LPC Coefficients Research Articles

The Role of Selected Speech Signal Characteristics in Discriminating Unipolar and Bipolar Disorders.

The objective of this study is to explore and enhance the diagnostic process of unipolar and bipolar disorders. The primary focus is on leveraging automated processes to improve the accuracy and accessibility of diagnosis. The study aims to introduce an audio corpus collected from patients diagnosed with these disorders, annotated using the Clinical Global Impressions Scale (CGI) by psychiatrists. Traditional diagnostic methods rely on the clinician's expertise and consideration of co-existing mental disorders. However, this study proposes the implementation of automated processes in the diagnosis, providing quantitative measures and enabling prolonged observation of patients. The paper introduces a speech signal pipeline for CGI state classification, with a specific focus on selecting the most discriminative features. Acoustic features such as prosodies, MFCC, and LPC coefficients are examined in the study. The classification process utilizes common machine learning methods. The results of the study indicate promising outcomes for the automated diagnosis of bipolar and unipolar disorders using the proposed speech signal pipeline. The audio corpus annotated with CGI by psychiatrists achieved a classification accuracy of 95% for the two-class classification. For the four- and seven-class classifications, the results were 77.3% and 73%, respectively, demonstrating the potential of the developed method in distinguishing different states of the disorders.

Comparison of different Classification techniques for the detection of Speech affected due to Respiratory disorders

Respiratory disorders are one of the common chronic disorders across the world. Any malfunctioning in the respiration process affects speech as speech and respiration go hand in hand. Analysis of speech parameter variation due to respiratory disorders is a prime research area in today’s scenario. In this paper, an efficient approach using a machine learning paradigm was used to detect the speech affected due to respiratory disorders. Various speech parameters were extracted (F1, F2, F3… etc.) using PRAAT software along with MFCC and LPC coefficients. Statistically, significant features were obtained to determine the predominant parameters. The extracted features were applied to different machine learning paradigms then different classification techniques were analyzed and compared to distinguish between normal and affected speech.5-fold,10-fold cross-validation, and hold-out data division protocols were used for evaluating the classification results. To evaluate the performance of the proposed methodology, Accuracy, sensitivity, specificity, and the area under receiver operating characteristics (AUC) were used. The result demonstrates that the speech parameters (F1, F2, F3…etc.) extraction method and the LPC method achieved a significantly higher classification accuracy, sensitivity, and specificity of 100% and AUC of 1, under the holdout method. While MFCC method achieved the highest classification accuracy of 83%, sensitivity of 100%, specificity of 67%, and AUC of 0.83.

SPEECH RECOGNITION APPLICATION AS AN ANIMATED OBJECT MOVEMENT CONTROLLER SYSTEM

Technological developments in the world have no boundaries. One of them is Speech Recognition. At first, words spoken by humans cannot be recognized by computers. To be recognizable, the word is processed using a specific method. Linear Predictive Coding Method (LPC) is a method used in this research to extract the characteristics of speech. The result of the LPC method is the LPC coefficient which is the number of LPC orders plus 1. The LPC coefficient is processed using Fast Fourier Transform (FFT) 512 to simplify the process of speech recognition. The results are then trained using Backpropagation Neural Network (BPNN) to recognize the spoken word. Speech recognition on the program is implemented as an animated object motion controller on the computer. The end result of this research is animated objects move in accordance with the spoken word. The optimal BPNN structure in this research is to use traingda training function, number of nodes 3, learning rate 0.05, epoch 1000, performance goal 0,00001. This structure can produce the smallest MSE value that is 0,000009957. So, this structure can recognize new words with 100% accuracy for trained data, 80% for the same respondents with trained data and reach 67.5% for new respondents.

Effects of stop voicing on long-distance acoustics

In Coleman (2003), the phonological feature [voice] was found to have a long-distance coarticulatory effect. However, voicing is not always realized as categorically present or absent. Davidson (2016) classifies partially voiced stops as having different patterns of voicing: bleed, trough, negative VOT, and hump. The present study answers the question: How do voicing patterns in stops affect the long-distance coarticulatory properties of the stop? Following the methodology of Coleman (2003), LPC coefficients, voicing, F0, harmonics-to-noise ratio, and z-score normalized RMS amplitude were measured in the read data from the Rainbow Passage in the Nationwide Speech Project (Clopper and Pisoni, 2006) in order to examine the stops and their coarticulatory effects. In the second sentence of the Rainbow Passage, the phones [b], [d], and [k] and their surrounding acoustics were measured in 10 millisecond frames that were aligned in time using dynamic time warping. A preliminary analysis using data visualization with smoothed conditional means reveals differences in the acoustics. There were several differences in both anticipatory and carryover coarticulation in the first through sixth and ninth through eleventh LPC coefficients. Subtle differences in acoustics as a result of long-distance coarticulation have implications for what makes both human and synthetic speech sound natural.

Automatic Identification of Cough Events from Acoustic Signals.

Cough is a common symptom of numerous respiratory diseases. In certain cases, such as asthma and COPD, early identification of coughs is useful for the management of these diseases. This paper presents an algorithm for automatic identification of cough events from acoustic signals. The algorithm is based on only four features of the acoustic signals including LPC coefficient, tonality index, spectral flatness and spectral centroid with a logistic regression model to label sound segments into cough and non-cough events. The algorithm achieves sensitivity of of 86.78%, specificity of 99.42%, and F1-score of 88.74%. Its high performance despite its small size of feature-space demonstrate its potential for use in remote patient monitoring systems for automatic cough detection using acoustic signals.

EOG angle recognition based on mixed LPC and DTW

In this thesis, EOG model is divided into four groups according to different scanning Angle, which is L1, L2, R1 and R2. Feature extraction adopts mixed LPC coefficient with the EOG waveform feature. This thesis adopts DTW, SVM and BP neural network on the EOG Angle recognition, each algorithm can achieve good recognition rate, but DTW is better. This thesis also designs an application that controls cartoon character movement through eye glancing to different Angles in different directions. That app can identify the eye movement type of the current data based on the template already stored and DTW algorithm.

Efficiency of chosen speech descriptors in relation to emotion recognition

This research paper presents parametrization of emotional speech using a pool of common features utilized in emotion recognition such as fundamental frequency, formants, energy, MFCC, PLP, and LPC coefficients. The pool is additionally expanded by perceptual coefficients such as BFCC, HFCC, RPLP, and RASTA PLP, which are used in speech recognition, but not applied in emotion detection. The main contribution of this work is the comparison of the accuracy performance of emotion detection for each feature type based on the results provided by both k-NN and SVM algorithms with 10-fold cross-validation. Analysis was performed on two different Polish emotional speech databases: voice performances by professional actors in comparison with the author’s spontaneous speech.

Effect of Glottal Excitation Interchange in Hindi and Dogri Languages

Voice acoustics is an active area of research which studies speaking voice and has gain popularity due to rapid advancements in digital signal processing. The shape of glottal excitation and the vocal tract may be speaker and language dependent. The objective of this paper is to study the effect of glottal excitation interchange on the quality and intelligibility in Hindi and Dogri languages. For this, recordings of six speakers (3 males and 3 females) were carried out in Dogri and Hindi languages. Cardinal vowels (/a/, /i/, /u/) were extracted from recordings of each speaker. Investigations were carried out by interchanging the glottal excitations corresponding to the vowels in the two languages for each speaker. The analysis of the results showed that interchange of excitation does not provide satisfactory quality of the synthesized speech in terms of identity and clarity of speech. Further, the synthesized speech is perceived as it was spoken in the original language. It was also observed that if any two of the parameters (excitation, gain, vocal tract LPC coefficients) are interchanged, the accent of the original language also changes. It means that minimum two of the three parameters are necessary to interchange for modifying the accent of the language under consideration.

TEXT-INDEPENDENT SPEAKER RECOGNITION USING COMBINED LPC AND MFC COEFFICIENTS

There are so many techniques for textindependent speaker recognition. However this text-independent speaker recognition is very difficult because the recognition is performed irrespective of what one he is saying. This paper presents a very simple approach to text independent recognition where the recognition is performed by using both LPC and MFC coefficients in parallel and the results of both methods are combined for best matching of the speaker. Here the ANN is used for classification.

Perceptual Speech Hashing Authentication Algorithm Based on Linear Prediction Analysis

According to the situation that traditional speech authentication algorithms are not appropriated for the present speech communication, we proposed a speech authentication algorithm of perceptual hashing which is based on linear forecast analysis (LPC), it can satisfy the requirement of the efficiency and the robustness for the speech authentication. Firstly, the LPC coefficients are optimized, based on the principle of LPC for speech signal, a new prediction coefficients is constituted through combing the energy and the LPC coefficients of a frame; then dividing the prediction coefficient matrix into smaller blocks, and do singular value decomposition (SVD) to the coefficients of each block. Finally, quantifying the formed singular value and getting perceptual hashing sequences. Experiments show that the proposed algorithm has good robustness for content preserving operations, and it doesn’t reduce the efficiency while meeting robustness, it can satisfy the real-time requirement of speech communication. DOI : http://dx.doi.org/10.11591/telkomnika.v12i4.4934

Estimation and Statistical Analysis of Human Voice Parameters to Investigate the Influence of Psychological Stress and to Determine the Vocal Tract Transfer Function of an Individual

In this study the principal focus is to examine the influence of psychological stress (both positive and negative stress) on the human articulation and to determine the vocal tract transfer function of an individual using inverse filtering technique. Both of these analyses are carried out by estimating various voice parameters. The outcomes of the analysis of psychological stress indicate that all the voice parameters are affected due to the influence of stress on humans. About 35 out of 51 parameters follow a unique course of variation from normal to positive and negative stress in 32% of the total analyzed signals. The upshot of the analysis is to determine the vocal tract transfer function for each vowel for an individual. The analysis indicates that it can be computed by estimating the mean of the pole zero plots of that individual’s vocal tract estimated for the whole day. Besides this, an analysis is presented to find the relationship between the LPC coefficients of the vocal tract and the vocal tract cavities. The results of the analysis indicate that all the LPC coefficients of the vocal tract are affected due to change in the position of any cavity.

Classificação de sinais de vozes saudáveis e patológicas por meio da combinação entre medidas da análise dinâmica não linear e codificação preditiva linear

Acoustic analysis has been suggested as a noninvasive aiding and low cost tool for laryngeal disease diagnosis. Several techniques are employed using either the linear model of speech production, or the nonlinear dynamic analysis of voice signals. The first method is based on source-filter theory, in which the source is the larynx and the filter is the vocal tract. In this model, the unvoiced sounds are modeled by a random noise source and the voiced ones by impulse train at the speaker fundamental frequency. In nonlinear approach, aspects of the human voice are considered, not explored in the linear model, such as temporal variation of the vocal tract shape, resonances associated with its physiology, losses due to friction in the vocal tract inner walls, sound radiation in the lips, nose coupling and dynamic behavior associated with vocal fold vibration. This work combines the two approaches and evaluates the performance in classifying the features individually, and from their combination. Eight measures are employed derived from the nonlinear dynamic analysis (correlation dimension, four entropy measures, Hurst exponent, the largest Lyapunov exponent and the first minimum of mutual information function), besides LPC coefficients obtained from linear predictive analysis. The results suggest the feasibility of the employed technique to discriminate between healthy and pathological voices in general, but also among specific laryngeal diseases as vocal fold edema, nodules and paralysis.

Noise Spectrum Estimation Using Line Spectral Frequencies for Robust Speech Recognition

This paper presents a novel method for estimating reliable noise spectral magnitude for acoustic background noise suppression where only a single microphone recording is available. The proposed method finds noise estimates from spectral magnitudes measured at line spectral frequencies (LSFs), under the observation that adjacent LSFs are near the peak frequencies and isolated LSFs are close to the relatively flattened valleys of LPC spectra. The parameters used in the proposed method are LPC coefficients, their corresponding LSFs, and the gain of LPC residual signals, so it suits well to LPC-based speech coders.

신경회로망 기반 고장 진단 시스템을 위한 고장 신호별 특징 벡터 결정 방법

As rotating machines play an important role in industrial applications such as aeronautical, naval and automotive industries, many researchers have developed various condition monitoring system and fault diagnosis system by applying various techniques such as signal processing and pattern recognition. Recently, fault diagnosis systems using artificial neural network have been proposed. For effective fault diagnosis, this paper used MLP(multi-layer perceptron) network which is widely used in pattern classification. Since using obtained signals without preprocessing as inputs of neural network can decrease performance of fault classification, it is very important to extract significant features of captured signals and to apply suitable features into diagnosis system according to the kinds of obtained signals. Therefore, this paper proposes the decision method of the proper feature vectors about each fault signal for neural-network-based fault diagnosis system. We applied LPC coefficients, maximum magnitudes of each spectral section in FFT and RMS(root mean square) and variance of wavelet coefficients as feature vectors and selected appropriate feature vectors as comparing error ratios of fault diagnosis for sound, vibration and current fault signals. From experiment results, LPC coefficients and maximum magnitudes of each spectral section showed 100 % diagnosis ratios for each fault and the method using wavelet coefficients had noise-robust characteristic.

Optimum HMM combined with vector Quantization for Hindi Speech word Recognition

This paper proposes an optimum speaker-independent, isolated word Hidden Markov Model (HMM) recognizer for the Hindi language. The recognition system is based on the combination of the vector quantization (VQ) technique at the acoustical level and the Markovian modeling at the recognition level. The recognizer consists of three modules – feature extraction, vector quantizer and HMM training and testing modules. The scheme proposed here firstly computes the acoustic features in terms of the Linear Predictive Cepstral LPC coefficients, Mel-Frequency Cepstral coefficients and delta MFCC along with noise and silence detection. Then, codebooks are created using VQ, and finally in the recognition phase, an optimum set of parameters are derived from different phases for getting the highest recognition score. The training and testing database consists of a set of 35 utterances of nine Indian cities/states and 35 utterances of nine digits spoken in Hindi by male and female speakers. The recognition rate was observed to be 98.61%.

LPC modelling and cepstral analysis applied to vocal fold pathology detection

Laryngeal pathologies are generally diagnosed using laryngoscopical exams, which are considered invasive to patients. Digital signal processing techniques are noninvasive and can be applied to perform an acoustic analysis for vocal quality assessment providing an objective diagnosis of pathological voices. This paper aims at specifying and evaluating the acoustic features for vocal fold edema through a parametric modelling based on the resonant structure of the human speech production mechanism by LPC and LPC-based cepstral coefficients and a nonparametric approach related to human auditory perception system by mel-frequency cepstral coefficients. A vector-quantising-trained distance classifier is used in the discrimination process.

Cepstral domain interpretations of line spectral frequencies

In this paper, we derive a relationship between a linear prediction (LP) polynomial and its corresponding line spectral frequencies (LSFs) in cepstral domain. We first obtain the cepstral representations of the symmetric and the antisymmetric polynomials constructed from the LP polynomial. The sum of these cepstra corresponds to the pseudo-cepstrum [H.K. Kim, S.H. Choi, H.S. Lee, On approximating line spectral frequencies to LPC cepstral coefficients, IEEE Trans. Acoust. Speech Audio Process. 8(2) (2000) 195–199]. Finally, LSFs can be computed by matching these two cepstra. Additionally, we derive the recursive relations between LSFs and the LPC coefficients, as well as between LSFs and LPC cepstral coefficients.

A Speech Packet Loss Concealment Method Using Linear Prediction

We proposed and evaluated a speech packet loss concealment method which predicts lost segments from speech included in packets either before, or both before and after the lost packet. The lost segments are predicted recursively by using linear prediction both in the forward direction from the packet preceding the loss, and in the backward direction from the packet succeeding the lost segment. Predicted samples in each direction are smoothed by averaging using linear weights to obtain the final interpolated signal. The adjacent segments are also smoothed extensively to significantly reduce the speech quality discontinuity between the interpolated signal and the received speech signal. Subjective quality comparisons between the proposed method and the the packet loss concealment algorithm described in the ITU standard G.711 Appendix I showed similar scores up to about 10% packet loss. However, the proposed method showed higher scores above this loss rate, with Mean Opinion Score rating exceeding 2.4, even at an extremely high packet loss rate of 30%. Packet loss concealment of speech degraded with G.729 coding, and babble noise mixed speech showed similar trends, with the proposed method showing higher qualities at high loss rates. We plan to further improve the performance by using adaptive LPC prediction order depending on the estimated pitch, and adaptive LPC bandwidth expansion depending on the consecutive number of repetitive prediction, among many other improvements. We also plan to investigate complexity reduction using gradient LPC coefficient updates, and processing delay reduction using adaptive forward/bidirectional prediction modes depending on the measured packet loss ratio.

Digital signals analysis with the LPC method

This paper concerns the issue of signal analysis by the linear predictive method. It is not only about frequency analysis, that is obtaining the LPC spectrum and comparing it with the Fourier one, but also about analysing prediction coefficients themselves. We will also discuss the program made by the authors of this paper, which beside foregoing functionalities, enables vocal tract visualization, modelled on the basis of LPC coefficients.

Phase quantization method and apparatus

A phase quantization method and apparatus in which the phase information of the input signal such as at the time of the sinusoidal synthesis encoding can be quantized efficiently. The phase of the input signal derived from speech signals from an input terminal 11 is found by a phase detection unit 12 and scalar-quantized by a scalar quantizer 13 . The spectral amplitude weighting k of each harmonics is calculated by a weighting calculation unit 18 based on the LPC coefficients from a terminal 17 . Using the weighting k, a bit allocation calculation unit 19 calculates an optimum number of quantization bits of respective harmonics to send the calculated optimum number to the scalar quantizer 13.