Search

Most of the state-of-the-art speaker recognition system use natural speech signal (i.e., real speech, spontaneous speech or contextual speech) from the subjects. In this paper, recognition of a person is attempted from his or her hum with the help of machines. This kind of application can be useful to design person-dependent Query-by-Humming (QBH) system and hence, plays an important role in music information retrieval (MIR) system. In addition, it can be also useful for other interesting speech technological applications such as human-computer interaction, speech prosody analysis of disordered speech, and speaker forensics. This paper develops new feature extraction technique to exploit perceptually meaningful (due to mel frequency warping to imitate human perception process for hearing) phase spectrum information along with magnitude spectrum information from the hum signal. In particular, the structure of state-of-the-art feature set, namely, Mel Frequency Cepstral Coefficients (MFCCs) is modified to capture the phase spectrum information. In addition, a new energy measure , namely, Variable length Teager Energy Operator (VTEO) is employed to compute subband energies of different time-domain subband signals (i.e., an output of 24 triangular-shaped filters used in the mel filterbank). We refer this proposed feature set as MFCC-VTMP (i.e., mel frequency cepstral coefficients to capture perceptually meaningful magnitude and phase information via VTEO)The polynomial classifier (which is in-principle similar to other discriminatively-trained classifiers such as support vector machine (SVM) with polynomial kernel) is used as the basis for all the experiments. The effectiveness of proposed feature set is evaluated and consistently found to be better than MFCCs feature set for several evaluation factors, such as, comparison with other phase-based features, the order of polynomial classifier, person (speaker) modeling approach (such as, GMM-UBM and i -vector), the dimension of feature vector, robustness under signal degradation conditions, static vs. dynamic features, feature discrimination measures and intersession variability. [ABSTRACT FROM AUTHOR]

Recent works on the vulnerability of automatic speaker verification (ASV) systems confirm that malicious spoofing attacks using synthetic speech can provoke significant increase in false acceptance rate. A reliable detection of synthetic speech is key to develop countermeasure for synthetic speech based spoofing attacks. In this paper, we targeted that by focusing on three major types of artifacts related to magnitude, phase and pitch variation, which are introduced during the generation of synthetic speech. We proposed a new approach to detect synthetic speech using score-level fusion of front-end features namely, constant Q cepstral coefficients (CQCCs), all-pole group delay function (APGDF) and fundamental frequency variation (FFV). CQCC and APGDF were individually used earlier for spoofing detection task and yielded the best performance among magnitude and phase spectrum related features, respectively. The novel FFV feature introduced in this paper to extract pitch variation at frame-level, provides complementary information to CQCC and APGDF. Experimental results show that the proposed approach produces the best stand-alone spoofing detection performance using Gaussian mixture model (GMM) based classifier on ASVspoof 2015 evaluation dataset. An overall equal error rate of 0.05% with a relative performance improvement of 76.19% over the next best-reported results is obtained using the proposed method. In addition to outperforming all existing baseline features for both known and unknown attacks, the proposed feature combination yields superior performance for ASV system (GMM with universal background model/i-vector) integrated with countermeasure framework. Further, the proposed method is found to have relatively better generalization ability when either one or both of copy-synthesized data and limited spoofing data are available a priori in the training pool. [ABSTRACT FROM AUTHOR]

In this paper we investigate the audio-visual processing of linguistic prosody, more precisely the detection of word prominence, and examine how the additional visual information can be used to increase the robustness when acoustic background noise is present. We evaluate the detection performance for each modality individually and perform experiments using feature and decision fusion. For the latter we also consider the adaptive fusion with fusion weights adjusted to the current acoustic noise level. Our experiments are based on a corpus with 11 English speakers which contains in addition to the speech signal also videos of the speakers’ heads. From the acoustic signal we extract features which are well known to capture word prominence like loudness, fundamental frequency and durational features. The analysis of the visual signal is based on features derived from the speaker’s rigid head movements and movements of the speaker’s mouth. We capture the rigid head movements by tracking the speaker’s nose. Via a two-dimensional Discrete Cosine Transform (DCT) calculated from the mouth region we represent the movements of the speaker’s mouth. The results show that the rigid head movements as well as movements inside the mouth region can be used to discriminate prominent from non-prominent words. The audio-only detection yields an Equal Error Rate (EER) averaged over all speakers of 13%. Based only on the visual features we obtain 20% of EER. When we combine the visual and the acoustic features we only see a small improvement compared to the audio-only detection for clean speech. To simulate background noise we added 4 different noise types at varying SNR levels to the acoustic stream. The results indicate that word prominence detection is quite robust against additional background noise. Even at a severe Signal to Noise Ratio (SNR) of − 10 dB the EER only rises to 35%. Despite this the audio-visual fusion leads to notable improvements for the detection from noisy speech. We observe relative reductions of the EER of up to 79%. [ABSTRACT FROM AUTHOR]

The paper describes a system for automatic speech recognition (ASR) that is benchmarked with data of the 3rd CHiME challenge, a dataset comprising distant microphone recordings of noisy acoustic scenes in public environments. The proposed ASR system employs various methods to increase recognition accuracy and noise robustness. Two different multi-channel speech enhancement techniques are used to eliminate interfering sounds in the audio stream. One speech enhancement method aims at separating the target speaker's voice from background sources based on non-negative matrix factorization (NMF) using variational Bayesian (VB) inference to estimate NMF parameters. The second technique is based on a time-varying minimum variance distortionless response (MVDR) beamformer that uses spatial information to suppress sound signals not arriving from a desired direction. Prior to speech enhancement, a microphone channel failure detector is applied that is based on cross-comparing channels using a modulation-spectral representation of the speech signal. ASR feature extraction employs the amplitude modulation filter bank (AMFB) that implicates prior information of speech to analyze its temporal dynamics. AMFBs outperform the commonly used frame splicing technique of filter bank features in conjunction with a deep neural network (DNN) based ASR system, which denotes an equivalent data-driven approach to extract modulation-spectral information. In addition, features are speaker adapted, a recurrent neural network (RNN) is employed for language modeling, and hypotheses of different ASR systems are combined to further enhance the recognition accuracy. The proposed ASR system achieves an absolute word error rate (WER) of 5.67% on the real evaluation test data, which is 0.16% lower compared to the best score reported within the 3rd CHiME challenge. [ABSTRACT FROM AUTHOR]

This paper gives an in-depth presentation of the multi-microphone speech recognition system we submitted to the 3rd CHiME speech separation and recognition challenge (CHiME-3) and its extension. The proposed system takes advantage of recurrent neural networks (RNNs) throughout the model from the front-end speech enhancement to the language modeling. Three different types of beamforming are used to combine multi-microphone signals to obtain a single higher-quality signal. The beamformed signal is further processed by a single-channel long short-term memory (LSTM) enhancement network, which is used to extract stacked mel-frequency cepstral coefficients (MFCC) features. In addition, the beamformed signal is processed by two proposed noise-robust feature extraction methods. All features are used for decoding in speech recognition systems with deep neural network (DNN) based acoustic models and large-scale RNN language models to achieve high recognition accuracy in noisy environments. Our training methodology includes multi-channel noisy data training and speaker adaptive training, whereas at test time model combination is used to improve generalization. Results on the CHiME-3 benchmark show that the full set of techniques substantially reduced the word error rate (WER). Combining hypotheses from different beamforming and robust-feature systems ultimately achieved 5.05% WER for the real-test data, an 84.7% reduction relative to the baseline of 32.99% WER and a 44.5% reduction from our official CHiME-3 challenge result of 9.1% WER. Furthermore, this final result is better than the best result (5.8% WER) reported in the CHiME-3 challenge. [ABSTRACT FROM AUTHOR]

This paper investigates the problem of measuring speech quality of received speech signal without employing the original speech signal. The problem of deterioration of the speech quality arises due to noise present in the surroundings. To this line, we propose a multiple time-instances (MTI) features-based approach for reference-free speech quality measurement model. A voice activity detector (VAD) is exploited primarily for calculating the number of active speech chunks of a speech signal. For these chunks and their successive combinations called here batches, multi-resolution auditory model (MRAM), mel-frequency cepstral coefficients (MFCC) and line spectral frequencies (LSF) features are extracted and called as MTI features. It is hypothesized that the MTI features are capable in capturing the distortions caused by time-localized effects of short-time transients, impulsive noise, and its differences from the plosive sounds. The MTI metric estimates (MTI-ME) are calculated corresponding to these MTI features employing the Gaussian mixture model (GMM) probabilistic technique. The overall objective speech quality of a speech signal is then determined as a linear combination of optimally weighted MTI-ME corresponding to distinct active speech chunks and their successive combinations, that is, batches of that speech signal. Minimum mean square error criterion or Pearson's correlation maximization criterion is employed for computing optimal weights. In addition, a deep neural network (DNN)-based speech quality model is also developed for calculating a single objective speech quality while considering all active speech chunks together. Pearson's correlation coefficient and weighted average correlation are exploited for evaluating the performance. Results demonstrate that the proposed model achieves promising improvement over the standard speech quality model (P.563) and improves correlation values by around 37%. • Addressing speech quality measuring problem by developing a reference-free speech quality model. • Developing a feature extraction technique incorporating distinct auditory features. • Developing a joint GMM training algorithm for computing objective speech quality. • Developing a deep neural network (DNN) framework for validating the proposed algorithm. • Demonstrating better performance as compared to the standard speech quality model (P.563). [ABSTRACT FROM AUTHOR]

The deep learning methodologies in state-of-the-art speaker recognition systems are predominantly limited to the extraction of recording level embeddings. This is usually followed by generative modeling of the embeddings to output the verification score. In this paper, we explore a fully neural approach where the neural model outputs the verification score directly, given the acoustic feature inputs. This model, termed as Siamese neural network (SiamNN), combines the embedding extraction and back-end modeling into a single processing pipeline. The back-end modeling is achieved using a neural approach to PLDA modeling, called neural probabilistic linear discriminant analysis (NPLDA). In the NPLDA model, the verification score is computed as a discriminative similarity function. The development of the single neural SiamNN model allows the joint optimization of all the modules using a verification cost. Several speaker recognition experiments are performed using SITW, VOiCES, and NIST SRE datasets where the proposed SiamNN model is shown to significantly improve over the state-of-art x-vector PLDA baseline system (relative improvements of up to 35% in the primary cost metric). We also provide a detailed analysis of the influence of hyper-parameters, choice of loss functions, and data sampling strategies for training the model. In particular, we highlight that the proposed soft detection cost function based optimization improves over other loss functions considered. [ABSTRACT FROM AUTHOR]

• The vowel transition regions lie in the junction between a consonant and vowel (CV) regions and between a vowel and a consonant (VC) regions. • We have proposed a method for determining the vowel transition regions based on the rate of change of formant frequencies using zero-time windowing and numerator of group-delay function. • Zero-time windowing derives the instantaneous formant frequencies accurately at every sample location due to contribution of that sample itself. • The numerator of group-delay function enhances the formant frequencies. • The proposed transition region detection method is evaluated on CV, CVC and TIMIT databases and it has shown significant improvement in the performance compared to the existing method. The vowel transition regions are the crucial landmarks in the speech signal. These vital regions are present at both ends of the vowel. They lie in the junction between a consonant and a vowel (CV) regions. This region plays an important role in numerous speech applications like speaker recognition, emotion conversion, speech rate modification, and CV unit recognition. The performance of these applications crucially depends on the accuracy of the estimation of vowel transition regions. In this paper, we have proposed a method for determining the transition regions based on the rate of change of formant frequencies using zero-time windowing and numerator of the group-delay function. Zero-time windowing derives the instantaneous formant frequencies accurately at every sample location due to the contribution of that sample itself. The numerator of the group-delay function enhances the formant frequencies. The proposed transition region detection method is evaluated on CV, and continuous speech databases recorded in the Hindi language. The proposed method has shown around 12% improvement in accuracy compared to the existing method. [ABSTRACT FROM AUTHOR]