustic features for ADD, including relative phase shift [20], [21], Cochlear Filter Cepstral Coef-ficients (CFCCs) [22], instantaneous frequency [22], [23], Linear Frequency Cepstral Coe ffi cients (LFCCs) [24], and Constant-Q Cepstral Coe ffi cients (CQCC) [25].”
FLADD: Federated Learning-Based Privacy Protection for Audio Deepfake Detection
Fan, C., Tang, Z., Xue, J., Cao, J., Lv, Z., Wen, Z., & Tao, J. (2026) · IEEE Transactions on Audio Speech and Language Processing · doi:10.1109/taslpro.2026.3727931 · Cited by 1
Generalized Replay Spoofing Countermeasure Based on Combining Local Subclassification Models
Altuwayjiri, S. M., Bchir, O., & Ismail, M. M. B. (2022) · Applied Sciences · doi:10.3390/app122211742 · Cited by 2
“In particular, Q Cepstral coefficients (CQCC) [4] and the linear frequency Cepstral coefficient (LFCC) [5] have been explored.”
BackgroundMethodInfluentialGeneralizing Voice Presentation Attack Detection to Unseen Synthetic Attacks and Channel Variation
Zhang, Y., Jiang, F., Zhu, G., Chen, X., & Duan, Z. (2023) · Advances in computer vision and pattern recognition · doi:10.1007/978-981-19-5288-3_15 · Cited by 2
Residual-based feature enhancement for forgery audio detection
Zheng, W., Ling, X., & Hai, H. (2023) · International Conference on Advanced Algorithms and Signal Image Processing (AASIP) · doi:10.1117/12.3005833 · Cited by 1
“The commonly used conventional methods include GMM[4] and i-vector[5], which have proven effective in culling out and countering deception in automatic speaker recognition from 2015-2021.”
MethodInfluentialQuefrency Domain Features with Residual Networks for Spoof Speech Detection
Kamble, M. R. (2023) · Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) · doi:10.1109/apsipaasc58517.2023.10317474 · Cited by 1
“D feature vector (including 40-∆ and 40 ∆∆ ) [10] CQCC : The CQCC features are extracted with a maximum frequency of F max = F NY Q , where F NY Q is the Nyquist frequency of 44.1kHz.”
BackgroundMethodSanghvi, H., & Mankad, S. H. (2023) · IEEE Region Humanitarian Technology Conference · doi:10.1109/r10-htc57504.2023.10461736 · Cited by 1
“…Fourier Transform (Chroma STFT), Bark-Frequency Cepstral Coefficients (BFCC) [11], Gammatone-Frequency Cepstral Coefficients (GFCC), Linear-Frequency Cepstral Coefficients (LFCC), Magnitude-based Spectral Root Cepstral Coefficients (MSRCC), and Mel-Frequency Cepstral Coefficients (MFCC) [12].”
BackgroundLiu, J., Cui, Z., & Zhou, Y. (2024) · Computers & Electrical Engineering · doi:10.1016/j.compeleceng.2024.109355 · Cited by 1
End-to-End User-Defined Keyword Spotting Using Shifted Delta Coefficients
Kesavaraj, V., Anuprabha, M., & Vuppala, A. K. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-78312-8_25 · Cited by 1
Reject Threshold Adaptation for Open-Set Model Attribution of Deepfake Audio
Yan, X., Yi, J., Tao, J., Chen, Y., Gu, H., Li, G., Zhou, J., Ren, Y., & Xu, T. (2024) · International Symposium on Chinese Spoken Language Processing (ISCSLP) · doi:10.1109/iscslp63861.2024.10800741 · Cited by 3
“The encoder adopts a simple ResNet-18 [24] architecture with input from low-dimensional LFCC features [25].”
AI Generated Speech Detection Using CNN
Demirörs, M., Akgün, T., & Özbayoğlu, A. M. (2024) · IEEE International Conference on Big Data (BigData) · doi:10.1109/bigdata62323.2024.10825541 · Cited by 1
Detecting Spoof Voices in Asian Non-Native Speech: An Indonesian and Thai Case Study
Adila, A., Mawalim, C. O., & Unoki, M. (2024) · Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) · doi:10.1109/apsipaasc63619.2025.10848707 · Cited by 2
“A study on synthetic speech detection in 2015 found that dynamic Linear Frequency Cepstral Coefficients (LFCC) with a GMM classifier performed best on ASVspoof evaluation sets [7].”
Audio Deepfake Detection: End-to-End training with powerful pretrained ASR
Mansoor, K. H., & Alam, M. (2024) · International Multi-Topic Conference · doi:10.1109/inmic64792.2024.11004412 · Cited by 2
“…may be STFT (Short Time Fourier Transform) based such as MFCC (Mel-Frequency Cepstral Coefficients), LFCC (Linear Frequency Cepstral Coefficients) [17], [18], [19] etc while the latent features are explored through various deep neural network models (CNN, RNN, ResNet) [20], [21], [22] which are…”
Improving Voice Spoofing Detection Through Extensive Analysis of Multicepstral Feature Reduction
de Souza, L. M., Guido, R. C., Contreras, R. C., Viana, M. S., & Bongarti, M. A. D. S. (2025) · Sensors · doi:10.3390/s25154821 · Cited by 5
On the Performance of Cepstral Features for Voice-Based Gender Recognition
Kanani, I., Shah, H., & Mankad, S. H. (2018) · Smart innovation, systems and technologies · doi:10.1007/978-981-13-1747-7_31 · Cited by 1
The GMM and I-Vector Systems Based on Spoofing Algorithms for Speaker Spoofing Detection
Tang, H., Lei, Z., Huang, Z., Gan, H., Yu, K., & Yang, Y. (2019) · Lecture notes in computer science · doi:10.1007/978-3-030-31456-9_55 · Cited by 2
Chadha, A. N., Abdullah, A. B., & Angeline, L. (2021) · International Journal of Advanced Computer Science and Applications · doi:10.14569/ijacsa.2021.0120894 · Cited by 3
“The LFCC features are found to be more robust than MFCC in terms of noisy speech as it performs well in the higher frequency region (comprising of VT features).”
MethodInfluentialEnd-to-End Voice Spoofing Detection Employing Time Delay Neural Networks and Higher Order Statistics
Alam, J., Fathan, A., & Kang, W. H. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-87802-3_2 · Cited by 2
Study of approaches to the synthesis and detection of cloned voices (DeepFake)
Ivanov, A. V., Primak, S. A., & Mazurenko, V. A. (2022) · Digital Technology Security · doi:10.17212/2782-2230-2022-3-62-80
Singh, A. K., Singh, P., & Nathwani, K. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2107.11412 · Cited by 3
“Comparison of various features for differentiating AI synthesised speech and human speech was presented in [3].”
BackgroundAudio Replay Attack Detection in Automated Speaker Verification
Anjee, P., Ghosh, S., Kodoor, S., & Shettar, R. (2018) · International Journal of Computer Applications · doi:10.5120/ijca2018916986
Voice conversion with parallel/non-parallel data and synthetic speech detection
Tian, X. (2019) · doi:10.32657/10220/47729
“A comparison work between these features for synthetic speech detection is performed in [102].”
BackgroundMethodResultInfluentialPerceptual Frequency Scale-based Filters for Replay Spoofing Attack Detection
Wickramasinghe, B., Ambikairajah, E., Le, P. N., & Epps, J. (2018) · International Conference on Information and Automation for Sustainability (ICIAfS) · doi:10.1109/iciafs.2018.8913356
“Other commonly used subband energy based features include rectangular filter cepstral coefficients (RFCCs), extracted from rectangular shaped filters that are placed linearly in frequency, and linear filter cepstral coefficients (LFCCs) [7, 8] which are similar to RFCCs, but use triangular shaped filters instead.”
MethodReplay Attack Detection Based on Spatial and Spectral Features of Stereo Signal
Yaguchi, R., Shiota, S., Ono, N., & Kiya, H. (2021) · Journal of Information Processing · doi:10.2197/ipsjjip.29.275 · Cited by 3
“The GMM of each system is trained with spectral features, CQCC [17] and LFCC [18], respectively.”
BackgroundMethodGeneralized Spoofing Detection Inspired from Audio Generation ArtifactsPreprint
Gao, Y., Vuong, T., Elyasi, M., Bharaj, G., & Singh, R. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2104.04111
“Previous studies on anti-spoofing attacks on ASV systems and synthetic speech detection evaluate various features [14, 23] and deep learning models [24] for detection performance.”
BackgroundBuettner, R., Gross, J., Roessler, P., Winter, J., Sauter, D., Baumgartl, H., & Ulrich, P. (2022) · Hawaii International Conference on System Sciences (HICSS) · doi:10.24251/hicss.2022.764
“Fourth, they designed a CNN based on Light CNN-29 [39] and ResNet-18 architecture.”
InfluentialA Novel Secure Speech Biometric Protection Method
Yang, L., Wu, Z., & Guo, J. (2022) · Lecture notes in computer science · doi:10.1007/978-3-030-97774-0_41
A Robust Pipeline based Deep Learning Approach to Detect Speech Attribution
Chakravarty, S., & Khandelwal, R. (2023) · IEEE International Conference for Convergence in Technology (I2CT) · doi:10.1109/i2ct57861.2023.10126219
“[3] Speechelo and Deep Neural Network are known to convert acoustics to TTS algorithm which are used for commercial purposes, like YouTube [4].”
BackgroundAudio DeepFake Detection Employing Multiple Parametric Exponential Linear Units
Alam, M. S., Fathan, A., & Alam, J. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-48312-7_25
A deep learning based design for speech signal deepfake generation and detection
Shrivastava, R., Malika, J., Garima, T., Kumar, G. S., & Ankita, N. (2024) · AIP conference proceedings · doi:10.1063/5.0190845 · Cited by 1
Ji, J., Xie, Y., & Yang, Y. (2023) · Conference proceedings/Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics · doi:10.1109/smc53992.2023.10393996
“The acoustic features we use in this study are linear frequency cepstrum coefficients (LFCC) [26].”
MethodXu, W., Zou, X., Hu, D., & Qian, S. (2024) · International Conference on Electronic Engineering and Information Systems (EEISS) · doi:10.1109/eeiss62553.2024.00030
Ankita, Shambhavi, Pushp, S., & Shahnawazuddin, S. (2024) · International Conference on Signal Processing and Communications (SPCOM) · doi:10.1109/spcom60851.2024.10631619
“Front-end acoustic features extracted using FDLP analysis have been effectively used in recognition of reverberant speech [7] and spoofing attacks [8], [9].”
MethodContext Encoded Multi-Modal Attention Network for Detecting Audio Spoofing
Ranjan, R., Mayank Vatsa, A., & Singh, R. (2024) · IEEE International Joint Conference on Biometrics (IJCB) · doi:10.1109/ijcb62174.2024.10744479 · Cited by 5
“In the detection pipeline, most of the research is focused on finding distinguishing features that capture artificial audio artifacts within speech signals [7, 31, 37].”
BackgroundAnalysis of Pathological Features for Spoof Detection
Aung, M. A. A., Naing, H. M. S., Hlaing, A. M., Pa, W. P., Galajit, K., & Mawalim, C. O. (2024) · International Conference on Speech Database and Assessments · doi:10.1109/o-cocosda64382.2024.10800195
Yang, J., Wu, J., Wen, H., & Gao, Z. (2024) · Innovative Applications of AI · doi:10.70695/skknxc45
Improving Fairness in Synthetic Speech Detectors
Yadav, A. K. S., Bhagtani, K., Bestagini, P., & Delp, E. J. (2024) · Asilomar Conference on Signals, Systems and Computers (ACSSC) · doi:10.1109/ieeeconf60004.2024.10942853 · Cited by 3
“Existing work shows that out of different approaches for detecting synthetic speech [1], [11], [12], [20], [23], the approaches processing features obtained from a self-supervised transformer network [12], [23], [29] pre-trained on large speech dataset [26] showed best detection performance and…”
Wang, Y., Wang, X., Leow, C. S., Qishan, Z., Li, M., & Nishizaki, H. (2025) · IEICE Transactions on Information and Systems · doi:10.1587/transinf.2025edp7044 · Cited by 2
Future-Proofing Multilingual Fake Speech Detection
Demirörs, M., Özbayoğlu, A. M., & Akgün, T. (2025) · doi:10.5121/csit.2025.151309
Fraud at the Interface: Synthetic Voice Exploits in Blockchain User Interaction LayersPreprint
Kachan, V. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5374534
Robust prosody modeling for synthetic speech detection
Cohen, A., Shyrman, D., Solonskyi, A., Frenkel, R., Krishtul, A., & Gal, O. (2025) · Speech Communication · doi:10.1016/j.specom.2025.103283 · Cited by 1
Cross-lingual XLSR-Wav2Vec2-based Speech Spoofing Detection for Vietnamese Speech
Nguyen, T.-T.-H., & Do, T.-N.-D. (2025) · International Conference on Multimedia Analysis and Pattern Recognition (MAPR) · doi:10.1109/mapr67746.2025.11133998 · Cited by 2
“Eight classifiers using RFCC, MFCC [7], IMFCC, LFCC [8], SSFC [9], SCFC, SCMC [10], and CQCC [4] were evaluated.”
A Comparative Analysis of Advanced Deep Learning Techniques for Detecting Deepfake Audio
Arya, G., Tyagi, S. S., Nayak, T., & Sharma, T. (2025) · Lecture notes in networks and systems · doi:10.1007/978-981-96-7131-1_6
Associative-Discriminative Fusion Networks for Synthetic Speech Detection
Mai, K., Chen, C., Feng, Y., & Chen, D. (2025) · Circuits Systems and Signal Processing · doi:10.1007/s00034-025-03427-9
“In synthetic speech detection, a variety of features are frequently utilized, including constant Q cepstral coefficients (CQCC) [ 46 ], linear frequency cepstral coefficient (LFCC) [38] and constant Q transform (CQT) spectrogram [ 57 ], etc.”
Detecting Bangla DeepFake Audio: A Dual Approach Using LSTM and WaveNet
Ayan, N. S., Dipto, M. A., Faria, S. R., Mitu, S. A., & Masud, S. M. R. A. (2025) · Lecture notes in networks and systems · doi:10.1007/978-981-96-9196-8_38
Wolley, S., Jones, K. O., Robinson, C., Chandler-Crnigoj, S. L., McColl, S. M., & Burrell, H. (2025) · Lecture notes on data engineering and communications technologies · doi:10.1007/978-981-96-7749-8_4
Bekiryazıcı, Ş., Hanilçi, C., & Ozcan, N. (2025) · Gümüşhane Üniversitesi Fen Bilimleri Enstitüsü Dergisi · doi:10.17714/gumusfenbil.1695733
Self-citationWang, X., Tan, Z., & Li, G. (2025) · PLoS ONE · doi:10.1371/journal.pone.0335240
“Sahidullah et al. [6] comprehensively reviewed features commonly employed for synthetic speech detection, including short-term power spectrum features, short-term phase features, and spectral features derived from long-term transforms.”
Speech emotional content recognition based on the brain emotional learning model
Ravanbakhsh, M., Setayeshi, S., & Ravanbakhsh, M. (2026) · Multimedia Tools and Applications · doi:10.1007/s11042-026-21252-3
“In this paper, the values of MFCC are calculated using the functions provided in [ 42 ].”
Analysis of Acoustic Anomalies and Speech Artefacts in Synthetic Content
Jędrasiak, K. (2026) · Bezpieczeństwo i Technika Pożarnicza · doi:10.12845/sft.67.1.2026.3
LENS-DF: Deepfake Detection and Temporal Localization for Long-Form Noisy Speech
Liu, X., Ge, W., Wang, X., & Yamagishi, J. (2025) · IEEE International Joint Conference on Biometrics (IJCB) · doi:10.1109/ijcb65343.2025.11411544 · Cited by 7
“Early approaches primarily relied on statistical models, such as Gaussian mixture models [30], combined with various acoustic features [24, 29].”
End-to-End Modeling for Speech Spoofing and Deepfake Detection
Tak, H. (2023) · theses.fr (ABES) · doi:10.70675/aa9fb6cbz3f6az45bcz933fzf95d89893e59 · Cited by 4
Advances in Privacy Preservation and Fairness Assessment for Voice Biometrics
Chouchane, O. (2024) · theses.fr (ABES) · doi:10.70675/27834b1az6b2bz404fza44ez2ee7f9ce2ea7
Chaptoukaev, H. (2025) · theses.fr (ABES) · doi:10.70675/d6c484f4zea27z4fedzac4dzd1a94b0332a6 · Cited by 1
Joyner John Raj, J., Kishore, A., & Saranya, S. (2026) · International Conference on Visual Analytics and Data Visualization (ICVADV) · doi:10.1109/icvadv67766.2026.11469806
Dynamic Spectrogram Analysis with Local-Aware Graph Networks for Audio Anti-Spoofing
Li, Y., Chen, C., Chen, D., Zeng, N., & Zeng, K. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11464633
Reducing False Wake-up Activations in Voice Assistants using Deep Liveness Detection
Kamble, M. R., Viswanathan, R., & Krishna, K. S. R. (2026) · National Conference on Communications · doi:10.1109/ncc68160.2026.11479070
“A. Feature Extraction 1) MFCC: Mel Frequency Cepstral Coefficients (MFCC) features are widely used for speech recognition, whisper detection, acoustic scene classification, etc [13], [14].”
REALM: A Robust Evidence-Based Audio Logic Matrix for Forensic Deepfake Detection
Albaqami, H., Alahmadi, L., & AbdulRahman, S. (2026) · IEEE International Conference on Communication Systems and Network Technologies (CSNT) · doi:10.1109/csnt69054.2026.11502113
“Current detection methods can be grouped into two categories: traditional handcrafted spectral methods [5], [6] and end-to-end deep learning architectures [7], [8].”
AI-Based Deepfake Voice Detection Using MFCC Features and Random Forest Classification
Kangane, A. S., Bhople, A. S., & Parulekar, W. R. (2026) · International Journal of Latest Technology in Engineering Management & Applied Science · doi:10.51583/ijltemas.2026.150500121
Interpreting SSL Representations for Spoof Detection: A WavLM Study
Mallat, M., Panariello, M., Todisco, M., Evans, N., & Larcher, A. (2026) · International Workshop on Biometrics and Forensics (IWBF) · doi:10.1109/iwbf68042.2026.11558152
“• Spectral (6): MFCC mean, spectral centroid, spectral roll-off, ZCR, spectral entropy, spectral flux, general signal descriptors used in CM [21].”
Verma, B. K., Giri, S., & Saxena, A. (2026) · International Conference on Innovations in Computational Intelligence (ICICI) · doi:10.1109/icici68867.2026.11564891
“These techniques did not generalise well to different datasets with changing synthesis approaches [5].”
The Generalization Gap: Do Audio Deepfake Detectors Actually Protect Against Modern Vishing?
Martínez-Echevarría, V. G., Palacios, R., López, G., & Gupta, A. D. (2026) · Electronics · doi:10.3390/electronics15132846 · Cited by 1
Exploring KAN Network with Multiple Basis Functions for Audio Anti-Spoofing
Badugu, V. K., Veesa, S., & Singh, M. (2026) · IEEE International Conference on Signals and Systems (ICSigSys) · doi:10.1109/icsigsys71160.2026.11613909
“Initial spoofing countermeasure systems relied on hand-crafted acoustic features such as CQCC [12] and LFCC [13] combined with GMM-based classifiers, and these approaches achieved strong performance on the ASVspoof 2015 dataset.”
Learning Nonlinear Acoustic Patterns with Residual KANs for ASV Anti-Spoofing System
Badugu, V. K., Veesa, S., & Singh, M. (2026) · IEEE International Conference on Signals and Systems (ICSigSys) · doi:10.1109/icsigsys71160.2026.11613907
“The proposed RKAN architecture out-performs the 3 baseline systems given in ASVspoof 2021 Challenge [10] i.e., LFCC-GMM [17], CQCC-GMM [18] and RawNet2 [16], shown in Table V.”
Cross-domain Fine-Tuned Wav2Vec 2.0 model with Bidirectional GRU for Audio Deepfake Detection
Sharma, B. B., Selwal, A., & Gupta, T. (2026) · International Conference on Electronics, Computers and Artificial Intelligence (ECAI) · doi:10.1109/ecai69016.2026.11613791
Daniyaruly, M., Smaiyl, A., & Zholmyrza, A. (2026) · International Conference on Cybersecurity, Digital Forensics, and AI Applications (ICCSDFAI) · doi:10.1109/iccsdfai70505.2026.11647876
“Sahidullah et al. [4] conducted a systematic comparison of spectral features for synthetic speech detection, finding that cepstral features performed adequately on matched-condition data but degraded sharply under channel mismatch — an early empirical hint at the domain sensitivity problem this…”
Detecting Voice Conversion Attacks Using Self-Supervised, Handcrafted, and Hybrid Audio Features
Nasr, N. A., Almutairi, A., Attia, M., Daoud, I., Alfares, Y., elsayed, H., Khan, M. U., & Koubâa, A. (2026) · IEEE Annual International Computer Software and Applications Conference (COMPSAC) · doi:10.1109/compsac69091.2026.00099
“For classification, the mean and standard deviation of each coefficient are combined, producing a 120-dimensional vector of features [6] .”
Signal Singularity Analysis for Robust Deep Audio Splicing Detection and Localization
Li, Y., Yang, Y., Li, Z., Liang, S., & Cao, Y. (2026) · International Conference on Cloud Computing and Big Data Analytics (ICCCBDA) · doi:10.1109/icccbda69573.2026.11676577
“LFCCs are therefore concatenated to provide complementary spectral cues [5, 7].”
Li, Y., Fang, Y., Xue, J., Milling, M., & Schuller, B. W. (2026) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2026.1928592
“These features could reveal subtle spectral artifacts introduced by synthetic voice generation (Sahidullah et al., 2015) or standard speech features (Hamza et al., 2022).”
Miroshnikov, N., Токарева, У., & Shulzhenko, A. D. (2026) · Informatics and Automation · doi:10.15622/ia.25.5.6
Beyond Known Speakers: Enhancing Spoofing Detection Generalization with VIVED
Wang, W., Zi, X., Prasad, M., Lipman, J., & Jha, S. K. (2026) · ACM Workshop on Deepfake, Deception, and Disinformation Security (3D-Sec) · doi:10.1145/3843205.3845632
Metadata
- Authors
- Sahidullah, M., Kinnunen, T., & Hanilçi, C.
- Affiliations
- University of Eastern Finland (FI)
- Venue
- Interspeech 2015
- Publisher
- ISCA
- Published
- 2015-09-06
- Pages
- 2087-2091
- ISSN
- 2308-457X (ISSN-L)
- OpenAlex
- W2295634712
- Cited by
- 382
- References
- 36
- Citations by year
- 2026: 23 · 2025: 41 · 2024: 42 · 2023: 50 · 2022: 38 · 2021: 51 · 2020: 35 · 2019: 41 · 2018: 15 · 2017: 29 · 2016: 15 · 2015: 2
- FWCI
- 11.46
- Type
- conference paper
- Language
- en
- Open access
- green
- erepo.uef.fi
- Licence
- other-oa
- Locations
- Interspeech (USB) (publishedVersion)
- UEF eRepo (University of Eastern Finland) (submittedVersion, other-oa)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Speech Recognition and Synthesis
- Physical Sciences › Computer Science › Signal Processing › Speech and Audio Processing
- Physical Sciences › Computer Science › Signal Processing › Biometric Identification and Security
- Keywords
- synthetic speech detection, spoofing attacks, anti-spoofing, speaker verification, voice conversion, speech synthesis, feature extraction
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.