Perceptual Linear Prediction Features Research Articles

Wind Sounds Classification Using Different Audio Feature Extraction Techniques

In this research, Different audio feature extraction technique are implemented and classification approaches are presented to classify seven types of wind. Where we applied features technique such as Zero Crossing Rate (ZCR) ,Fast Fourier Transformation (FFT), Linear predictive coding (LPC), Perceptual Linear Prediction (PLP). We know that some of these methods are good with human voices, but we tried to apply them here to characterize the wind audio content. The CNN classification method is implemented to determine the class of input wind sound signal. Experimental results show that each of these extraction feature methods are gave different results, but classification accuracy that are obtained of PLP features proven to have the best results.

Performance Analysis of various Front-end and Back End Amalgamations for Noise-robust DNN-based ASR

Introduction: An Automatic Speech Recognition (ASR) system enables to recognize the speech utterances and thus can be used to convert speech into text for various purposes. These systems are deployed in different environments such as clean or noisy and are used by all ages or types of people. These also present some of the major difficulties faced in the development of an ASR system. Thus, an ASR system need to be efficient, while also being accurate and robust. Our main goal is to minimize the error rate during training as well as testing phases, while implementing an ASR system. Performance of ASR depends upon different combinations of feature extraction techniques and back-end techniques. In this paper, using a continuous speech recognition system, the performance comparison of different combinations of feature extraction techniques and various types of back-end techniques has been presented Methods: Hidden Markov Models (HMMs), Subspace Gaussian Mixture Models (SGMMs) and Deep Neural Networks (DNNs) with DNN-HMM architecture, namely Karel’s, Dan’s and Hybrid DNN-SGMM architecture are used at the back-end of the implemented system. Mel frequency Cepstral Coefficient (MFCC), Perceptual Linear Prediction (PLP), and Gammatone Frequency Cepstral coefficients (GFCC) are used as feature extraction techniques at the front-end of the proposed system. Kaldi toolkit has been used for the implementation of the proposed work. The system is trained on the Texas Instruments-Massachusetts Institute of Technology (TIMIT) speech corpus for English language Results: The experimental results show that MFCC outperforms GFCC and PLP in noiseless conditions, while PLP tends to outperform MFCC and GFCC in noisy conditions. Furthermore, the hybrid of Dan’s DNN implementation along with SGMM performs the best for the back-end acoustic modeling. The proposed architecture with PLP feature extraction technique in the front end and hybrid of Dan’s DNN implementation along with SGMM at the back end outperforms the other combinations in a noisy environment. Conclusion: Automatic Speech recognition has numerous applications in our lives like Home automation, Personal assistant, Robotics etc. It is highly desirable to build an ASR system with good performance. The performance Automatic Speech Recognition is affected by various factors which include vocabulary size, whether system is speaker dependent or independent, whether speech is isolated, discontinuous or continuous, adverse conditions like noise. The paper presented an ensemble architecture that uses PLP for feature extraction at the front end and a hybrid of SGMM + Dan’s DNN in the backend to build a noise robust ASR system Discussion: The presented work in this paper discusses the performance comparison of continuous ASR systems developed using different combinations of front-end feature extraction (MFCC, PLP, and GFCC) and back-end acoustic modeling (mono-phone, tri-phone, SGMM, DNN and hybrid DNN-SGMM) techniques. Each type of front-end technique is tested in combination with each type of back-end technique. Finally, it compares the results of the combinations thus formed, to find out the best performing combination in noisy and clean conditions

Deep neural network architectures for dysarthric speech analysis and recognition

This paper investigates the ability of deep neural networks (DNNs) to improve the automatic recognition of dysarthric speech through the use of convolutional neural networks (CNNs) and long short-term memory (LSTM) neural networks. Dysarthria is one of the most common speech communication disorders associated with neurological impairments that can drastically reduce the intelligibility of speech. The aim of the present study is twofold. First, it compares three different input features for training and testing dysarthric speech recognition systems. These features are the mel-frequency cepstral coefficients (MFCCs), mel-frequency spectral coefficients (MFSCs), and the perceptual linear prediction features (PLPs). Second, the performance of the CNN- and LSTM-based architectures is compared against a state-of-the-art baseline system based on hidden Markov models (HMMs) and Gaussian mixture models (GMMs) to determine the best dysarthric speech recognizer. Experimental results show that the CNN-based system using perceptual linear prediction features provides a recognition rate that can reach 82%, which constitutes relative improvement of 11% and 32% when compared to the performance of LSTM- and GMM-HMM-based systems, respectively.

A Language Identification System using Hybrid Features and Back-Propagation Neural Network

Language Identification (LID) is accurate identification of the unknown language by comparison of speech biometrics of test speech sample and language models accumulated beforehand. This paper presents and encourages the use of hybrid robust feature extraction techniques for spoken language identification (LID) system. In the feature extraction stage, different techniques are applied individually such as Mel frequency cepstral coefficients (MFCCs), perceptual linear prediction features (PLP), relative perceptual linear prediction features (RASTA-PLP). Later, performance of our LID system based on several combinations of the different features (hybrid features) are investigated such as MFCC, PLP, combined with their 1st order derivatives, MFCC + RASTA-PLP, MFCC + SDC (Shifted delta cepstral coefficients). Language identification phase or classification utilizes feed forward back-propagation neural network (FFBPNN) and comparison is based on two learning algorithms: the Levenberg–Marquardt “trainlm” and the scaled conjugate gradient “trainscg”. A comparative analysis in terms of performance is done between different hybrid feature extraction techniques and their individual counterparts. Results clearly indicates that improved performance is obtained with hybrid features with “trainlm” learning algorithm as compared to their individual counterparts. The results are very promising with MFCC-RASTA-PLP hybrid feature extraction technique in comparison to the other hybrid feature extraction techniques with overall accuracy of 94.6% and a minimum test error rate of 0.10. The efficiency of proposed hybrid approaches is determined by simulating several experiments on a user defined language database of speech signals in the working platform of MATLAB.

Analysis and Detection of Pathological Voice Using Glottal Source Features

Automatic detection of voice pathology enables objective assessment and earlier intervention for the diagnosis. This study provides a systematic analysis of glottal source features and investigates their effectiveness in voice pathology detection. Glottal source features are extracted using glottal flows estimated with the quasi-closed phase (QCP) glottal inverse filtering method, using approximate glottal source signals computed with the zero frequency filtering (ZFF) method, and using acoustic voice signals directly. In addition, we propose to derive mel-frequency cepstral coefficients (MFCCs) from the glottal source waveforms computed by QCP and ZFF to effectively capture the variations in glottal source spectra of pathological voice. Experiments were carried out using two databases, the Hospital Universitario Príncipe de Asturias (HUPA) database and the Saarbrücken Voice Disorders (SVD) database. Analysis of features revealed that the glottal source contains information that discriminates normal and pathological voice. Pathology detection experiments were carried out using support vector machine (SVM). From the detection experiments it was observed that the performance achieved with the studied glottal source features is comparable or better than that of conventional MFCCs and perceptual linear prediction (PLP) features. The best detection performance was achieved when the glottal source features were combined with the conventional MFCCs and PLP features, which indicates the complementary nature of the features.

Khasi Speech Recognition using Hidden Markov Model with Different Spectral Features: A Comparison

In this work, various feature extraction techniques are implemented for a speech data corpus in Khasi Language. This paper provides a comparative performance analysis of widely used spectral features like Mel-Frequency Cepstrum Coefficients (MFCC), Perceptual Linear Prediction (PLP), Linear Predictive Cepstrum Coefficient (LPCC), Linear Predictive Coefficient (LPC) and Linear Predictive Reflection Coefficients (LPREFC). For each features we have used coefficient dimensions of static, delta and acceleration evaluated for the recognition of words spoken in Khasi language, a language spoken in the Khasi and Jaintia hills districts of Meghalaya. A recorder based continuous speech database has been collected as per the standard procedure and data processing was done using wave surfer. From the analysis, it was observed that for sampling rate of 16 KHz and 8 KHz the word recognition accuracy of MFCC and PLP features are better than LPCC, LPC and LPREFC features under all circumstances.

Vocal Tract Length Normalization using a Gaussian mixture model framework for query-by-example spoken term detection

A speech spectrum is known to be changed by the variations in the length of the vocal tract of a speaker. This is because of the fact that speech formants are inversely related to the vocal tract length (VTL). The process of compensating spectral variation due to the length of the vocal tract is known as Vocal Tract Length Normalization (VTLN). VTLN is a very important speaker normalization technique for speech recognition and related tasks. In this paper, we used Gaussian Posteriorgram (GP) of VTL-warped spectral features for a Query-by-Example Spoken Term Detection (QbE-STD) task. This paper presents the use of a Gaussian Mixture Model (GMM) framework for VTLN warping factor estimation. In particular, the presented GMM framework does not require phoneme-level transcription. We observed the correlation between the VTLN warping factor estimates obtained via a supervised HMM-based approach and an unsupervised GMM-based approach. In addition, a phoneme recognition and speaker de-identification tasks were conducted using GMM-based VTLN warping factor estimates. For QbE-STD, we considered three spectral features, namely, Mel Frequency Cepstral Coefficients (MFCC), Perceptual Linear Prediction (PLP), and MFCC-TMP (which uses Teager Energy Operator (TEO) to exploit implicitly magnitude and phase information in the MFCC framework). Linear frequency scaling variations for VTLN warping factor are incorporated into these three cepstral representations for the QbE-STD task. Similarly, VTL-warped Gaussian posteriorgram improved the Maximum Term Weighted Value by 0.021 (i.e., 2.1%), and 0.015 (i.e., 1.5%), for MFCC and PLP feature sets, respectively, on the evaluation set of the MediaEval SWS 2013 corpus. The better performance is primarily due to VTLN warping factor estimation using unsupervised GMM framework. Finally, the effectiveness of the proposed VTL-warped GP is presented to rescore using various detection sources, such as depth of detection valley, Self-Similarity Matrix, Pseudo Relevance Feedback and weighted mean features.

Continuous Punjabi speech recognition model based on Kaldi ASR toolkit

In this paper, continuous Punjabi speech recognition model is presented using Kaldi toolkit. For speech recognition, the extraction of Mel frequency cepstral coefficients (MFCC) features and perceptual linear prediction (PLP) features were extracted from Punjabi continuous speech samples. The performance of automatic speech recognition (ASR) system for both monophone and triphone model i.e., tri1, tri2 and tri3 model using N-gram language model is reported. The performance of ASR system were computed in terms of word error rate (WER). A significant reduction in WER was observed using the tri phone model over mono phone model ASR .Also the performance of ASR using tri3 model is improved over tri2 model and the performance of tri2 model is improved over tri1 model ASR. Further, it was found that MFCC feature provides higher speech recognition accuracy than PLP features for continuous Punjabi speech.

Long-Term Critical Band Energy-Based Feature Set for Dialect Identification Using a Neuro-Fuzzy Approach

Dialects of languages demonstrate dependency on both speaker and sound-unit (phone)-related information, which encompasses the problem of dialect identification (DID) under the domain of language identification (LID). The DID task is more complicated than conventional LID, and it has been established that conventional acoustic features like perceptual linear prediction (PLP) and mel frequency cepstral coefficient (MFCC) features-which carry only phone-unit-related information-are not sufficient to address the problem of DID. The authors explore raw log critical band energy (LCBE) information obtained from critical band analysis of speech signals, which effectively carries both speaker and phone-unit-related information. A nonlinear feature extractor using multilayer perceptron (MLP) is designed to model the critical band information. Further, a neuro-fuzzy classifier (NFC) is configured to classify feature vectors into different dialectal classes to discriminate between finer variations. The objective is to investigate perceptually oriented information obtained from all critical bands to distinguish dialectal speech and the applicability of NFC for such problems. Experimental results are shown in terms of classification accuracy of four dialects of Assamese language, mostly spoken in Northeast India. A few baseline systems are developed using PLP and MFCC features along with a Gaussian Mixture Model (GMM)-based classifier. Experimental results prove the strength of the MLP-based nonlinear mapping of critical band information for dialect discrimination compared to the PLP-based autoregressive approximation and MFCC-based cepstral domain version of critical band energy.

Multitaper perceptual linear prediction features of voice samples to discriminate healthy persons from early stage Parkinson diseased persons

The performance of multitaper perceptual linear prediction (PLP) features of speech samples to discriminate healthy and early stage Parkinson diseased subjects is investigated in this paper. The PLP features are conventionally obtained by computing the power spectrum using a single tapered Hamming window. This estimated spectrum exhibits large variance which can be reduced by computing the weighted average of power spectra obtained using a set of tapered windows, leading to multitaper spectral estimation. In this investigation, two multitaper techniques namely Sine wave taper and Thomson multitaper along with the conventional single taper windowing are investigated. Artificial Neural network is then used to classify the PLP features extracted by applying the three types of window tapers on the speech signals of healthy and early stage Parkinson affected people and their respective performances are compared. The results show more accuracy using the multitaper techniques when compared with the conventional single taper technique. It is seen that the accuracy obtained using Sine wave tapers as well as Thomson multitaper is maximum for five tapers. An improvement in the recognition accuracy by 7.5% using the Sine tapers and by 6.9% using the Thomson tapers is obtained when compared with the conventional method. An improvement in other performance measures like Equal error rate, False positive rate, False negative rate, Sensitivity and Specificity is also observed in the multitaper techniques.

Emotion recognition using MLP and GMM for Oriya language

Emotion recognition of human beings is one of the major challenges in the modern complicated world of political and criminal scenario. In this paper an attempt is taken to recognise two classes of speech emotions as high arousal like angry, surprise and low arousal like sad and bore. Linear prediction coefficients (LPC), Mel-frequency cepstral coefficient (MFCC) and perceptual linear prediction (PLP) features are used for emotions recognition using multilayer perceptron (MLP) and Gaussian mixture model (GMM) classifier. Two different databases of four emotions, one of five children and other one of a professional actor has been used in this work. Emotion recognition performance of LPC, PLP and MFCC features has been compared with two classifiers, MLP and GMM. MFCC features with MLP classifier and PLP features with GMM classifier has performed best in their respective categories.

Selection of Features for Emotion Recognition from Speech

Background/Objective: Speech is one of the modes for Human Computer Interface (HCI). Speech contains message to convey as well as the speaker characteristics such as speaker identity and emotional state of the speaker. Recently, researchers are taking more interest in the emotional parameters of speech signals which helps to improve the functionality of HCI. This research focus on selecting features which helps to identify the emotion of the speaker. Methods/Statistical Analysis: Mel Frequency Cepstrum Coefficient (MFCC), Linear Prediction Cepstrum Coefficient (LPCC) and Perceptual Linear Predictive (PLP) methods are used to extract the features. Each emotion is modeled as one Hidden Markov Model (HMM) using Hidden Markov Tool Kit (HTK tool kit). The Beagle Bone Black (BBB) board is chosen for the implementation because of the form factor. Findings: The results indicate that MFCC features gives 100% accuracy for surprise emotion, PLP features gives 100% accuracy for anger emotion and LPCC features give 100% accuracy for fear emotion. Conclusion/Improvement: A hybrid feature extraction method should be devised to detect all emotions with 100% accuracy.

Speaker Adaptation on Hidden Markov Model using MFCC and Rasta-PLP and Comparative Study

This work compares the performance of the Mel-Frequency Cepstral Coefficient (MFCC) and Perceptual Linear Prediction (PLP) features for developing a text-dependent speaker identification system. Continuously spoken Hindi speech sentences have been used to train the HMM models using HTK toolkit for each speaker separately. The experiments have been performed using a set of 200 continuously spoken sentences with vocabulary of 20000 isolated words using a database of 100 speakers. The results show an accuracy of 92.26% recognition when PLP features have been used and accuracy of 91.18% for MFCC features. A confusion matrix has been created for all the 20 test speakers based on the recognition scores obtained for each of these speakers and their confusion with other speakers. Performance has been compared in the closed set and open set conditions of testing and as it is expected, the performance in the closed set condition is far better than the open set. We propose that if PLP features are used in place of MFCC, they may provide improvement in speaker identification accuracy by reducing the cases of false acceptance.

Applying Bottle Neck Feature for Vietnamese speech recognition

In the paper, the basic idea of Bottle Neck Feature (BNF) and the process how to extract BNF are presented. We apply BNF for Vietnamese speech recognition with five layers MLP network of different sizes for the first hidden layer. Input features to extract BNF feature are Perceptual Linear Prediction (PLP) and Mel Frequency Cepstral Coefficient (MFCC). The experiments are carried out on a data set of VOV (Voice of Vietnam). The results show that using BNF for Vietnamese speech recognition, a WER (Word Error Rate) is improved up to 6-7% comparing to the baseline system, and MFCC feature gives a better result than PLP feature.

Multitaper MFCC and PLP features for speaker verification using i-vectors

In this paper we study the performance of the low-variance multi-taper Mel-frequency cepstral coefficient (MFCC) and perceptual linear prediction (PLP) features in a state-of-the-art i-vector speaker verification system. The MFCC and PLP features are usually computed from a Hamming-windowed periodogram spectrum estimate. Such a single-tapered spectrum estimate has large variance, which can be reduced by averaging spectral estimates obtained using a set of different tapers, leading to a so-called multi-taper spectral estimate. The multi-taper spectrum estimation method has proven to be powerful especially when the spectrum of interest has a large dynamic range or varies rapidly. Multi-taper MFCC features were also recently studied in speaker verification with promising preliminary results. In this study our primary goal is to validate those findings using an up-to-date i-vector classifier on the latest NIST 2010 SRE data. In addition, we also propose to compute robust perceptual linear prediction (PLP) features using multitapers. Furthermore, we provide a detailed comparison between different taper weight selections in the Thomson multi-taper method in the context of speaker verification. Speaker verification results on the telephone (det5) and microphone speech (det1, det2, det3 and det4) of the latest NIST 2010 SRE corpus indicate that the multi-taper methods outperform the conventional periodogram technique. Instead of simply averaging (using uniform weights) the individual spectral estimates in forming the multi-taper estimate, weighted averaging (using non-uniform weights) improves performance. Compared to the MFCC and PLP baseline systems, the sine-weighted cepstrum estimator (SWCE) based multitaper method provides average relative reductions of 12.3% and 7.5% in equal error rate, respectively. For the multi-peak multi-taper method, the corresponding reductions are 12.6% and 11.6%, respectively. Finally, the Thomson multi-taper method provides error reductions of 9.5% and 5.0% in EER for MFCC and PLP features, respectively. We conclude that both the MFCC and PLP features computed via multitapers provide systematic improvements in recognition accuracy.

Robust speech detection in real acoustic backgrounds with perceptually motivated features

The current study presents an analysis of the robustness of a speech detector in real background sounds. One of the most important aspects of automatic speech/nonspeech classification is robustness in the presence of strongly varying external conditions. These include variations of the signal-to-noise ratio as well as fluctuations of the background noise. These variations are systematically evaluated by choosing different mismatched conditions between training and testing of the speech/nonspeech classifiers. The detection performance of the classifier with respect to these mismatched conditions is used as a measure of robustness and generalisation. The generalisation towards un-trained SNR conditions and unknown background noises is evaluated and compared to a matched baseline condition. The classifier consists of a feature front-end, which computes amplitude modulation spectral features (AMS), and a support vector machine (SVM) back-end. The AMS features are based on Fourier decomposition over time of short-term spectrograms. Mel-frequency cepstral coefficients (MFCC) as well as relative spectral features (RASTA) based on perceptual linear prediction (PLP) serve as baseline. The results show that RASTA-filtered PLP features perform best in the matched task. In the generalisation tasks however, the AMS features emerge as more robust in most cases, while MFCC features are outperformed by both other feature types. In a second set of experiments, a hierarchical approach is analysed which employs a background classification step prior to the speech/nonspeech classifier in order to improve the robustness of the detection scores in novel backgrounds. The background sounds used are recorded in typical everyday scenarios. The hierarchy provides a benefit in overall performance if the robust AMS features are employed. The generalisation capabilities of the hierarchy towards novel backgrounds and SNRs is found to be optimal when a limited number of training backgrounds is used (compared to the inclusion of all available background data). The best backgrounds in terms of generalisation capabilities are found to be backgrounds in which some component of speech (such as unintelligible background babble) is present, which corroborates the hypothesis that the AMS features provide a decomposition of signals which is by itself very suitable for training very general speech/nonspeech detectors. This is also supported by the finding that the SVMs combined with RASTA-PLPs require nonlinear kernels to reach a similar performance as the AMS patterns with linear kernels.

Generalized perceptual linear prediction features for animal vocalization analysis

A new feature extraction model, generalized perceptual linear prediction (gPLP), is developed to calculate a set of perceptually relevant features for digital signal analysis of animal vocalizations. The gPLP model is a generalized adaptation of the perceptual linear prediction model, popular in human speech processing, which incorporates perceptual information such as frequency warping and equal loudness normalization into the feature extraction process. Since such perceptual information is available for a number of animal species, this new approach integrates that information into a generalized model to extract perceptually relevant features for a particular species. To illustrate, qualitative and quantitative comparisons are made between the species-specific model, generalized perceptual linear prediction (gPLP), and the original PLP model using a set of vocalizations collected from captive African elephants (Loxodonta africana) and wild beluga whales (Delphinapterus leucas). The models that incorporate perceptional information outperform the original human-based models in both visualization and classification tasks.

Exploitation of spectral variance to improve robustness in speech recognition

With the aim of improving noise robustness of speech recognition an approach that exploits the variance information in spectral sub-bands is presented. The variance based features are used in combination with the normally used Mel-frequency cepstral coefficients (MFCC), and experimental results show that the combined features outperform MFCC alone, perceptual linear prediction features and entropy based features.