Baseline Mel Frequency Cepstral Coefficients Research Articles

Classification of functional dysphonia using the tunable Q wavelet transform

Functional dysphonia (FD) refers to an abnormality in voice quality in the absence of an identifiable lesion. In this paper, we propose an approach based on the tunable Q wavelet transform (TQWT) to automatically classify two types of FD (hyperfunctional dysphonia and hypofunctional dysphonia) from a healthy voice using the acoustic voice signal. Using TQWT, voice signals were decomposed into sub-bands and the entropy values extracted from the sub-bands were utilized as features for the studied 3-class classification problem. In addition, the Mel-frequency cepstral coefficient (MFCC) and glottal features were extracted from the acoustic voice signal and the estimated glottal source signal, respectively. A convolutional neural network (CNN) classifier was trained separately for the TQWT, MFCC and glottal features. Experiments were conducted using voice signals of 57 healthy speakers and 113 FD patients (72 with hyperfunctional dysphonia and 41 with hypofunctional dysphonia) taken from the VOICED database. These experiments revealed that the TQWT features yielded an absolute improvement of 5.5% and 4.5% compared to the baseline MFCC features and glottal features, respectively. Furthermore, the highest classification accuracy (67.91%) was obtained using the combination of the TQWT and glottal features, which indicates the complementary nature of these features.

FEATURE EXTRACTION ALGORITHM USING NEW CEPSTRAL TECHNIQUES FOR ROBUST SPEECH RECOGNITION

In this work, we propose a novel feature extraction algorithm that improves the robustness of automatic speech recognition (ASR) systems in the presence of various types of noise. The proposed algorithm uses a new cepstral technique based on the differential power spectrum (DPS) instead of the power spectrum (PS), the algorithm replaces the logarithmic non linearity by the power function. In order to reduce cepstral coefficients mismatches between training and testing conditions, we used the mean and variance normalization, then we apply auto-regression movingaverage filtering (MVA) in the cepstral domain. The ASR experiments were conducted using two databases, the first is LASA digit database designed for recognition the isolated Arabic digits in the presence of different types of noise. The second is Aurora 2 noisy speech database designed to recognize connected English digits in various operating environments. The experimental results show a substantial improvement from the proposed algorithm over the baseline Mel Frequency Cepstral Coefficients (MFCC), the relative improvement is the 28.92% for LASA database and is the 44.43% for aurora 2 database. The performance of our proposed algorithm was tested and verified by extensive comparisons with the state-of-the-art noise-robust features in aurora 2.

Improved Speech-Signal Based Frequency Warping Scale for Cepstral Feature in Robust Speaker Verification System

Development of automatic speaker verification system (ASV) for real-world applications remains a major challenge. In this paper, we propose an improved speech-signal-based frequency warping scale to extract cepstral features from the speech signal for ASV application. The proposed scale is a modified version of the speech-signal-based scale, successfully used in speech recognition application, an allied domain. It uses spectral entropy weighted power spectral density to extract speaker specific attributes. This is complementary to fixed scale based mel frequency cepstral coefficient (MFCC) for different emphasis given to spectral regions. The work uses fusion based approach to exploit the complementarity of static MFCC and proposed feature. The performances of the ASV system that uses MFCC and the proposed technique are evaluated in clean and various noisy conditions on publicly available NIST SRE databases. Noise database (NOISEX-92) is used to simulate the noisy environment. The ASV system developed from the proposed feature extraction method shows slightly improved performance than baseline MFCC and SFCC (speech-signal-based frequency cepstral coefficient) based techniques in clean condition and up to 38.15% and 17.15%, respectively in noisy conditions. The fusion-based approach further improves the performance of ASV system with up to 53.85% and 36.22% relative improvement over baseline MFCC and SFCC based feature extraction methods, respectively.

Text-Independent Speaker Identification Through Feature Fusion and Deep Neural Network

Speaker identification refers to the process of recognizing human voice using artificial intelligence techniques. Speaker identification technologies are widely applied in voice authentication, security and surveillance, electronic voice eavesdropping, and identity verification. In the speaker identification process, extracting discriminative and salient features from speaker utterances is an important task to accurately identify speakers. Various features for speaker identification have been recently proposed by researchers. Most studies on speaker identification have utilized short-time features, such as perceptual linear predictive (PLP) coefficients and Mel frequency cepstral coefficients (MFCC), due to their capability to capture the repetitive nature and efficiency of signals. Various studies have shown the effectiveness of MFCC features in correctly identifying speakers. However, the performances of these features degrade on complex speech datasets, and therefore, these features fail to accurately identify speaker characteristics. To address this problem, this study proposes a novel fusion of MFCC and time-based features (MFCCT), which combines the effectiveness of MFCC and time-domain features to improve the accuracy of text-independent speaker identification (SI) systems. The extracted MFCCT features were fed as input to a deep neural network (DNN) to construct the speaker identification model. Results showed that the proposed MFCCT features coupled with DNN outperformed existing baseline MFCC and time-domain features on the LibriSpeech dataset. In addition, DNN obtained better classification results compared with five machine learning algorithms that were recently utilized in speaker recognition. Moreover, this study evaluated the effectiveness of one-level and two-level classification methods for speaker identification. The experimental results showed that two-level classification presented better results than one-level classification. The proposed features and classification model for identifying a speaker can be widely applied to different types of speaker datasets.

An Investigation of a Feature-Level Fusion for Noisy Speech Emotion Recognition

Because one of the key issues in improving the performance of Speech Emotion Recognition (SER) systems is the choice of an effective feature representation, most of the research has focused on developing a feature level fusion using a large set of features. In our study, we propose a relatively low-dimensional feature set that combines three features: baseline Mel Frequency Cepstral Coefficients (MFCCs), MFCCs derived from Discrete Wavelet Transform (DWT) sub-band coefficients that are denoted as DMFCC, and pitch based features. Moreover, the performance of the proposed feature extraction method is evaluated in clean conditions and in the presence of several real-world noises. Furthermore, conventional Machine Learning (ML) and Deep Learning (DL) classifiers are employed for comparison. The proposal is tested using speech utterances of both of the Berlin German Emotional Database (EMO-DB) and Interactive Emotional Dyadic Motion Capture (IEMOCAP) speech databases through speaker independent experiments. Experimental results show improvement in speech emotion detection over baselines.

Improved filter bank on multitaper framework for robust Punjabi-ASR system

Robustness of the automatic speech recognition (ASR) system relies upon the accuracy of feature extraction and classification in training phase. The mismatch between training and testing conditions during classification of large feature vectors causes a low performance. In this paper, the issue of robustness of acoustic information is addressed for practical Punjabi dataset. Traditional feature extraction approaches: mel frequency cepstral coefficients (MFCC) and gammatone frequency cepstral coefficients (GFCC) face the issue of high variance with leakage of spectral information. Also, handling of the huge number of feature information creates chaos for large speech vocabulary. To overcome this dilemma, a Principal component analysis (PCA) based multi-windowing technique is proposed with the incorporation of baseline GFCC and MFCC based feature approaches after the tuning of taper parameter. The proposed integrated approaches result in better feature vectors, which are further processed using differential evolution + hidden Markov model (DE + HMM) based modelling classifier. The integrated approaches show substantial performance for word recognition as compared to the conventional or fused feature extraction systems.

Filterbank Optimization for Text-Dependent Speaker Verification by Evolutionary Algorithm Using Spline-Defined Design Parameters

Mel frequency cepstral coefficients (MFCCs) have been the most predominantly used spectral features in many a speech-based application. It was primarily introduced to address speech recognition and was later adopted for various other applications such as speaker recognition and emotion recognition. Several findings, in recent times, suggest that Mel-scale filterbank, which is primarily inspired by human perception phenomenon, may not be the most optimum one for speaker recognition. Working in the same direction, this study attempts optimization of filterbank design for text-dependent speaker verification. Motivated by the success of evolutionary computations in the related fields, an evolutionary algorithm is used to carry out this optimization process. This brings into effect data-driven learning of the design parameters and is hypothesized to yield filterbanks which would suit the specific task of speaker-phrase discrimination. The filterbanks have been optimized for the task of text-dependent speaker verification in general, and also for specific cases of speakers and phrases. The proposed filterbank results in relative equal error rate reduction of up to 39.41% with respect to the baseline MFCCs.

The application of Gammatone frequency cepstral coefficients for forensic voice comparison under noisy conditions

ABSTRACT Compared with humans, who have more powerful auditory ability in discriminating and identifying speakers in noisy environments, traditional forensic automatic speaker recognizers do not perform well when dealing with noisy recordings. This paper proposes a GMM-UBM Forensic Automatic Speaker Recognition (FASR) System to reduce the effect of noise on performance. The system uses Gammatone Frequency Cepstral Coefficients (GFCC) based on an auditory periphery model and also incorporates a Principal Component Analysis (PCA) algorithm. The system was tested and validated using Mandarin voice databases compromised with different levels of white noise and office noise. The performance of the system was compared with a baseline system using Mel Frequency Cepstral Coefficients (MFCC) and also PCA under the same conditions. The results show that the performance of the combined GFCC system achieved a substantial improvement when compared with the baseline MFCC system under conditions of a high level of office noise.

Speaker Identification for OFDM-Based Aeronautical Communication System

Although a lot of research has been done on speaker identification in the presence of noise and channel variation, to the best of our knowledge, no work has been reported for aeronautical applications. In this paper, we aim to fulfill this goal by developing a Speaker Identification System (SIS) for future aeronautical communications systems. Furthermore, we present a novel feature extraction scheme based on multi-resolution analysis. The proposed features called SMFCC use Mel Frequency Cepstral Coefficients (MFCCs) features of stationary wavelet transform sub-bands. The extracted features are modeled using the i-vector approach, and support-vector machines are adopted as a back-end classifier. The performance of the proposed SIS is evaluated using two publicly available databases. Comparison of the proposed approach with the baseline MFCC feature extraction shows the feasibility and the robustness of the proposed method. Besides the noise reduction, the identification accuracy is improved by about 12% at higher signal-to-noise ratios and reaches 97.33% as compared to 88.33% using MFCC for ATCOSIM database.

Structure of pauses in speech in the context of speaker verification and classification of speech type

Statistics of pauses appearing in Polish as a potential source of biometry information for automatic speaker recognition were described. The usage of three main types of acoustic pauses (silent, filled and breath pauses) and syntactic pauses (punctuation marks in speech transcripts) was investigated quantitatively in three types of spontaneous speech (presentations, simultaneous interpretation and radio interviews) and read speech (audio books). Selected parameters of pauses extracted for each speaker separately or for speaker groups were examined statistically to verify usefulness of information on pauses for speaker recognition and speaker profile estimation. Quantity and duration of filled pauses, audible breaths, and correlation between the temporal structure of speech and the syntax structure of the spoken language were the features which characterize speakers most. The experiment of using pauses in speaker biometry system (using Universal Background Model and i-vectors) resulted in 30 % equal error rate. Including pause-related features to the baseline Mel-frequency cepstral coefficient system has not significantly improved its performance. In the experiment with automatic recognition of three types of spontaneous speech, we achieved 78 % accuracy, using GMM classifier. Silent pause-related features allowed distinguishing between read and spontaneous speech by extreme gradient boosting with 75 % accuracy.

Channel and noise robustness of articulatory features in a deep neural net based speech recognition system

Articulatory features (AFs) are known to provide an invariant representation of speech, which is expected to be robust against channel and noise degradations. This work presents a deep neural network (DNN)—hidden Markov model (HMM) based acoustic model where articulatory features are used in addition to mel-frequency cepstral coefficients (MFCC) for the Aurora-4 speech recognition task. AFs were generated using a DNN trained layer-by-layer using synthetic speech data. Comparison between baseline mel-filterbank energy (MFB) features, MFCCs and fusion of articulatory feature with MFCCs show that articulatory features helped to increase the noise and channel robustness of the DNN-HMM acoustic model, indicating that articulatory representation does provide an invariant representation of speech.

Wavelet basis selection for enhanced speech parametrization in speaker verification

We study the inherent properties of nine wavelet functions and subsequently evaluate their applicability as basis functions in a speech parametrization scheme that is advantageous for speaker verification. Particularly, the inherent properties of nine candidate basis functions are initially analysed and their advantages and disadvantages are discussed. Subsequently, all candidates are employed in a well-proven speech parametrization scheme, and the resulting speech features are computed. Finally, these speech features are evaluated in a common experimental set-up on the speaker verification task. The experimental results, obtained on two well-known speaker recognition databases, show that the Battle-Lemarie wavelet function is the most advantageous one, among all other functions evaluated here, since it leads to the most beneficial speech descriptors. When compared to the baseline Mel-frequency cepstral coefficients (MFCC), a relative reduction of the equal error rate by 4.2 % was observed on the 2001 NIST speaker recognition evaluation database, and by 2.3 % on the Polycost speaker recognition database.

Direct Recovery of Clean Speech Using a Hybrid Noise Suppression Algorithm for Robust Speech Recognition System

A new log-power domain feature enhancement algorithm named NLPS is developed. It consists of two parts, direct solution of nonlinear system model and log-power subtraction. In contrast to other methods, the proposed algorithm does not need prior speech/noise statistical model. Instead, it works by direct solution of the nonlinear function derived from the speech recognition system. Separate steps are utilized to refine the accuracy of estimated cepstrum by log-power subtraction, which is the second part of the proposed algorithm. The proposed algorithm manages to solve the speech probability distribution function (PDF) discontinuity problem caused by traditional spectral subtraction series algorithms. The effectiveness of the proposed filter is extensively compared using the standard database, AURORA2. The results show that significant improvement can be achieved by incorporating the proposed algorithm. The proposed algorithm reaches a recognition rate of over 86% for noisy speech (average from SNR 0 dB to 20 dB), which means a 48% error reduction over the baseline Mel-frequency Cepstral Coefficient (MFCC) system.

A temporal frequency warped (TFW) 2D psychoacoustic filter for robust speech recognition system

In this paper, a novel hybrid feature extraction algorithm is proposed, which implements forward masking, lateral inhibition, and temporal integration with a simple 2D psychoacoustic filter. The proposed algorithm consists of two key parts, the 2D psychoacoustic filter and cepstral mean variance normalization (CMVN). Mathematical derivation is provided to show the correctness of the 2D psychoacoustic filter based on the characteristic functions of masking effects. The effectiveness of the proposed algorithm is tested on the AURORA2 database. Extensive comparison is made against lateral inhibition (LI), forward masking (FM), CMVN, RASTA filter, the ETSI standard advanced front-end feature extraction algorithm (AFE), and the temporal warped 2D psychoacoustic filter. Experimental results show significant improvements from the proposed algorithm, a relative improvement of nearly 46.78% over the baseline mel-frequency cepstral coefficients (MFCC) system in noisy conditions.

Feature selection using singular value decomposition and QR factorization with column pivoting for text-independent speaker identification

Selection of features is one of the important tasks in the application like Speaker Identification (SI) and other pattern recognition problems. When multiple features are extracted from the same frame of speech, it is expected that a feature vector would contain redundant features. Redundant features confuse the speaker model in multidimensional space resulting in degraded performance by the system. Careful selection of potential features can remove this redundancy while helping to achieve the higher rate of accuracy at lower computational cost. Although the selection of features is difficult without having exhaustive search, this paper proposes an alternative and straight forward technique for feature selection using Singular Value Decomposition (SVD) followed by QR Decomposition with Column Pivoting (QRcp). The idea is to capture the most salient part of the information from the speakers’ data by choosing those features that can explain different dimensions showing minimal similarities (or maximum acoustic variability) among them in orthogonal sense. The performances after selection of features using proposed criterion have been compared with using Mel-frequency Cepstral Coefficients (MFCC), Linear Frequency (LF) Cepstral Coefficients (LFCC) and a new feature proposed in this paper that is based on Gaussian shaped filters on mel-scale. It is shown that proposed SVD-QRcp based feature selection outperforms F-Ratio based method and the proposed feature extraction tool is superior to baseline MFCC & LFCC.

Constructing Modulation Frequency Domain-Based Features for Robust Speech Recognition

Data-driven temporal filtering approaches based on a specific optimization technique have been shown to be capable of enhancing the discrimination and robustness of speech features in speech recognition. The filters in these approaches are often obtained with the statistics of the features in the temporal domain. In this paper, we derive new data-driven temporal filters that employ the statistics of the modulation spectra of the speech features. Three new temporal filtering approaches are proposed and based on constrained versions of linear discriminant analysis (LDA), principal component analysis (PCA), and minimum class distance (MCD), respectively. It is shown that these proposed temporal filters can effectively improve the speech recognition accuracy in various noise-corrupted environments. In experiments conducted on Test Set A of the Aurora-2 noisy digits database, these new temporal filters, together with cepstral mean and variance normalization (CMVN), provide average relative error reduction rates of over 40% and 27% when compared with baseline Mel frequency cepstral coefficient (MFCC) processing and CMVN alone, respectively.

Advances in unsupervised audio classification and segmentation for the broadcast news and NGSW corpora

The problem of unsupervised audio classification and segmentation continues to be a challenging research problem which significantly impacts automatic speech recognition (ASR) and spoken document retrieval (SDR) performance. This paper addresses novel advances in 1) audio classification for speech recognition and 2) audio segmentation for unsupervised multispeaker change detection. A new algorithm is proposed for audio classification, which is based on weighted GMM Networks (WGN). Two new extended-time features: variance of the spectrum flux (VSF) and variance of the zero-crossing rate (VZCR) are used to preclassify the audio and supply weights to the output probabilities of the GMM networks. The classification is then implemented using weighted GMM networks. Since historically there have been no features specifically designed for audio segmentation, we evaluate 16 potential features including three new proposed features: perceptual minimum variance distortionless response (PMVDR), smoothed zero-crossing rate (SZCR), and filterbank log energy coefficients (FBLC) in 14 noisy environments to determine the best robust features on the average across these conditions. Next, a new distance metric, T/sup 2/-mean, is proposed which is intended to improve segmentation for short segment turns (i.e., 1-5 s). A new false alarm compensation procedure is implemented, which can compensate the false alarm rate significantly with little cost to the miss rate. Evaluations on a standard data set-Defense Advanced Research Projects Agency (DARPA) Hub4 Broadcast News 1997 evaluation data-show that the WGN classification algorithm achieves over a 50% improvement versus the GMM network baseline algorithm, and the proposed compound segmentation algorithm achieves 23%-10% improvement in all metrics versus the baseline Mel-frequency cepstral coefficients (MFCC) and traditional Bayesian information criterion (BIC) algorithm. The new classification and segmentation algorithms also obtain very satisfactory results on the more diverse and challenging National Gallery of the Spoken Word (NGSW) corpus.