Gammatone Frequency Cepstral Coefficients Research Articles

Adaptive wavelet thresholding with robust hybrid features for text-independent speaker identification system

The robustness of speaker identification system over additive noise channel is crucial for real-world applications. In speaker identification (SID) systems, the extracted features from each speech frame are an essential factor for building a reliable identification system. For clean environments, the identification system works well; in noisy environments, there is an additive noise, which is affect the system. To eliminate the problem of additive noise and to achieve a high accuracy in speaker identification system a proposed algorithm for feature extraction based on speech enhancement and a combined features is presents. In this paper, a wavelet thresholding pre-processing stage, and feature warping (FW) techniques are used with two combined features named power normalized cepstral coefficients (PNCC) and gammatone frequency cepstral coefficients (GFCC) to improve the identification system robustness against different types of additive noises. Universal Background Model Gaussian Mixture Model (UBM-GMM) is used for features matching between the claim and actual speakers. The results showed performance improvement for the proposed feature extraction algorithm of identification system comparing with conventional features over most types of noises and different SNR ratios.

An Underwater Acoustic Target Recognition Method Based on Restricted Boltzmann Machine.

This article focuses on an underwater acoustic target recognition method based on target radiated noise. The difficulty of underwater acoustic target recognition is mainly the extraction of effective classification features and pattern classification. Traditional feature extraction methods based on Low Frequency Analysis Recording (LOFAR), Mel-Frequency Cepstral Coefficients (MFCC), Gammatone-Frequency Cepstral Coefficients (GFCC), etc. essentially compress data according to a certain pre-set model, artificially discarding part of the information in the data, and often losing information helpful for classification. This paper presents a target recognition method based on feature auto-encoding. This method takes the normalized frequency spectrum of the signal as input, uses a restricted Boltzmann machine to perform unsupervised automatic encoding of the data, extracts the deep data structure layer by layer, and classifies the acquired features through the BP neural network. This method was tested using actual ship radiated noise database, and the results show that proposed classification system has better recognition accuracy and adaptability than the hand-crafted feature extraction based method.

ROBUST HYBRID FEATURES BASED TEXT INDEPENDENT SPEAKER IDENTIFICATION SYSTEM OVER NOISY ADDITIVE CHANNEL

Robustness of speaker identification systems over additive noise is crucial for real-world applications. In this paper, two robust features named Power Normalized Cepstral Coefficients (PNCC) and Gammatone Frequency Cepstral Coefficients (GFCC) are combined together to improve the robustness of speaker identification system over different types of noise. Universal Background Model Gaussian Mixture Model (UBM-GMM) is used as a feature matching and a classifier to identify the claim speakers. Evaluation results show that the proposed hybrid feature improves the performance of identification system when compared to conventional features over most types of noise and different signal-to-noise ratios.

Individual Violin Recognition Method Combining Tonal and Nontonal Features

Individual recognition among instruments of the same type is a challenging problem and it has been rarely investigated. In this study, the individual recognition of violins is explored. Based on the source–filter model, the spectrum can be divided into tonal content and nontonal content, which reflects the timbre from complementary aspects. The tonal/nontonal gammatone frequency cepstral coefficients (GFCC) are combined to describe the corresponding spectrum contents in this study. In the recognition system, Gaussian mixture models–universal background model (GMM–UBM) is employed to parameterize the distribution of the combined features. In order to evaluate the recognition task of violin individuals, a solo dataset including 86 violins is developed in this study. Compared with other features, the combined features show a better performance in both individual violin recognition and violin grade classification. Experimental results also show the GMM–UBM outperforms the CNN, especially when the training data are limited. Finally, the effect of players on the individual violin recognition is investigated.

Mixture linear prediction Gammatone Cepstral features for robust speaker verification under transmission channel noise

In this paper, we present a Mixture Linear Prediction based approach for robust Gammatone Cepstral Coefficients extraction (MLPGCCs). The proposed method provides performance improvement of Automatic Speaker Verification (ASV) using i-vector and Gaussian Probabilistic Linear Discriminant Analysis GPLDA modeling under transmission channel noise. The performance of the extracted MLPGCCs was evaluated using the NIST 2008 database where a single channel microphone recorded conversational speech. The system is analyzed in the presence of different channel transmission noises such as Additive White Gaussian (AWGN) and Rayleigh fading at various Signals to Noise Ratio (SNR) levels. The evaluation results show that the MLPGCCs features are a promising way for the ASV task. Indeed, the speaker verification performance using the MLPGCCs proposed features is significantly improved compared to the conventional Gammatone Frequency Cepstral Coefficients (GFCCs) and Mel Frequency Cepstral Coefficients (MFCCs) features. For speech signals corrupted with AWGN noise at SNRs ranging from (-5 dB to 15 dB), we obtain a significant reduction of the Equal Error Rate (EER) ranging from 9.41% to 6.65% and 3.72% to 1.50%, compared with conventional MFCCs and GFCCs features respectively. In addition, when the test speech signals are corrupted with Rayleigh fading channel we achieve an EER reduction ranging from 23.63% to 7.8% and from 10.88% to 6.8% compared with conventional MFCCs and GFCCs, respectively. We also found that the combination of GFCCs and MLPGCCs gives the highest performance of speaker verification system. The best performance combination achieved is around EER from 0.43% to 0.59% and 1.92% to 3.88%.

Audio content analysis for unobtrusive event detection in smart homes

Environmental sound signals are multi-source, heterogeneous, and varying in time. Many systems have been proposed to process such signals for event detection in ambient assisted living applications. Typically, these systems use feature extraction, selection, and classification. However, despite major advances, several important questions remain unanswered, especially in real-world settings. This paper contributes to the body of knowledge in the field by addressing the following problems for ambient sounds recorded in various real-world kitchen environments: (1) which features and which classifiers are most suitable in the presence of background noise? (2) what is the effect of signal duration on recognition accuracy? (3) how do the signal-to-noise-ratio and the distance between the microphone and the audio source affect the recognition accuracy in an environment in which the system was not trained? We show that for systems that use traditional classifiers, it is beneficial to combine gammatone frequency cepstral coefficients and discrete wavelet transform coefficients and to use a gradient boosting classifier. For systems based on deep learning, we consider 1D and 2D Convolutional Neural Networks (CNN) using mel-spectrogram energies and mel-spectrograms images as inputs, respectively, and show that the 2D CNN outperforms the 1D CNN. We obtained competitive classification results for two such systems. The first one, which uses a gradient boosting classifier, achieved an F1-Score of 90.2% and a recognition accuracy of 91.7%. The second one, which uses a 2D CNN with mel-spectrogram images, achieved an F1-Score of 92.7% and a recognition accuracy of 96%.

A Hybrid Meta-Heuristic Feature Selection Method for Identification of Indian Spoken Languages From Audio Signals

With the recent advancements in the fields of machine learning and artificial intelligence, spoken language identification-based applications have been increasing in terms of the impact they have on the day-to-day lives of common people. Western countries have been enjoying the privilege of spoken language recognition-based applications for a while now, however, they have not gained much popularity in multi-lingual countries like India owing to various complexities. In this paper, we have addressed this issue by attempting to identify different Indian languages based on various well-known features like Mel-Frequency Cepstral Coefficient (MFCC), Linear Prediction Coefficient (LPC), Discrete Wavelet Transform (DWT), Gammatone Frequency Cepstral Coefficient (GFCC) as well as a few deep learning architecture based features like i-vector and x-vector extracted from the audio signals. After comparing the initial results, it is observed that the combination of MFCC and LPC produces the best results. Then we have developed a new nature-inspired feature selection (FS) algorithm by hybridizing Binary Bat Algorithm (BBA) with Late Acceptance Hill-Climbing (LAHC) to select the optimal subset from the said feature vectors in order to reduce the model complexity and help it train faster. Using Random Forest (RF) classifier, we have achieved an accuracy of 92.35% on Indic TTS database developed by IIT-Madras, and an accuracy of 100% on the Indic Speech database developed by the Speech and Vision Laboratory (SVL) IIIT-Hyderabad. The proposed algorithm is also found to outperform many standard meta-heuristic FS algorithms. The source code of this work is available at: https://github.com/CodeChef97dotcom/Feature-Selection

Robust Speaker Recognition Based on Single-Channel and Multi-Channel Speech Enhancement

Deep neural network (DNN) embeddings for speaker recognition have recently attracted much attention. Compared to i-vectors, they are more robust to noise and room reverberation as DNNs leverage large-scale training. This article addresses the question of whether speech enhancement approaches are still useful when DNN embeddings are used for speaker recognition. We investigate single- and multi-channel speech enhancement for text-independent speaker verification based on x-vectors in conditions where strong diffuse noise and reverberation are both present. Single-channel (monaural) speech enhancement is based on complex spectral mapping and is applied to individual microphones. We use masking-based minimum variance distortion-less response (MVDR) beamformer and its rank-1 approximation for multi-channel speech enhancement. We propose a novel method of deriving time-frequency masks from the estimated complex spectrogram. In addition, we investigate gammatone frequency cepstral coefficients (GFCCs) as robust speaker features. Systematic evaluations and comparisons on the NIST SRE 2010 retransmitted corpus show that both monaural and multi-channel speech enhancement significantly outperform x-vector's performance, and our covariance matrix estimate is effective for the MVDR beamformer.

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.

Robustness Speaker Recognition Based on Feature Space in Clean and Noisy Condition

Background & Objective: Speaker Recognition (SR) techniques have been developed into a relatively mature status over the past few decades through development work. Existing methods typically use robust features extracted from clean speech signals, and therefore in idealized conditions can achieve very high recognition accuracy. For critical applications, such as security and forensics, robustness and reliability of the system are crucial. Methods: The background noise and reverberation as often occur in many real-world applications are known to compromise recognition performance. To improve the performance of speaker verification systems, an effective and robust technique is proposed to extract features for speech processing, capable of operating in the clean and noisy condition. Mel Frequency Cepstrum Coefficients (MFCCs) and Gammatone Frequency Cepstral Coefficients (GFCC) are the mature techniques and the most common features, which are used for speaker recognition. MFCCs are calculated from the log energies in frequency bands distributed over a mel scale. While GFCC has been acquired from a bank of Gammatone filters, which was originally suggested to model human cochlear filtering. This paper investigates the performance of GFCC and the conventional MFCC feature in clean and noisy conditions. The effects of the Signal-to-Noise Ratio (SNR) and language mismatch on the system performance have been taken into account in this work. Conclusion: Experimental results have shown significant improvement in system performance in terms of reduced equal error rate and detection error trade-off. Performance in terms of recognition rates under various types of noise, various Signal-to-Noise Ratios (SNRs) was quantified via simulation. Results of the study are also presented and discussed.

Underwater Acoustic Target Recognition: A Combination of Multi-Dimensional Fusion Features and Modified Deep Neural Network

A method with a combination of multi-dimensional fusion features and a modified deep neural network (MFF-MDNN) is proposed to recognize underwater acoustic targets in this paper. Specifically, due to the complex and changeable underwater environment, it is difficult to describe underwater acoustic signals with a single feature. The Gammatone frequency cepstral coefficient (GFCC) and modified empirical mode decomposition (MEMD) are developed to extract multi-dimensional features in this paper. Moreover, to ensure the same time dimension, a dimension reduction method is proposed to obtain multi-dimensional fusion features in the original underwater acoustic signals. Then, to reduce redundant features and further improve recognition accuracy, the Gaussian mixture model (GMM) is used to modify the structure of a deep neural network (DNN). Finally, the proposed underwater acoustic target recognition method can obtain an accuracy of 94.3% under a maximum of 800 iterations when the dataset has underwater background noise with weak targets. Compared with other methods, the recognition results demonstrate that the proposed method has higher accuracy and strong adaptability.

Enhanced Automatic Speech Recognition System Based on Enhancing Power-Normalized Cepstral Coefficients

Many new consumer applications are based on the use of automatic speech recognition (ASR) systems, such as voice command interfaces, speech-to-text applications, and data entry processes. Although ASR systems have remarkably improved in recent decades, the speech recognition system performance still significantly degrades in the presence of noisy environments. Developing a robust ASR system that can work in real-world noise and other acoustic distorting conditions is an attractive research topic. Many advanced algorithms have been developed in the literature to deal with this problem; most of these algorithms are based on modeling the behavior of the human auditory system with perceived noisy speech. In this research, the power-normalized cepstral coefficient (PNCC) system is modified to increase robustness against the different types of environmental noises, where a new technique based on gammatone channel filtering combined with channel bias minimization is used to suppress the noise effects. The TIDIGITS database is utilized to evaluate the performance of the proposed system in comparison to the state-of-the-art techniques in the presence of additive white Gaussian noise (AWGN) and seven different types of environmental noises. In this research, one word is recognized from a set containing 11 possibilities only. The experimental results showed that the proposed method provides significant improvements in the recognition accuracy at low signal to noise ratios (SNR). In the case of subway noise at SNR = 5 dB, the proposed method outperforms the mel-frequency cepstral coefficient (MFCC) and relative spectral (RASTA)–perceptual linear predictive (PLP) methods by 55% and 47%, respectively. Moreover, the recognition rate of the proposed method is higher than the gammatone frequency cepstral coefficient (GFCC) and PNCC methods in the case of car noise. It is enhanced by 40% in comparison to the GFCC method at SNR 0dB, while it is improved by 20% in comparison to the PNCC method at SNR −5dB.

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.

Native Language Identification in Very Short Utterances Using Bidirectional Long Short-Term Memory Network

Native language identification (NLI) is the task of identifying the first language of a user based on their speech or written text in a second language. In this paper, we propose the use of spectrogram- and cochleagram-based features extracted from very short speech utterances (0.8 s on average) to infer the native language of an Urdu speaker. The bidirectional long short-term memory (BLSTM) neural networks are adopted for the classification of utterances among the native languages. A set of experiments is carried out for the network architecture search and the system’s accuracy is evaluated on the validation data set. Overall accuracy of 74.81% and 71.61% is achieved using the Mel-frequency cepstral coefficients (MFCC) and Gammatone frequency cepstral coefficients (GFCC), respectively. Moreover, the optimized MFCC feature-based BLSTM network and GFCC feature-based BLSTM network are merged together to take advantage of both the feature sets. The experiments show that the performance of the merged network surpasses the individual BLSTM networks and accuracy of 75.69% is achieved on the evaluation data. The effect of test data duration is also analyzed (from 0.27 s to 1.5 s); in addition, it is observed that with very short duration as 0.4 s, an accuracy of over 50% can be achieved.

An investigation on the degradation of different features extracted from the compressed American English speech using narrowband and wideband codecs

Speech coding facilitates speech compression without perceptual loss that results in the elimination or deterioration of both speech and speaker specific features used for a wide range of applications like automatic speaker and speech recognition, biometric authentication, prosody evaluations etc. The present work investigates the effect of speech coding in the quality of features which include Mel Frequency Cepstral Coefficients, Gammatone Frequency Cepstral Coefficients, Power-Normalized Cepstral Coefficients, Perceptual Linear Prediction Cepstral Coefficients, Rasta-Perceptual Linear Prediction Cepstral Coefficients, Residue Cepstrum Coefficients and Linear Predictive Coding-derived cepstral coefficients extracted from codec compressed speech. The codecs selected for this study are G.711, G.729, G.722.2, Enhanced Voice Services, Mixed Excitation Linear Prediction and also three codecs based on compressive sensing frame work. The analysis also includes the variation in the quality of extracted features with various bit-rates supported by Enhanced Voice Services, G.722.2 and compressive sensing codecs. The quality analysis of extracted epochs, fundamental frequency and formants estimated from codec compressed speech was also performed here. In the case of various features extracted from the output of selected codecs, the variation introduced by Mixed Excitation Linear Prediction codec is the least due to its unique method for the representation of excitation. In the case of compressive sensing based codecs, there is a drastic improvement in the quality of extracted features with the augmentation of bit rate due to the waveform type coding used in compressive sensing based codecs. For the most popular Code Excited Linear Prediction codec based on Analysis-by-Synthesis coding paradigm, the impact of Linear Predictive Coding order in feature extraction is investigated. There is an improvement in the quality of extracted features with the order of linear prediction and the optimum performance is obtained for Linear Predictive Coding order between 20 and 30, and this varies with gender and statistical characteristics of speech. Even though the basic motive of a codec is to compress single voice source, the performance of codecs in multi speaker environment is also studied, which is the most common environment in majority of the speech processing applications. Here, the multi speaker environment with two speakers is considered and there is an augmentation in the quality of individual speeches with increase in diversity of mixtures that are passed through codecs. The perceptual quality of individual speeches extracted from the codec compressed speech is almost same for both Mixed Excitation Linear Prediction and Enhanced Voice Services codecs but regarding the preservation of features, the Mixed Excitation Linear Prediction codec has shown a superior performance over Enhanced Voice Services codec.

Optimizing Integrated Features for Hindi Automatic Speech Recognition System

Abstract An automatic speech recognition (ASR) system translates spoken words or utterances (isolated, connected, continuous, and spontaneous) into text format. State-of-the-art ASR systems mainly use Mel frequency (MF) cepstral coefficient (MFCC), perceptual linear prediction (PLP), and Gammatone frequency (GF) cepstral coefficient (GFCC) for extracting features in the training phase of the ASR system. Initially, the paper proposes a sequential combination of all three feature extraction methods, taking two at a time. Six combinations, MF-PLP, PLP-MFCC, MF-GFCC, GF-MFCC, GF-PLP, and PLP-GFCC, are used, and the accuracy of the proposed system using all these combinations was tested. The results show that the GF-MFCC and MF-GFCC integrations outperform all other proposed integrations. Further, these two feature vector integrations are optimized using three different optimization methods, particle swarm optimization (PSO), PSO with crossover, and PSO with quadratic crossover (Q-PSO). The results demonstrate that the Q-PSO-optimized GF-MFCC integration show significant improvement over all other optimized combinations.

Visual Speech Recognition Using Optical Flow and Hidden Markov Model

The present work proposes audio-visual speech recognition with the use of Gammatone frequency cepstral coefficient (GFCC) and optical flow (OF) features with Hindi speech database. The OF refers to the distribution of apparent velocities of brightness pattern movements in an image. In this technique, OF is determined without extracting the location and contours of pair of lips of individual speaker. The visual features as horizontal component and vertical components of flow velocities have been calculated. Furthermore, the visual features are combined with audio features using early integration method followed by classification using hidden Markov model. The isolated Hindi digits were evaluated for their recognition performance using GFCC features not only in clean environment but also tested under noisy environment and compared with existing Mel frequency cepstral coefficient (MFCC) features. The GFCC shows almost comparable result with MFCC in clean environment; however, its performance goes down in noisy environment. Futhermore, the visual features obtained by the OF analysis when combine with GFCC audio features give significant improvement of ~ 12%, ~ 12%, and ~ 14% at different SNRs (5 dB, 10 dB, and 20 dB, respectively) in recognition performance under noisy environment.

Dereverberation binaural source separation using deep learning

This paper reported a deep-learning based binaural separation using gammatone-frequency cepstral coefficient (GFCC) and multi-resolution cochleagram (MRCG) as spectral features and interaural time difference (ITD) and interaural level difference (ILD) as spatial features. A binary mask was estimated by deep neural network (DNN) binary classifier that used the features as a training data and ideal ratio mask (IRM) as a training target. In the experiment, a male speaker as a target speech at azimuth 0o and a female speaker as a masker speech at azimuth 30o, 20o, 10o, and 5o in rooms with 0.32, 0.47, 0.68, and 0.89 s reverberation time (RT60). As a training process, 50 mixtures were used for each condition experiment. The classifier contained two hidden layers, 200 binary neurons and 50 epoch, and Restricted Boltzmann Machine (RBM) was used as pre-training process. The RBM and the classifier learning rate was 1 up to 0.001 from epoch 1 to epoch 50. The sound quality results indicated by 88% intelligibility of STOI method which means that the separated sound is easily understood by the listener. The MOS value is 2.8 which means the sentence is clear but spectral distortion is slightly annoying.

Detecting Manic State of Bipolar Disorder Based on Support Vector Machine and Gaussian Mixture Model Using Spontaneous Speech.

ObjectiveThis study was aimed to compare the accuracy of Support Vector Machine (SVM) and Gaussian Mixture Model (GMM) in the detection of manic state of bipolar disorders (BD) of single patients and multiple patients. Methods21 hospitalized BD patients (14 females, average age 34.5±15.3) were recruited after admission. Spontaneous speech was collected through a preloaded smartphone. Firstly, speech features [pitch, formants, mel-frequency cepstrum coefficients (MFCC), linear prediction cepstral coefficient (LPCC), gamma-tone frequency cepstral coefficients (GFCC) etc.] were preprocessed and extracted. Then, speech features were selected using the features of between-class variance and within-class variance. The manic state of patients was then detected by SVM and GMM methods. ResultsLPCC demonstrated the best discrimination efficiency. The accuracy of manic state detection for single patients was much better using SVM method than GMM method. The detection accuracy for multiple patients was higher using GMM method than SVM method. ConclusionSVM provided an appropriate tool for detecting manic state for single patients, whereas GMM worked better for multiple patients’ manic state detection. Both of them could help doctors and patients for better diagnosis and mood state monitoring in different situations.

A Multiobjective Learning and Ensembling Approach to High-Performance Speech Enhancement With Compact Neural Network Architectures

In this study, we propose a novel deep neural network (DNN) architecture for speech enhancement (SE) via a multiobjective learning and ensembling (MOLE) framework to achieve a compact and lowlatency design, while maintaining good performance in quality evaluations. MOLE follows the boosting concept when combining weak models into a strong classifier and consists of two compact DNNs. The first, called the multiobjective learning DNN (MOL-DNN), takes multiple features, such as log-power spectra (LPS), mel-frequency cepstral coefficients (MFCCs) and Gammatone frequency cepstral coefficients (GFCCs) to predict a multiobjective set that includes clean speech feature, dynamic noise feature, and ideal ratio mask (IRM). The second, called the multiobjective ensembling DNN (MOE-DNN), takes the learned features from MOL-DNN as inputs and separately predicts clean LPS and IRM, clean MFCC and IRM, and clean GFCC and IRM using three sets of weak regression functions. Finally, a postprocessing operation can be applied to the estimated clean features by leveraging the multiple targets learned from both the MOL-DNN and the MOE-DNN. On speech corrupted by 15 noise types not seen in model training the SE results show that the MOLE approach, which features a small model size and low run-time latency, can achieve consistent improvements over both DNN- and long short-term memory (LSTM)-based techniques in terms of all the objective metrics evaluated in this study for all three cases (the input contexts contain 1-frame, 4-frame and 7-frame instances). The 1-frame MOLE-based SE system outperforms the DNN-based SE system with a 7-frame input expansion at a 3-frame delay and also achieves better performance than the LSTM-based SE system with 4-frame, no delay expansion by including only 3 previous frames, and with 170 times less processing latency.