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 201–300 of 383
Speech Audio Deepfake Detection via Convolutional Neural Networks
Valente, L. P., de Souza, M. M. S., & da Rocha, A. M. (2024) · IEEE Conference on Evolving and Adaptive Intelligent Systems (EAIS) · doi:10.1109/eais58494.2024.10569111 · Cited by 13
SpecViT: A Custom Vision-Transformer based Approach for Audio Deepfake Detection
Modak, S., Das, A. K., & Naskar, R. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10889022 · Cited by 12
End-to-end spoofing speech detection and knowledge distillation under noisy conditions
Liu, P., Zhang, Z., & Yang, Y. (2021) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn52387.2021.9534312 · Cited by 8
Voice Spoofing Countermeasure Based on Spectral Features to Detect Synthetic Attacks Through LSTM
Qadir, G., Zareen, S., Hassan, F., & Rahman, A. U. (2022) · International Journal of Innovations in Science and Technology · doi:10.33411/ijist/2021030512 · Cited by 12
How to Boost Anti-Spoofing with X-Vectors
Ma, X., Zhang, S., Huang, S., Gao, J., Hu, Y., & He, L. J. (2023) · IEEE Spoken Language Technology Workshop (SLT) · doi:10.1109/slt54892.2023.10022504 · Cited by 7
“Two proposed methods, one is to integrate transformed segmental x-vectors and then concatenate with linear frequency cesptral coeffcient (LFCC) [14] features as the input of anti-spoofing system, another is to splice x-vector with the output of one embedding layer in anti-spoofing system.”
MethodDifferential convolutional network for noise mask estimation
Dişken, G. (2023) · Applied Acoustics · doi:10.1016/j.apacoust.2023.109568 · Cited by 8
Tackling Spoofing-Aware Speaker Verification with Multi-Model Fusion
Wu, H., Kang, J., Meng, L., Zhang, Y., Wu, X., Wu, Z., Lee, H.-Y., & Meng, H. M. L. (2022) · The Speaker and Language Recognition Workshop (Odyssey) · doi:10.21437/odyssey.2022-13 · Cited by 7
Synthesized Speech Detection Based on Spectrogram and Convolutional Neural Networks
Nosek, T., Suzić, S., Papic, B., & Jakovljevic, N. M. (2019) · Telecommunications Forum (TELFOR) · doi:10.1109/telfor48224.2019.8971215 · Cited by 11
“First system uses linear frequency cepstral coefficients [18] while the second one uses constant Q cepstral coefficients [19].”
MethodMankad, S. H., & Garg, S. (2020) · Progress in Artificial Intelligence · doi:10.1007/s13748-020-00216-0 · Cited by 7
Compact Time-Domain Representation for Logical Access Spoofed Audio
Karo, M., Yeredor, A., & Lapidot, I. (2024) · IEEE/ACM Transactions on Audio, Speech, and Language Processing · doi:10.1109/taslp.2023.3341000 · Cited by 6
“Motivated by the feature extraction process of cepstral co-efficients [21], such as Mel Frequency Cepstral Coefficients (MFCC), Linear Frequency Cepstral Coefficients (LFCC), and Constant-Q Cepstral Coefficients (CQCC) [22], [23], different filterbanks were applied to the audio samples.”
MethodMulti Pattern Features-Based Spoofing Detection Mechanism Using One Class Learning
Üstübioğlu, B., Tahaoğlu, G., Üstübıoğlu, A., Ulutaş, G., Amerini, I., & Kılıç, M. (2024) · IEEE Access · doi:10.1109/access.2024.3447572 · Cited by 7
“The comparison with the state-of-the-arts with the four baseline systems CQCC-GMM [4], LFCC-GMM [42], LFCC-LCNN [43], RawNet2 [44] and UR-AIR [45].”
RawTFNet: A Lightweight CNN Architecture for Speech Anti-Spoofing
Yang, X. X., Dang, T., & Das, R. K. (2025) · Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) · doi:10.1109/apsipaasc65261.2025.11249383 · Cited by 12
“Feature-based methods use pre-designed representations such as linear-frequency cepstral coefficients (LFCC) [7], cochlear filter cepstral coefficient and instantaneous frequency [8] and constant-Q transform (CQT) based representations [9], [10].”
Replay Attack Detection for Speaker Verification Using Different Features Level Fusion System
Jana, S., Yashwanth, V. S., Dheeraj, K. V. N., Balaji, S., Bharath, K. P., & Rajesh Kumar, M. (2021) · Innovations in Power and Advanced Computing Technologies (i-PACT) · doi:10.1109/i-pact52855.2021.9696686 · Cited by 7
Audio Deepfakes: Feature Extraction and Model Evaluation for Detection
Bhukya, R. K., Raj, A., & Raja, D. N. (2024) · International Conference for Emerging Technology (INCET) · doi:10.1109/incet61516.2024.10593405 · Cited by 7
“Subsequently, ASVspoof2019 [19] emphasized countermeasures, specifically in ASV and spoofed audio detection (SAD) tasks.”
BackgroundDiffSSD: A Diffusion-Based Dataset For Speech Forensics
Bhagtani, K., Yadav, A. K. S., Bestagini, P., & Delp, E. J. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10889450 · Cited by 18
“The inputs to synthetic speech detection methods are processed either by Gaussian Mixture Model (GMM) [1], [37], neural networks such as ResNet [27] or transformer networks [26], [38].”
Use of Claimed Speaker Models for Replay Detection
Suthokumar, G., Sriskandaraja, K., Sethu, V., Wijenayake, C. A., Ambikairajah, E., & Li, H. (2018) · Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) · doi:10.23919/apsipa.2018.8659510 · Cited by 7
“This feature set has been shown to be effective in spoofing detection for not only replay attacks but also for speech synthesis and voice conversion attacks [17], [28].”
BackgroundA Comparison of Features for Replay Attack Detection
Xie, Z., Zhang, W., Chen, Z., & Xu, X. (2019) · Journal of Physics Conference Series · doi:10.1088/1742-6596/1229/1/012079 · Cited by 6
“In terms of the classifiers, simple classifiers such as Gaussian mixture models (GMMs) are commonly used [7, 14].”
BackgroundInfluentialASLNet: An Encoder-Decoder Architecture for Audio Splicing Detection and Localization
Zhang, Z., Zhao, X., & Yi, X. (2022) · Security and Communication Networks · doi:10.1155/2022/8241298 · Cited by 13
“It has been shown choosing an appropriate acoustic feature for fake speech detection is important [26], so the acoustic features need to be selected for ASLNet.”
BackgroundExploration of Bonafide and Spoofed Audio Classification Using Machine Learning Models
Babu, N., Kumar, P., Mathew, J., & Satija, U. (2022) · IEEE India Conference · doi:10.1109/indicon56171.2022.10039908 · Cited by 6
Audio Stream Analysis for Deep Fake Threat Identification
Jędrasiak, K. (2024) · Civitas et Lex · doi:10.31648/cetl.9684 · Cited by 6
Voice liveness detection under feature fusion and cross-environment scenario
Garg, S., & Mankad, S. H. (2020) · Multimedia Tools and Applications · doi:10.1007/s11042-020-09281-y · Cited by 5
Capsule Network based End-to-end System for Detection of Replay Attacks
Ouyang, M., Das, R. K., Yang, J., & Li, H. (2021) · International Symposium on Chinese Spoken Language Processing (ISCSLP) · doi:10.1109/iscslp49672.2021.9362111 · Cited by 8
“org/ cepstral coefficients (LFCC), subband spectral flux coefficients and spectral centroid frequency coefficients [14].”
BackgroundClassification of Speech with and without Face Mask using Acoustic FeaturesPreprint
Das, R. K., & Li, H. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2010.03907 · Cited by 6
“However, LFCC and CQCC emerge as strong acoustic feature based single systems that suggest the importance of linear filterbanks and long-term information.”
BackgroundMethodInfluentialWu, Y., Hua, H., Li, Z., & Zhang, S. (2025) · Electronics · doi:10.3390/electronics14122456 · Cited by 12
Mishra, J., Chhibber, M., Shim, H.-J., & Kinnunen, T. (2025) · Computer Speech & Language · doi:10.1016/j.csl.2025.101840 · Cited by 20
“As for the detection solutions that have emerged through the past decade, the research focus has shifted from hand-crafting features [17, 18] to design of deep neural network architectures [19, 20].”
Self-citationCBC-Based Synthetic Speech Detection
Yang, J., He, Q., Hu, Y., & Pan, W. (2019) · International Journal of Digital Crime and Forensics · doi:10.4018/ijdcf.2019040105 · Cited by 5
Deepfake Detection System for the ADD Challenge Track 3.2 Based on Score Fusion
Zhang, Y., Lu, J., Wang, X., Li, Z., Xiao, R., Wang, W., Li, M., & Zhang, P. (2022) · doi:10.1145/3552466.3556528 · Cited by 6
“The input features of CNN based classifiers are two-dimensional features, including short-time Fourier transform (STFT) spectrograms, constant-Q transform (CQT) spectrograms, mel-scale transforms (MSTFT), and cepstral features such as constant-Q cepstral coefficients (CQCC) [43] and Linear Frequency Cepstral Coefficie…”
Attentional Fusion TDNN for Spoof Speech Detection
Wu, L., & Jiang, Y. (2022) · International Conference on Pattern Recognition and Artificial Intelligence (PRAI) · doi:10.1109/prai55851.2022.9904136 · Cited by 8
“CQCC [4] to train a model for extracting feature embedding of speech, then use the information to make judgements, such as SVM or GMM [5], [6].”
MethodFeature Extraction Techniques for Deep Learning based Speech Classification
Chakravarty, S., Khandelwal, R., & Dhote, K. M. (2023) · International Conference on Computing, Communication and Networking Technologies (ICCCNT) · doi:10.1109/icccnt56998.2023.10307237 · Cited by 6
“Noteworthy algorithms like PLP, Viterbi search, deep neural networks, and WFST outline [3] are commonly employed in the synthesis of natural-sounding speech.”
MethodSarkar, A. K., Sahidullah, M., Tan, Z., & Kinnunen, T. (2017) · Interspeech (USB) · doi:10.21437/interspeech.2017-1758 · Cited by 7
“We use a two-class GMM-based method [18, 19] using two separate GMMs [20] trained independently with real human and spoofed data.”
MethodSelf-citationHitting Three Birds with One System: A Voice-Based CAPTCHA for the Modern User
Shah, M. A., & Harras, K. A. (2018) · IEEE International Conference on Web Services (ICWS) · doi:10.1109/icws.2018.00040 · Cited by 4
“acoustic modeling and analysis for a variety of different applications and have recently been shown to perform quite well at discriminating between natural and forged speech[12], [33], [29].”
BackgroundVoice Liveness Detection for Medical Devices
Hao, B., & Hei, X. (2018) · Advances in medical technologies and clinical practice book series · doi:10.4018/978-1-5225-7525-2.ch005 · Cited by 8
Overview of methods for extracting acoustic speech features in speaker recognition
Sudjenkova, A. (2020) · Transaction of Scientific Papers of the Novosibirsk State Technical University · doi:10.17212/2307-6879-2019-3-4-139-164 · Cited by 4
“Идея о том, что при такой фильтрации может быть упущена значимая информация, воплощена в комплементарном к MFCC методе – обратных мел-частотных кепстральных коэффициентах (IMFCC – inverted mel-frequency cepstral coefficients) [11].”
InfluentialSpeech spoofing detection using SVM and ELM technique with acoustic features
Rahmeni, R., Aicha, A. B., & BenAyed, Y. (2020) · International Conference on Advanced Technologies for Signal and Image Processing · doi:10.1109/atsip49331.2020.9231799 · Cited by 8
“ASV systems need to find out the enrolled speaker’s identity and compare it against other models [2] [3].”
BackgroundKumar, S. R., & Bharathi, B. (2022) · Circuits Systems and Signal Processing · doi:10.1007/s00034-022-01957-0 · Cited by 5
Altuwayjiri, S. M., Bchir, O., & Ismail, M. M. B. (2022) · International Journal of Advanced Computer Science and Applications · doi:10.14569/ijacsa.2022.0130596 · Cited by 6
“For this purpose, the authors in [42] conducted an empirical comparison of 19 different features to determine the most appropriate one for classifying spoofing versus genuine records.”
BackgroundMethodInfluentialDefending Your Voice: Adversarial Attack on Voice ConversionPreprint
Huang, C., Lin, Y. Y., Lee, H.-Y., & Lee, L.-S. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2005.08781 · Cited by 5
“Detecting any of such artifacts or defending against such activities is thus increasingly important [6, 7, 8, 9], which applies equally to voice conversion.”
BackgroundSpeech Emotion Recognition Using Magnitude and Phase Features
Shankar, D. R., Manjula, R. B., & Biradar, R. C. (2024) · SN Computer Science · doi:10.1007/s42979-024-02833-1 · Cited by 4
“In this work, we have used standard CQCC parameters [27], such as, 96 bin/octave, 16 constant samples in the first octave to 256.”
A Frequency Bin Analysis of Distinctive Ranges Between Human and Deepfake Generated Voices
Maltby, H., Wall, J. A., Glackin, C., Moniri, M., Cannings, N., & Salami, I. (2024) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn60899.2024.10650554 · Cited by 8
“…of features specifically accentuating high-frequency discriminative patterns and audio artefacts [8] [9], with Inverse Mel Frequency Cepstral Coefficients (IMFCC) showing promising performance for unseen generative algorithms [10] as it accentuates the higher frequencies instead of the lower ones.”
Modified gammatone frequency cepstral coefficients to improve spoofing detection
Das, K. A., George, K. K., Kumar, C. S., Veni, S., & Panda, A. (2016) · International Conference on Advances in Computing, Communications and Informatics (ICACCI) · doi:10.1109/icacci.2016.7732024 · Cited by 7
“In this work, we focus only on the detection of voice converted and synthesized speech.”
Speaker Embedding Extraction with Multi-feature Integration Structure
Li, Z., Lu, H., Zhou, J., Li, L., & Hong, Q. (2019) · Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) · doi:10.1109/apsipaasc47483.2019.9023103 · Cited by 5
“The ASVspoof2019 officially [21] revealed the baseline systems with two kinds of features (LFCC [22] and CQCC [23]) and two baseline systems shared a same GMM classifier.”
BackgroundConstant‐Q magnitude–phase coefficients extraction for synthetic speech detection
Yang, J., Lin, P., & He, Q. (2020) · IET Biometrics · doi:10.1049/iet-bmt.2018.5100 · Cited by 5
“Sahidullah et al. [16] used inverted MFCCs. Paul et al. [17] proposed several features based on the power spectrum for synthetic speech detection, for example, speech-signal frequency cepstral coefficients and mel-warped overlapped block transformation.”
BackgroundInfluentialExploiting Magnitude and Phase Aware Deep Neural Network for Replay Attack Detection
Phapatanaburi, K., Buayai, P., Naktong, W., & Srinonchat, J. (2020) · ECTI Transactions on Electrical Engineering Electronics and Communications · doi:10.37936/ecti-eec.2020182.240341 · Cited by 4
“In this paper, MFCC is utilized as a magnitude-based feature for DNN input for the researcher to use the information and a reliable means of detecting replay attack [11].”
MethodBPCNN: Bi-Point Input for Convolutional Neural Networks in Speaker Spoofing Detection
Yoon, S.-H., & Yu, H.-J. (2022) · Sensors · doi:10.3390/s22124483 · Cited by 6
“The score of each utterance is computed by averaging the scores of all the segments in the utterance.”
MethodImproved One-class Learning for Voice Spoofing Detection
Li, L., Xue, X., Peng, H., Ren, Y., & Zhao, M. (2023) · Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) · doi:10.1109/apsipaasc58517.2023.10317117 · Cited by 5
“Well-designed acoustic features have achieved good performance, such as Linear Frequency Cepstral Coefficients (LFCC) [6], constant Q transform (CQT) [7], etc. Gaussian Mixture Model (GMM) [8]–[10] and deep neural network-based classifiers [11]–[14] are the mainstream back-end classifiers.”
BackgroundEnhancing Synthesized Speech Detection with Dual Attention Using Features Fusion
Wang, B., Ma, Y., Tang, Y., Wang, R., & Zhang, M. (2023) · doi:10.1109/ccat59108.2023.00026 · Cited by 5
“%) t-DCF CQCC-GMM [24] 15.62 0.4974 LFCC-GMM [25] 19.3 0.5758 LFCC-LCNN [26] 9.26 0.3445 RawNet2 [11] 9.5 0.4257 LFCC-ECAPA-TDNN [27] 5.46 0.3094 SSL-AASIST [16] 4.48 0.3094 Our 4.12 0.3008 is 0.73% lower than inputting them into the Resnet network, demonstrating the effectiveness of self-attention.”
InfluentialContrastive Learning-Based Speech Spoofing Detection for Multimedia Security in Edge Intelligence
Sun, J., Deng, X., Liu, S., Fan, X., Huang, Y., He, Y., Wu, C., & Park, J. H. (2024) · ACM Transactions on Multimedia Computing Communications and Applications · doi:10.1145/3698773 · Cited by 6
“However, these methods come with high computational complexity, rendering them unsuitable for speech spoofing detections with high real-time requirements.”
Feature Fusion and Voiceprint-Based Access Control for Wireless Insulin Pump Systems
Ping, Y., Hao, B., Hei, X., Tu, Y., Du, X., & Wu, J. (2019) · IEEE Access · doi:10.1109/access.2019.2937805 · Cited by 5
“The original implementations of RFCC, SCFC, SCMC, and SCMC only use linear scale [27] and the RPS (only) uses MEL, we extend the implementations in [27], [39] and apply linear scale, MEL and inverted IMEL filter banks to these features and form”
BackgroundMethodInfluentialInvestigating Feature Reduction Strategies for Replay Antispoofing in Voice Biometrics
Mankad, S. H., Garg, S., Patel, M., & Adalja, H. (2019) · Lecture notes in computer science · doi:10.1007/978-3-030-34872-4_44 · Cited by 4
Improving replay attack detection by combination of spatial and spectral features
Yaguchi, R., Shiota, S., Ono, N., & Kiya, H. (2019) · Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) · doi:10.1109/apsipaasc47483.2019.9023181 · Cited by 4
“The GMM of each system is trained with acoustic features, CQCC [12] and LFCC [13], respectively.”
MethodTowards robust audio spoofing detection: a detailed comparison of traditional and learned featuresPreprint
Balamurali, B. T., Lin, K. W. E., Lui, S., Chen, J., & Herremans, D. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1905.12439 · Cited by 4
“The procedure for calculating LFCCs is similar to that of RFCCs [41], however, it uses triangular filters instead of rectangular filters [19], [42].”
BackgroundMethodInfluentialContreras, R. C., Viana, M. S., Fonseca, E. S., Bongarti, M. A. D. S., Toygar, Ö., & Guido, R. C. (2025) · Computers & Electrical Engineering · doi:10.1016/j.compeleceng.2025.110570 · Cited by 3
Convolutional kolmogorov-arnold network(CKAN) based spoofing countermeasure System(SCMS)
Badugu, V. K., & Singh, M. (2026) · International Journal of Speech Technology · doi:10.1007/s10772-026-10258-9 · Cited by 3
“while we have obtained better results in comparision with three baseline systems (Todisco et al., 2017 ; Sahidullah et al., 2015 ; Tak et al., 2021 ) with proposed architecture with Convolutional Kolmogorov Arnold Network uses the log Mel spectrogram provides a compact, perceptually relevant…”
Wang, X., Yamagishi, J., Todisco, M., Delgado, H., Nautsch, A., Evans, N., Sahidullah, M., Vestman, V., Kinnunen, T., Lee, K. A., et al. (2019) · Graduate School and Research Center in Digital Science (EURECOM) · Cited by 6
Self-citationTelesonar: Robocall Alarm System by Detecting Echo Channel and Breath Timing
Yan, Z., Tan, R., Song, Q., & Lu, C. X. (2022) · ACM Conference on Embedded Networked Sensor Systems (SenSys) · doi:10.1145/3560905.3568500 · Cited by 5
“DNN-based approaches [22, 27, 35, 36] have been proposed to verify the voice genuity.”
MethodContreras, R. C., Viana, M. S., & Guido, R. C. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-42508-0_27 · Cited by 3
From Voices to Beats: Enhancing Music Deepfake Detection by Identifying Forgeries in Background
Wei, Z., Ye, D., Deng, J., & Lin, Y. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10890293 · Cited by 11
“Specifically, we evaluated performance using various feature extraction techniques, including LFCC [18], Rawnet, Wav2Vec2, and our proposed Hybrid Frontend.”
BIM-Based Adversarial Attacks Against Speech Deepfake Detectors
Wang, W. E., Salvi, D., Negroni, V., Leonzio, D. U., Bestagini, P., & Tubaro, S. (2025) · Electronics · doi:10.3390/electronics14152967 · Cited by 9
On the Use of Fisher Vector Encoding for Voice Spoofing Detection
Alam, J. (2019) · International Conference on Ubiquitous Computing and Ambient Intelligence (UCAmI) · doi:10.3390/proceedings2019031037 · Cited by 3
“The most effective countermeasures for spoofing detection are the CQT-based cepstral coefficients (CQCC) [10], linear- frequency cepstral coefficients (LFCC), product spectrum-based cepstral coefficients (PSCC) [11–13], all pole group delay cepstral coefficients (APGDC) [13,22], linear prediction residual cepstral coe…”
BackgroundDetecting Replay Attacks Using Single-Channel Audio: The Temporal Autocorrelation of Speech
Lee, S.-K., Tsao, Y., & Wang, H. (2022) · Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) · doi:10.23919/apsipaasc55919.2022.9980046 · Cited by 4
Robust Audio Anti-Spoofing System Based on Low-Frequency Sub-Band Information
Li, M., & Zhang, X.-P. S. (2023) · IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA) · doi:10.1109/waspaa58266.2023.10248132 · Cited by 8
“In the recent literature, the hand-crafted acoustic features, such as Mel-frequency Cepstral Coefficients (MFCC) [6, 7], Constant Q Cepstral Coefficients (CQCC) [8, 9], have been widely used in the Deepfake detectors to distinguish the spoofing audio based on the speech quality and similarity [10].”
MethodSpeech Spoofing Detection Based on Graph Attention Networks with Spectral and Temporal Information
Zhang, P., Chen, Y., Li, M., Zhao, H., Zhang, J., Wang, F., & Wu, X. (2023) · ACM International Conference on Multimedia in Asia (MMAsia) · doi:10.1145/3595916.3626406 · Cited by 5
“Traditional hand-designed features such as MFCC, LPC, and LPCC [25] are widely used in speech spoofing detection, but they may not adequately capture subtle differences in speech signals [4].”
SV-DeiT: Speaker Verification with DeiTCap Spoofing Detection
Ranjan, R., Mayank Vatsa, A., & Singh, R. (2023) · IEEE International Joint Conference on Biometrics (IJCB) · doi:10.1109/ijcb57857.2023.10449121 · Cited by 9
“Sahidullah et al. [23] compared the acoustics features that are used by spoof detection systems.”
BackgroundMOS-FAD: Improving Fake Audio Detection Via Automatic Mean Opinion Score Prediction
Zhou, W., Yang, Z., Chu, C., Li, S., Dabre, R., Zhao, Y., & Kawahara, T. (2024) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp48485.2024.10446041 · Cited by 12
“Research has shown that well-designed acoustic features can effectively distinguish fake audio from real audio [12, 13, 14, 15].”
BackgroundWeighted I-Vector Based Text-Independent Speaker Verification System
Mohammadi, M., & Mohammadi, H. R. S. (2019) · Iranian Conference on Electrical Engineering · doi:10.1109/iraniancee.2019.8786420 · Cited by 3
Using Multi-Resolution Feature Maps with Convolutional Neural Networks for Anti-Spoofing in ASV
Wang, Q., Lee, K. A., & Koshinaka, T. (2020) · The Speaker and Language Recognition Workshop (Odyssey) · doi:10.21437/odyssey.2020-20 · Cited by 4
“Recently, the use of high time-frequency resolution features has become a popular approach [10–12].”
BackgroundVoiceprint-Based Access Control for Wireless Insulin Pump Systems
Hao, B., Hei, X., Tu, Y., Du, X., & Wu, J. (2018) · IEEE International Conference on Mobile Ad-Hoc and Sensor Systems (MASS) · doi:10.1109/mass.2018.00046 · Cited by 3
“We focused on short-term power spectrum features (except CQCC) and evaluated 9 different features [20], [21]: Mel Frequency Cepstral Coefficients (MFCC), Inverted MFCC (IMFCC), Linear Frequency”
MethodTowards Development of Smart and Reliable Voice Based Personal Assistants
Mankad, S. H., Shah, V., & Garg, S. (2018) · IEEE Region 10 Conference (TENCON) · doi:10.1109/tencon.2018.8650091 · Cited by 3
“We used cepstral features with filterbanks of two different shapes [14]: (i) Rectangular filterbanks (ii) Triangular filterbanks.”
MethodEffects of Waveform PMF on Anti-Spoofing Detection
Lapidot, I., & Bonastre, J.-F. (2019) · Interspeech (USB) · doi:10.21437/interspeech.2019-2607 · Cited by 10
“Different features were proposed for anti-spoofing systems [9].”
BackgroundAttention Network with GMM Based Feature for ASV Spoofing Detection
Lei, Z., Yu, H., Yang, Y., & Ma, M. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-86608-2_50 · Cited by 3
On the contribution of the voice texture for speech spoofing detection
Rahmeni, R., Aicha, A. B., & BenAyed, Y. (2019) · International Conference on Sciences and Techniques of Automatic Control and Computer Engineering · doi:10.1109/sta.2019.8717297 · Cited by 3
“These features are used in training phase in order to develop a model able to distinguish between genuine and spoofed speech [10]– [12].”
MethodAn Investigation about the Scalability of the Spoofing Detection System
Suthokumar, G., Sriskandaraja, K., Sethu, V., Wijenayake, C. A., & Ambikairajah, E. (2018) · International Conference on Information and Automation for Sustainability (ICIAfS) · doi:10.1109/iciafs.2018.8913369 · Cited by 6
“Linear frequency cepstral coefficients (LFCC) [16], IMFCC [16], CQCC [17] and SCC[18] features have been shown to be effective front-ends for synthetic speech detection.”
BackgroundInfluentialAdversarial Post-Processing of Voice Conversion against Spoofing Detection
Ding, Y., Zhang, J.-X., Liu, L., Jiang, Y., Hu, Y., & Ling, Z.-H. (2020) · Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) · Cited by 5
End-to-end anti-spoofing with RawNet2Preprint
Tak, H., Patino, J., Todisco, M., Nautsch, A., Evans, N., & Larcher, A. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2011.01108 · Cited by 3
“These choices are motivated by the success of linear and inverseMel [21] front-ends for anti-spoofing.”
BackgroundOne-class learning towards generalized voice spoofing detection.Preprint
Zhang, Y., Jiang, F., & Duan, Z. (2020) · arXiv (Cornell University) · Cited by 3
The Interpretable and Transferable Adversarial Attack against Synthetic Speech Detectors
Deng, J., Ye, D., Li, J., Liu, Z., Tang, L., & Zhang, Y. (2025) · ACM Transactions on Multimedia Computing Communications and Applications · doi:10.1145/3727341 · Cited by 2
“To detect this kind of attack, researchers have proposed a variety of countermeasures [15, 25, 28].”
Ge, Z., Xu, X., Guo, H., & Schuller, B. W. (2025) · IEEE Transactions on Audio Speech and Language Processing · doi:10.1109/taslpro.2025.3581019 · Cited by 2
“…aims to investigate feasible handcrafted acoustic features for ADD, including relative phase shift [28], [29], Cochlear Filter Cepstral Coefficients (CFCCs) [30], instantaneous frequency [30], [31], Linear Frequency Cepstral Coefficients (LFCCs) [32], and Constant-Q Cepstral Coefficients (CQCC) [33].”
RPRA-ADD: Forgery Trace Enhancement-Driven Audio Deepfake Detection
Fu, R., Wang, X., Wen, Z., Tao, J., Xie, Y., Wang, Z., Qiang, C., Liu, X., Fan, C., Li, C., et al. (2025) · IEEE Transactions on Audio Speech and Language Processing · doi:10.1109/taslpro.2025.3646051 · Cited by 5
“Tosignificantlyaccentuatethefeaturespacedistributiondivergencesbetweenrealandfake Tosignificantlyaccentuatethefeaturespacedistributiondivergencesbetweenrealandfake audio,theMulti-stageDispersedEnhancementLoss(MDEL)isdesigned,whichemploysadispersalstrategyinmulti-stagefea-turespaces.”
Kasapoglu, B., & Koç, T. (2019) · National Conference on Electrical and Electronics Engineering (ELECO) · doi:10.23919/eleco47770.2019.8990465 · Cited by 2
“) [7, 8] and phase spectrum based features, like group delay (GD) [9, 10].”
BackgroundSpoofed Speech Detection with Weighted Phase Features and Convolutional Networks
Dişken, G. (2022) · Archives of Acoustics · doi:10.24425/aoa.2022.141648 · Cited by 2
“Besides the MFCC and CQCC features, many other magnitude-based features can be found in the literature for spoof detection ( Font et al ., 2017; Sahidullah et al ., 2015).”
BackgroundMethodSynthetic Speech Spoofing Detection Based on Online Hard Example Mining
Hu, C., Zhou, R., & Yuan, Q. (2023) · IEEE Access · doi:10.1109/access.2023.3311849 · Cited by 4
“Conventional front-end approaches employ digital signal processing algorithms such as linear frequency cepstral coefficients (LFCC) [8], Constant-Q transform (CQT), and Constant Q Cepstral Coefficients (CQCC) [9] to extract acoustic features.”
MethodPatel, U., Bhilare, S., & Hati, A. (2024) · Pattern Analysis and Applications · doi:10.1007/s10044-024-01269-w · Cited by 3
“…speaker identification by Highly imperceptible adversarial Disturbances (FoolHD) [4] attack methods for generating adversarial examples, taking (1) [39] Linear scale Triangular Cepstral High CQCC [40] Constant Q transform Linear Cepstral High, Low LPCC [41] Linear predictive coding Linear…”
Chaudhari, A., Shedge, D., Bairagi, V. K., & Nanthaamornphong, A. (2024) · Symmetry · doi:10.3390/sym16070788 · Cited by 2
“This work also evaluated IMFCC, SCMC, SCFC, RFCC [3] and APGDF [28] based features.”
BackgroundMethodLeveraging distance information for generalized spoofing speech detection
Lu, J., Zhang, Y., Li, Z., Shang, Z., Wang, W., & Zhang, P. (2025) · Computer Speech & Language · doi:10.1016/j.csl.2025.101804 · Cited by 3
Leveraging Neural Vocoder Artifacts for Improved Synthetic Speech Detection
Xue, J., Cui, S., & Zhang, W. (2024) · International Conference on Sensors, Signal and Image Processing (SSIP) · doi:10.1145/3725949.3725971 · Cited by 2
“Since TTS and VC systems rely on vocoders for speech synthesis, early research centered on [17] identifying acoustic differences between vocoder-generated and natural speech.”
Presentation attack detection in voice biometrics
Korshunov, P., & Marcel, S. (2017) · Institution of Engineering and Technology eBooks · doi:10.1049/pbse004e_ch10 · Cited by 2
SHNU Anti-spoofing Systems for ASVspoof 2019 Challenge
Feng, Z., Tong, Q., Long, Y., Wei, S., Yang, C., & Zhang, Q. (2019) · Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) · doi:10.1109/apsipaasc47483.2019.9023319 · Cited by 3
“To detect the synthetic and voice converted speech, several new features have been proposed and proved to be very effective, such as the Constant Q Cepstral Coefficients (CQCC) [9], the Inverted Mel Frequency Cepstral Coefficients (IMFCC) features [10], the high-dimensional magnitude and phase features [11], the spoof…”
MethodStudy of speech features robustness for speaker verification application in noisy environments
Mohammadi, M., & Mohammadi, H. R. S. (2016) · International Symposium on Telecommunications · doi:10.1109/istel.2016.7881869 · Cited by 4
“The results indicate that PNCC and IMFCC features outperform MFCC and LFCC.”
BackgroundMethodInfluentialReplay spoofing countermeasures using high spectro-temporal resolution features
Alluri, K. N. R. K. R., & Vuppala, A. K. (2019) · International Journal of Speech Technology · doi:10.1007/s10772-019-09602-z · Cited by 2
Smartphone Audio Replay Attacks Dataset
Mandalapu, H., Ramachandra, R., & Busch, C. (2021) · International Workshop on Biometrics and Forensics (IWBF) · doi:10.1109/iwbf50991.2021.9465096 · Cited by 2
Chaudhari, A., & Shedge, D. K. (2021) · International Conference for Emerging Technology (INCET) · doi:10.1109/incet51464.2021.9456331 · Cited by 2
“An extensive experimentation carried out in [8] demonstrates that information content in high-frequency Fig.”
BackgroundMultilingual Audio-Visual Smartphone Dataset and Evaluation
Mandalapu, H., N, A. R. P., Ramachandra, R., Rao, K. S., Mitra, P., Prasanna, S. R. M., & Busch, C. (2021) · IEEE Access · doi:10.1109/access.2021.3125485 · Cited by 6
“The initial approach of LFCCs is used for the detection of synthetic speech in [34].”
Identification of Synthetic Spoofed Speech with Deep Capsule Network
Mao, T., Yan, D., Gong, Y., & Wang, R. (2022) · Communications in computer and information science · doi:10.1007/978-981-19-0523-0_17 · Cited by 2
Improving Robustness of Speech Anti-Spoofing System Using Resnext with Neighbor Filters
Zhang, Z., Zhao, X., & Yi, X. (2022) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme52920.2022.9859858 · Cited by 4
Fisher Vectors for Biometric Presentation Attack Detection
Gonzalez-Soler, L. J., Gomez‐Barrero, M., Patino, J., Kamble, M. R., Todisco, M., & Busch, C. (2023) · Advances in computer vision and pattern recognition · doi:10.1007/978-981-19-5288-3_18 · Cited by 2
Mankad, S. H., Garg, S., Patel, V., & Patwa, N. (2023) · Journal of Ambient Intelligence and Humanized Computing · doi:10.1007/s12652-023-04684-9 · Cited by 2
“The process for obtaining LFCC and IMFCC features differ in placement of filters, and is described in Fig.”
InfluentialShort-Utterance-Based Children’s Speaker Verification in Low-Resource Conditions
Aziz, S., Ankita, A., & Shahnawazuddin, S. (2023) · Circuits Systems and Signal Processing · doi:10.1007/s00034-023-02535-8 · Cited by 4
Protecting Voice-Controlled Devices Against LASER Injection Attacks
Ali, H., Khuttan, D., Refat, R. U. D., & Malik, H. M. A. (2023) · IEEE International Workshop on Information Forensics and Security · doi:10.1109/wifs58808.2023.10374658 · Cited by 3
“Baseline 1 used Q cepstral coefficients (CQCCs) as features [12] and baseline 2 used linear frequency cepstral coefficients (LFCCs) [20] as features.”
MethodNovel Energy Separation Based Frequency Modulation Features for Spoofed Speech Classification
Kamble, M. R., & Patil, H. A. (2017) · International Conference on Advances in Pattern Recognition · doi:10.1109/icapr.2017.8593041 · Cited by 2
“This might suggest that the feature extraction has much more important role to design of spoofing countermeasures rather than the advanced or complex classifiers [17].”
BackgroundReplay Spoofing Countermeasure Using Autoencoder and Siamese Network on ASVspoof 2019 ChallengePreprint
Adiban, M., Sameti, H., & Shehnepoor, S. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1910.13345 · Cited by 2
“proposed method in the baseline system is based on CQCC and Linear Frequency Cepstral Coefficients (LFCC) [51] features and GMM classifiers.”
Method
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.