Automatic Speaker Verification Research Articles

Voice Spoofing Countermeasure Based on Spectral Features to Detect Synthetic Attacks Through LSTM

With the growing number of voice-controlled devices, it is necessary to address the potential vulnerabilities of Automatic Speaker Verification (ASV) against voice spoofing attacks such as Physical Access (PA) and Logical Access (LA) attacks. To improve the reliability of ASV systems, researchers have developed various voice spoofing countermeasures. However, it is hard for the voice anti-spoofing systems to effectively detect the synthetic speech attacks that are generated through powerful spoofing algorithms and have quite different statistical distributions. More importantly, the speedy improvement of voice spoofing structures is producing the most effective attacks that make ASV structures greater vulnerable to stumble on those voice spoofing assaults. In this paper, we proposed a unique voice spoofing countermeasure which is successful to hit upon the LA attacks (i.e., artificial speech and transformed speech) and classify the spoofing structures by the usage of Long Short-Term Reminiscence (LSTM). The novel set of spectral features i.e., Mel-Frequency Cepstral Coefficients (MFCC), Gammatone Cepstral Coefficients (GTCC), and spectral centroid are capable to seize maximum alterations present in the cloned audio. The proposed system achieved remarkable accuracy of 98.93%, precision of 100%, recall of 92.32%, F1-score of 96.01%, and an Equal Error Rate (EER) of 1.30%. Our method achieved 8.5% and 7.02% smaller EER than the baseline methods such as Constant-Q Cepstral Coefficients (CQCC) using Gaussian Mixture Model (GMM) and Linear Frequency Cepstral Coefficients (LFCC) using GMM, respectively. We evaluated the performance of the proposed system on the standard dataset i.e., ASVspoof2019 LA. Experimental results and comparative analysis with other existing state-of-the-art methods illustrate that our method is reliable and effective to be used for the detection of voice spoofing attacks.

Uncertainty assessment for detection of spoofing attacks to speaker verification systems using a Bayesian approach

There has been tremendous progress in automatic speaker verification systems over the last decade. Still, spoofing attacks pose a significant challenge to their deployment. Even though there are various attack techniques such as voice conversion and speech synthesis, replay attacks pose one of the most important types since they can be done without significant expertise in speech technology. Moreover, replay attacks are hard to detect because they are done with simple replay of the original audio. The problem has gained more attention since the introduction of the ASV spoof 2017 challenge, which included a well-designed database with realistic replay attack conditions. Even though many different deep network types and acoustic features were proposed since the challenge, one key issue, which is model uncertainty around the neural networks’ decision is largely ignored. This is a result of using the softmax function with the cross-entropy loss, which is widely used in many domains. Here, we propose using evidential deep learning, which is a recently proposed method that is rapidly gaining popularity, for assessing the model uncertainty around the network’s decision. Experimental results show that the investigated network architectures perform better in terms of equal error rate with the new loss function. Moreover, reliability of measured uncertainty is shown by filtering samples out of the test set using the Bayesian uncertainty measure, which resulted with a consistent decrease in EER with decreasing threshold.

Improving the Adversarial Robustness for Speaker Verification by Self-Supervised Learning

Previous works have shown that automatic speaker verification (ASV) is seriously vulnerable to malicious spoofing attacks, such as replay, synthetic speech, and recently emerged adversarial attacks. Great efforts have been dedicated to defending ASV against replay and synthetic speech; however, only a few approaches have been explored to deal with adversarial attacks. All the existing approaches to tackle adversarial attacks for ASV require the knowledge for adversarial samples generation, but it is impractical for defenders to know the exact attack algorithms that are applied by the in-the-wild attackers. This work is among the first to perform adversarial defense for ASV without knowing the specific attack algorithms. Inspired by self-supervised learning models (SSLMs) that possess the merits of alleviating the superficial noise in the inputs and reconstructing clean samples from the interrupted ones, this work regards adversarial perturbations as one kind of noise and conducts adversarial defense for ASV by SSLMs. Specifically, we propose to perform adversarial defense from two perspectives: 1) adversarial perturbation purification and 2) adversarial perturbation detection. The purification module aims at alleviating the adversarial perturbations in the samples and pulling the contaminated adversarial inputs back towards the decision boundary. Experimental results show that our proposed purification module effectively counters adversarial attacks and outperforms traditional filters from both alleviating the adversarial noise and maintaining the performance of genuine samples. The detection module aims at detecting adversarial samples from genuine ones based on the statistical properties of ASV scores derived by a unique ASV integrating with different number of SSLMs. Experimental results show that our detection module helps shield the ASV by detecting adversarial samples. Both purification and detection methods are helpful for defending against different kinds of attack algorithms. Moreover, since there is no common metric for evaluating the ASV performance under adversarial attacks, this work also formalizes evaluation metrics for adversarial defense considering both purification and detection based approaches into account. We sincerely encourage future works to benchmark their approaches based on the proposed evaluation framework.

Optimizing Tandem Speaker Verification and Anti-Spoofing Systems

As automatic speaker verification (ASV) systems are vulnerable to spoofing attacks, they are typically used in conjunction with spoofing countermeasure (CM) systems to improve security. For example, the CM can first determine whether the input is human speech, then the ASV can determine whether this speech matches the speaker's identity. The performance of such a tandem system can be measured with a tandem detection cost function (t-DCF). However, ASV and CM systems are usually trained separately, using different metrics and data, which does not optimize their combined performance. In this work, we propose to optimize the tandem system directly by creating a differentiable version of t-DCF and employing techniques from reinforcement learning. The results indicate that these approaches offer better outcomes than finetuning, with our method providing a 20% relative improvement in the t-DCF in the ASVSpoof19 dataset in a constrained setting.

Task-Specific Optimization of Virtual Channel Linear Prediction-Based Speech Dereverberation Front-End for Far-Field Speaker Verification

Developing a single-microphone speech denoising or dereverberation front-end for robust automatic speaker verification (ASV) in noisy far-field speaking scenarios is challenging. To address this problem, we present a novel front-end design that involves a recently proposed extension of the weighted prediction error (WPE) speech dereverberation algorithm, the virtual acoustic channel expansion (VACE)-WPE. It is demonstrated experimentally in this study that unlike the conventional WPE algorithm, the VACE-WPE can be explicitly trained to cancel out both late reverberation and background noise. To build the front-end, the VACE-WPE is first independently (pre)trained to produce "noisy" dereverberated signals. Subsequently, given a pretrained speaker embedding model, the VACE-WPE is additionally fine-tuned within a task-specific optimization (TSO) framework, causing the speaker embedding extracted from the processed signal to be similar to that extracted from the "noise-free" target signal. Moreover, to extend the application of the proposed front-end to more general, unconstrained "in-the-wild" ASV scenarios beyond controlled far-field conditions, we propose a distortion regularization method for the VACE-WPE within the TSO framework. The effectiveness of the proposed approach is verified on both far-field and in-the-wild ASV benchmarks, demonstrating its superiority over fully neural front-ends and other TSO methods in various cases.

Use of Speaker Recognition Approaches for Learning and Evaluating Embedding Representations of Musical Instrument Sounds

Constructing an embedding space for musical instrument sounds that can meaningfully represent new and unseen instruments is important for downstream music generation tasks such as multi-instrument synthesis and timbre transfer. The framework of Automatic Speaker Verification (ASV) provides us with architectures and evaluation methodologies for verifying the identities of unseen speakers, and these can be repurposed for the task of learning and evaluating a musical instrument sound embedding space that can support unseen instruments. Borrowing from state-of-the-art ASV techniques, we construct a musical instrument recognition model that uses a SincNet front-end, a ResNet architecture, and an angular softmax objective function. Experiments on the NSynth and RWC datasets show our model’s effectiveness in terms of equal error rate (EER) for unseen instruments, and ablation studies show the importance of data augmentation and the angular softmax objective. Experiments also show the benefit of using a CQT-based filterbank for initializing SincNet over a Mel filterbank initialization. Further complementary analysis of the learned embedding space is conducted with t-SNE visualizations and probing classification tasks, which show that including instrument family labels as a multi-task learning target can help to regularize the embedding space and incorporate useful structure, and that meaningful information such as playing style, which was not included during training, is contained in the embeddings of unseen instruments.

Text-independent Speaker Verification Using Hybrid Convolutional Neural Networks

Automatic speaker verification is an active research area for more than four decades, and the technology has gradually upgraded for real application. In this paper, a hybrid convolutional neural network (CNN) model is proposed where a combination of the 3D CNN & 2D CNN model is used for speaker verification in the text-independent scenario. For speaker verification, this novel convolutional neural network architecture was built to capture and discard speaker and non-speaker information at the same time. In the training process, the network is trained to differentiate between different identities of a speaker to establish the background model. The model development of the speaker is one of the important aspects. Most conventional techniques employed the d-vector system to create speaker models by means of an average of the features collected from the speaker utterance. Here a hybrid of convolutional neural networks model is utilized in the development and registration phases for building a speaker model. The approach suggested exceeds the existing methods of speaker verification.

Static\u2013dynamic features and hybrid deep learning models based spoof detection system for ASV

Detection of spoof is essential for improving the performance of current scenario of Automatic Speaker Verification (ASV) systems. Empowerment to both frontend and backend parts can build the robust ASV systems. First, this paper discuses performance comparison of static and static–dynamic Constant Q Cepstral Coefficients (CQCC) frontend features by using Long Short Term Memory (LSTM) with Time Distributed Wrappers model at the backend. Second, it performs comparative analysis of ASV systems built using three deep learning models LSTM with Time Distributed Wrappers, LSTM and Convolutional Neural Network at backend and using static–dynamic CQCC features at frontend. Third, it discusses implementation of two spoof detection systems for ASV by using same static–dynamic CQCC features at frontend and different combination of deep learning models at backend. Out of these two, the first one is a voting protocol based two-level spoof detection system that uses CNN, LSTM model at first level and LSTM with Time Distributed Wrappers model at second level. The second one is a two-level spoof detection system with user identification and verification protocol, which uses LSTM model for user identification at first level and LSTM with Time Distributed Wrappers for verification at the second level. For implementing the proposed work, a variation in ASVspoof 2019 dataset has been used to introduce all types of spoofing attacks such as Speech Synthesis (SS), Voice Conversion (VC) and replay in single set of dataset. The results show that, at frontend, static–dynamic CQCC feature outperform static CQCC features and at the backend, hybrid combination of deep learning models increases accuracy of spoof detection systems.

Lightweight and irreversible speech pseudonymization based on data-driven optimization of cascaded voice modification modules

In this paper, we propose a speech pseudonymization framework that utilizes cascaded and superposition-based voice modification modules. With increasing opportunities to use spoken dialogue systems nowadays, research regarding protecting the privacy of speaker information encapsulated in speech data is attracting attention. Pseudonymization, which is one method for voice privacy protection, aims to keep the intelligibility of speech while simultaneously suppressing speaker-specific information. One motivation of our framework is to achieve a reliable pseudonymization performance with light computation. To do this, we utilize the advantages of both machine learning-based and signal processing-based approaches. The advantages are (1) using signal processing-based methods parameterized with few hyperparameters and (2) using machine learning-based optimization to optimize all hyperparameters on the basis of black-box systems consisting of automatic speaker verification and automatic speech recognition. Our method of cascading signal processing modules, which are jointly optimized in a data-driven manner, can pseudonymize speech in a lightweight manner. Additionally, we discuss irreversible pseudonymization approaches and propose a superposition approach, yet another pseudonymization method that is more irreversible than the cascade method in terms of estimating the adequate parameters to recover the original signal. From the experimental results conducted under the VoicePrivacy 2020 protocols, we can demonstrate that (1) our cascade method succeeds in deteriorating the speaker recognition rate by over 24% while simultaneously improving the speech recognition rate by approximately 8% compared with a signal processing-based baseline system of VoicePrivacy 2020 and that (2) our superposition method works comparable to our cascade method in terms of pseudonymization performance.

A review on state-of-the-art Automatic Speaker verification system from spoofing and anti-spoofing perspective

Background/Objectives: The anti-spoofing measures are blooming with an aim to protect the Automatic Speaker Verification systems from susceptible spoofing attacks. This review is an amalgam of the possible attack types, the datasets required, the renowned feature representation techniques, modeling algorithms involving machine learning, and score normalization techniques. Method/Findings: A detailed analysis of existing datasets is carried based on the total speaker samples, the number of speakers, and source of availability- open or licensed. This may foster choosing the right dataset for building the anti-spoofing frameworks. Further, the feature extraction schemes are elaborated with an intention to cover the vast span of features existing in various parts of raw speech for obtaining speaker-specific traits. Further, the machine learning algorithms ranging from discriminative to generative to mixed form are explored for seeking the right algorithm in specific attack conditions. On the whole, these analyses of existing features and machine learning algorithms together contribute to classifying the unknown test samples as genuine or spoofed. The score normalization techniques are also considered in this review to avoid any misclassifications and ultimately reduce the False Acceptance Ratios. The performance of any anti-spoofing speaker verification system may be evaluated using standard objective measures such are Equal Error Rate, False positive ratios, and graphical plots. These measures are briefly explained in this review. Overall, the critical analysis of individual methods-feature extraction, machine learning, score normalization, and all the anti-spoofing datasets are also discussed for giving a kick-start to any researcher beginning to explore in this direction. The shortcomings and risks involved in building an enhanced speaker verification system that is robust to almost all the attack types are listed in this article. The review of studies conducted so far has led to vital future directions that are enlisted in the concluding remarks of the article. Keywords: Automatic Speaker Verification; Spoofed Detection, Anti­Spoofing, Voice Conversion, Speech Synthesis, Replay Speech

Improving the potential of Enhanced Teager Energy Cepstral Coefficients (ETECC) for replay attack detection

In the scope of voice biometrics, the term replay attack, (RA) refers to the dishonest attempt made by an impostor to spoof someone else’s identity by replaying the subject’s previously recorded speech close to the Automatic Speaker Verification (ASV) system under attack. State-of-the-art strategies for RA detection, such as the Enhanced Teager Energy Cepstral Coefficients (ETECC), have shown promising results due to their precision in measuring energy from high frequency components of speech, as a function of two recently defined concepts: signal mass and Enhanced Teager Energy Operator (ETEO). Nevertheless, since the replay mechanism prominently deteriorates the speech signal spectrum just in those spectral zones, we propose the association of ETEO with different strategies to further improve the previous results in getting effective countermeasures against RAs. Specifically, comprehensive evaluations which include a detailed mathematical analysis, a simulation on amplitude and frequency modulated (AM–FM) signals, and a spectrographic inspection involving different filterbank structures, along with their experimental results, are provided in this paper. In addition, ETEO-derived features are contrasted to existing feature sets by using Paraconsistent Feature Engineering (PFE) for feature ranking, expanding our previously published results. Lastly, experiments are performed with ASVSpoof-2017 version 2.0 dataset, Realistic Replay Attack Microphone Array Speech Corpus (ReMASC), BTAS-2016, dataset, ASVSpoof-2019 challenge dataset, and ASVSpoof-2015 challenge dataset, considering Gaussian Mixture Models (GMMs), Convolutional Neural Networks (CNNs) and Light-CNN architectures as being the classifiers. The standalone ETECC-GMM system showed the best performance by producing equal error rates (EERs) of 5.55% and 10.75% on development and evaluation sets, respectively.

Adjustable deterministic pseudonymization of speech

While public speech resources become increasingly available, there is a growing interest to preserve the privacy of the speakers, through methods that anonymize the speaker information from speech while preserving the spoken linguistic content. In this paper, a method for pseudonymization (reversible anonymization) of speech is presented, that allows to obfuscate the speaker identity in untranscribed running speech. The approach manipulates the spectro-temporal structure of the speech to simulate a different length and structure of the vocal tract by modifying the formant locations, as well as by altering the pitch and speaking rate. The method is deterministic and partially reversible, and the changes are adjustable on a continuous scale. The method has been evaluated in terms of (i) ABX listening experiments, and (ii) automatic speaker verification and speech recognition. ABX experimental results indicate that the speaker identifiability among forced choice pairs reduced from over 90% to less than 70% through pseudonymization, and that de-pseudonymization was partially effective. An evaluation on the VoicePrivacy 2020 challenge data showed that the proposed approach performs better than the signal processing based baseline method that uses McAdams coefficient and performs slightly worse than the neural source filtering based baseline method. Further analysis showed that the proposed approach: (i) is comparable to the neural source filtering baseline based method in terms of phone posterior feature based objective intelligibility measure, (ii) preserves formant tracks better than the McAdams based method, and (iii) preserves paralinguistic aspects such as dysarthria in several speakers.

Machine Learning Mitigants for Speech Based Cyber Risk

Statistical analysis of speech is an emerging area of machine learning. In this paper, we tackle the biometric challenge of Automatic Speaker Verification (ASV) of differentiating between samples generated by two distinct populations of utterances, those of an authentic human voice and those generated by a synthetic one. Solving such an issue through a statistical perspective foresees the definition of a decision rule function and a learning procedure to identify the optimal classifier. Classical state-of-the-art countermeasures rely on strong assumptions such as stationarity or local-stationarity of speech that may be atypical to encounter in practice. We explore in this regard a robust non-linear and non-stationary signal decomposition method known as the Empirical Mode Decomposition combined with the Mel-Frequency Cepstral Coefficients in a novel fashion with a refined classifier technique known as multi-kernel Support Vector machine. We undertake significant real data case studies covering multiple ASV systems using different datasets, including the ASVSpoof 2019 challenge database. The obtained results overwhelmingly demonstrate the significance of our feature extraction and classifier approach versus existing conventional methods in reducing the threat of cyber-attack perpetrated by synthetic voice replication seeking unauthorised access.

Automatic speaker verification systems and spoof detection techniques: review and analysis

Automatic speaker verification systems and spoof detection techniques: review and analysis

A survey on presentation attack detection for automatic speaker verification systems: State-of-the-art, taxonomy, issues and future direction

The emergence of biometric technology provides enhanced security compared to the traditional identification and authentication techniques that were less efficient and secure. Despite the advantages brought by biometric technology, the existing biometric systems such as Automatic Speaker Verification (ASV) systems are weak against presentation attacks. A presentation attack is a spoofing attack launched to subvert an ASV system to gain access to the system. Though numerous Presentation Attack Detection (PAD) systems were reported in the literature, a systematic survey that describes the current state of research and application is unavailable. This paper presents a systematic analysis of the state-of-the-art voice PAD systems to promote further advancement in this area. The objectives of this paper are two folds: (i) to understand the nature of recent work on PAD systems, and (ii) to identify areas that require additional research. From the survey, a taxonomy of voice PAD and the trend analysis of recent work on PAD systems were built and presented, whereby the recent and relevant articles including articles from Interspeech and ICASSP Conferences, mostly indexed by Scopus, published between 2015 and 2021 were considered. A total of 172 articles were surveyed in this work. The findings of this survey present the limitation of recent works, which include spoof-type dependent PAD. Consequently, the future direction of work on voice PAD for interested researchers is established. The findings of this survey present the limitation of recent works, which include spoof-type dependent PAD. Consequently, the future direction of work on voice PAD for interested researchers is established.

Children's speaker verification in low and zero resource conditions

Our efforts towards developing an automatic speaker verification (ASV) system for child speakers are presented in this paper. For the majority of the languages, children's speech data for training the ASV system is either unavailable (zero-resource) or very limited (low-resource). Under low- and zero-resource conditions, developing an ASV system becomes a very challenging problem. To overcome this issue, we have studied the effectiveness of in-domain and out-of-domain data augmentation in this work. Speed and pitch modifications of children's speech are employed for synthetically creating data in the case of in-domain data augmentation. On the other hand, a limited amount of adults' speech is used when out-of-domain data augmentation is performed. Using adults' speech leads to severe acoustic mismatch due to dissimilarity in the attributes of speech data from adult and child speakers. To address this drawback, speech data from adult speakers are subjected to voice conversion (VC) to alter the acoustic attributes. A cycle-consistent generative adversarial network is used in this work for voice conversion. Voice conversion renders adults' speech perceptually similar to children's speech. The voice converted adults' data can then be used for augmentation, ensuring that the acoustic mismatch is minimal. To study the effectiveness of proposed data augmentation techniques experimentally, x-vector-based ASV system architecture is employed. At the same time, the role of i-vector is also studied in this paper. As a consequence of data augmentation, both equal error rate and minimum decision cost function are reduced significantly in low- and zero-resource conditions. At the same time, employing i-vectors for modeling speaker characteristics is noted to be superior. Finally, we have also presented a detailed study on the effect of data augmentation with child speakers' age variation.

Automatic speaker verification from affective speech using Gaussian mixture model based estimation of neutral speech characteristics

Intra-speaker variability, caused by emotional speech, is a real threat to the performance of speaker recognition systems. In fact, as human beings, we are constantly changing our emotional state. While many efforts have been made to increase automatic speaker verification (ASV) robustness towards channel effects or spoofing attacks, only a handful of studies have addressed the detrimental consequences of affective speech. In this paper, we propose a new method to minimize the mismatch between neutral and affective speech. To this end, a Gaussian mixture model is used to learn a prior probability distribution of the neutral speech for a given speaker (i.e., characterizing his/her source space). This knowledge is then used to minimize the differences between target (affective) and source (neutral) spaces. The proposed method is validated across four multi-lingual emotional datasets. Experimental results show a consistent improvement in performance across eight emotional states, with significant reductions of equal error rate relative to the baseline.

ASVspoof 2019: Spoofing Countermeasures for the Detection of Synthesized, Converted and Replayed Speech

The ASVspoof initiative was conceived to spearhead research in anti-spoofing for automatic speaker verification (ASV). This paper describes the third in a series of bi-annual challenges: ASVspoof 2019. With the challenge database and protocols being described elsewhere, the focus of this paper is on results and the top performing single and ensemble system submissions from 62 teams, all of which out-perform the two baseline systems, often by a substantial margin. Deeper analyses shows that performance is dominated by specific conditions involving either specific spoofing attacks or specific acoustic environments. While fusion is shown to be particularly effective for the logical access scenario involving speech synthesis and voice conversion attacks, participants largely struggled to apply fusion successfully for the physical access scenario involving simulated replay attacks. This is likely the result of a lack of system complementarity, while oracle fusion experiments show clear potential to improve performance. Furthermore, while results for simulated data are promising, experiments with real replay data show a substantial gap, most likely due to the presence of additive noise in the latter. This finding, among others, leads to a number of ideas for further research and directions for future editions of the ASVspoof challenge.

Improving short utterance speaker verification by combining MFCC and Entrocy in Noisy conditions

Short utterance and background noise represent great challenging for speaker verification due to the mismatch and limited training and/or retrieve data. A remarkable performance using matched training and testing conditions generally could be achieved in automatic speaker verification. However, mismatched noisy and short utterances conditions attend to drop the results significantly. Furthermore, the performance is significantly affected by the features extraction. The most common features in this field of the study are Mel-Frequency Cepstral Coefficients (MFCCs). With a noise presents in the background and short utterances, MFCC performance could not be reliable without a support feature. To address this, a new feature ‘Entrocy’ for accurate and robust speaker verification under limited data and noisy environments is proposed and employed to support MFCC coefficients. Entrocy feature represents the Fourier Transform of the Entropy that calculates the fluctuation of the information in the sound segments over time. The resulting Entrocy features are combined with MFCC functionality to generate a composite feature, which is tested using the Gaussian Mixture Model (GMM) recognition method. The suggested method was conducted out over a range of signal/noise ratios and utterances were truncating into shorts (2, 3, 4, 5, 6, 8, and 10s) for verification. The proposed method has shown strong robustness in the challenging of background noise and limited testing data and they consistently perform better than the well-known MFCC.

LSTM and CNN based ensemble approach for spoof detection task in automatic speaker verification systems

Nowadays, fingerprint and retina scans are the most reliable and widely used biometric authentication systems. Another emerging biometric approach for authentication, though vulnerable, is speech-based systems. However, speech-based systems are more prone to spoofing attacks. Through such malicious attacks, an unauthentic person tries to present himself as legitimate user, in order to acquire illegitimate advantage. Therefore, these attacks, created by synthesis or conversion of speech, pose an enormous threat to the reliable functioning of automatic speaker verification (ASV) authentication systems. The work presented in this paper tries to address this problem by using deep learning (DL) methods and ensemble of different neural networks. The first model is a combination of time-distributed dense layers and long short-term memory (LSTM) layers. The other two deep neural networks (DNNs) are based on temporal convolution (TC) and spatial convolution (SC). Finally, an ensemble model comprising of these three DNNs has also been analysed. All these models are analysed with Mel frequency cepstral coefficients (MFCC), inverse Mel frequency cepstral coefficients (IMFCC) and constant Q cepstral coefficients (CQCC) at the frontend, where the proposed ensemble performs best with CQCC features. The proposed work uses ASVspoof 2015 and ASVspoof 2019 datasets for training and testing, with the evaluation set having speech synthesis (SS) and voice conversion (VC) attacked utterances. Performance of proposed system trained with ASVspoof 2015 dataset degrades with evaluation set of ASVspoof 2019 dataset, whereas performance of the same system improves when training is also done with the ASVspoof 2019 dataset. Also, a joint ASVspoof 1519 dataset is created to add more variations into the single dataset, and it has been observed that the trained ensemble with this joint dataset performs even better, while performing evaluation using single datasets. This research promotes the development of systems that can cope with completely unknown data in testing. It has been further observed that the models can reach promising results, paving the way for future research in this domain using DL.