Linear Predictive Coding Parameters Research Articles

Neural Nets Distributed On Microcontrollers Using Metaheuristic Parallel Optimization Algorithm

Metaheuristic algorithms are powerful methods for solving compute intensive problems. neural Networks, when trained well, are great at prediction and classification type of problems. Backpropagation is the most popular method utilized to obtain the weights of Neural Nets though it has some limitations of slow convergence and getting stuck in a local minimum. In order to overcome these limitations, in this paper, a hybrid method combining the parallel distributed bat algorithm with backpropagation is proposed to compute the weights of the Neural Nets. The aim is to use the hybrid method in applications of a distributed nature. Our study uses the Matlab® software and Arduino® microcontrollers as a testbed. To test the performance of the testbed, an application in the area of speech recognition is carried out. Due to the resource limitations of Arduino microcontrollers, the core speech pre-processing of LPC (linear predictive coding) feature extractions are done in Matlab® and only the LPC parameters are passed to the Neural Nets, which are implemented on Arduino microcontrollers. The experimental results show that the proposed scheme does produce promising results.

Neural Nets Distributed on Microcontrollers using Metaheuristic Parallel Optimization Algorithm

Metaheuristic algorithms are powerful methods for solving compute intensive problems. neural Networks, when trained well, are great at prediction and classification type of problems. Backpropagation is the most popular method utilized to obtain the weights of Neural Nets though it has some limitations of slow convergence and getting stuck in a local minimum. In order to overcome these limitations, in this paper, a hybrid method combining the parallel distributed bat algorithm with backpropagation is proposed to compute the weights of the Neural Nets. The aim is to use the hybrid method in applications of a distributed nature. Our study uses the Matlab® software and Arduino® microcontrollers as a testbed. To test the performance of the testbed, an application in the area of speech recognition is carried out. Due to the resource limitations of Arduino microcontrollers, the core speech pre-processing of LPC (linear predictive coding) feature extractions are done in Matlab® and only the LPC parameters are passed to the Neural Nets, which are implemented on Arduino microcontrollers. The experimental results show that the proposed scheme does produce promising results.

Linear Predictive Approaches Separate Field Potentials in Animal Model of Parkinson's Disease.

Parkinson's disease (PD) causes impaired movement and cognition. PD can involve profound changes in cortical and subcortical brain activity as measured by electroencephalography or intracranial recordings of local field potentials (LFP). Such signals can adaptively guide deep-brain stimulation (DBS) as part of PD therapy. However, adaptive DBS requires the identification of triggers of neuronal activity dependent on real time monitoring and analysis. Current methods do not always identify PD-related signals and can entail delays. We test an alternative approach based on linear predictive coding (LPC), which fits autoregressive (AR) models to time-series data. Parameters of these AR models can be calculated by fast algorithms in real time. We compare LFPs from the striatum in an animal model of PD with dopamine depletion in the absence and presence of the dopamine precursor levodopa, which is used to treat motor symptoms of PD. We show that in dopamine-depleted mice a first order AR model characterized by a single LPC parameter obtained by LFP sampling at 1 kHz for just 1 min can distinguish between levodopa-treated and saline-treated mice and outperform current methods. This suggests that LPC may be useful in online analysis of neuronal signals to guide DBS in real time and could contribute to DBS-based treatment of PD.

Line spectral frequencies modeling by a mixture of von Mises–Fisher distributions

Efficient quantization of the linear predictive coding (LPC) parameters plays a key role in parametric speech coding. The line spectral frequency (LSF) representation of the LPC parameters has found its applications in speech model quantization. In practical implementations of vector quantization (VQ), probability density function optimized VQ has been shown to be more efficient than the VQ based on training data. In this paper, we present the LSF parameters by a unit vector form, which has directional characteristics. The underlying distribution of this unit vector variable is modeled by a von Mises–Fisher mixture model (VMM). An optimal inter-component bit allocation strategy is proposed based on high rate theory and a distortion-rate (D-R) relation is derived for the VMM based-VQ (VVQ). Experimental results show that the VVQ outperforms the recently introduced Dirichlet mixture model-based VQ and the conventional Gaussian mixture model-based VQ, in terms of modeling performance and D-R relation.

A novel hierarchical decomposition vector quantization method for high-order LPC parameters

The paper investigates vector quantization coding of high-order (e.g., 20th-50th order) linear prediction coding (LPC) parameters, and proposes a novel hierarchical decomposition vector quantization method for a scalable speech coding framework with variable orders of LPC analysis. Instead of vector quantizing the whole group of LPC parameters in the linear spectral frequency (LSF) domain directly, the proposed method decomposes the high-order LPC model into several low-order (e.g., 10th-order) LPC models, and vector quantizes them in the LSF domain separately. For the decomposition, the high-order LPC model is converted into a group of reflection coefficients at first, and then the group is split into several subgroups and converted into multiple low-order LPC models. It is shown that the proposed method is naturally suitable for a scalable coding framework where the information of the decomposed low-order LPC models can be encoded into a multi-layered bitstream and can be combined in a progressive way to recover the high-order LPC information. Experiments in a scalable coding framework with variable LPC analysis orders (10-50) reveal that, compared to a direct vector quantization scheme, the proposed method can reduce the size of the codebook and the number of coding bits significantly, and can also efficiently reduce the computation cost.

High-Rate Vector Quantization for Noisy Channels With Applications to Wideband Speech Spectrum Compression

This paper considers the high-rate performance of source coding for noisy discrete symmetric channels with random index assignment (IA). Accurate analytical models are developed to characterize the expected distortion performance of vector quantization (VQ) for a large class of distortion measures. It is shown that when the point density is continuous, the distortion can be approximated as the sum of the source quantization distortion and the channel-error induced distortion. Expressions are also derived for the continuous point density that minimizes the expected distortion. Next, for the case of mean squared error distortion, a more accurate analytical model for the distortion is derived by allowing the point density to have a singular component. The extent of the singularity is also characterized. These results provide analytical models for the expected distortion performance of both conventional VQ as well as for channel-optimized VQ. As a practical example, compression of the linear predictive coding parameters in the wideband speech spectrum is considered, with the log spectral distortion as performance metric. The theory is able to correctly predict the channel error rate that is permissible for operation at a particular level of distortion.

Vector quantization of LSF parameters with a mixture of dirichlet distributions

Quantization of the linear predictive coding parameters is an important part in speech coding. Probability density function (PDF)-optimized vector quantization (VQ) has been previously shown to be more efficient than VQ based only on training data. For data with bounded support, some well-defined bounded-support distributions (e.g., the Dirichlet distribution) have been proven to outperform the conventional Gaussian mixture model (GMM), with the same number of free parameters required to describe the model. When exploiting both the boundary and the order properties of the line spectral frequency (LSF) parameters, the distribution of LSF differences LSF can be modelled with a Dirichlet mixture model (DMM). We propose a corresponding DMM based VQ. The elements in a Dirichlet vector variable are highly mutually correlated. Motivated by the Dirichlet vector variable's neutrality property, a practical non-linear transformation scheme for the Dirichlet vector variable can be obtained. Similar to the Karhunen-Loeve transform for Gaussian variables, this non-linear transformation decomposes the Dirichlet vector variable into a set of independent beta-distributed variables. Using high rate quantization theory and by the entropy constraint, the optimal inter- and intra-component bit allocation strategies are proposed. In the implementation of scalar quantizers, we use the constrained-resolution coding to approximate the derived constrained-entropy coding. A practical coding scheme for DVQ is designed for the purpose of reducing the quantization error accumulation. The theoretical and practical quantization performance of DVQ is evaluated. Compared to the state-of-the-art GMM-based VQ and recently proposed beta mixture model (BMM) based VQ, DVQ performs better, with even fewer free parameters and lower computational cost

An algorithm for token-by-token linear predictive coding analysis of oral monophthongs.

Linear Predictive Coding (LPC) is the most reliable vowel formant estimation method. LPC trajectories closely approximate formants and are preferable as a means of avoiding human bias. Nevertheless, formant analysis by LPC is not error free, as it can be affected by fundamental frequency, noise and vocal tract morphology. Error minimization is usually achieved through manual checking of LPC results with spectral contour. An algorithm to find optimal LPC parameters (Burg method) on a token by token basis was developed and implemented as a Praat script. The algorithm minimizes inter- and intra-vowel formant variability without knowledge of vowel identity. Algorithm results were compared with results of manual adjustment of LPC parameters on a trial and error basis on a corpus of 180 stressed and unstressed oral monophthongs of Brazilian Portuguese from a male speaker. First and second formants linear correlations are r2 = 0.98 and 0.99 respectively. The areas of /i, a, u/ vocalic triangles (mean values as vertices) of manual and automatic measurements showed only 4.5% difference. These results suggest the algorithm can be generalized to any speaker and speed up the measurement of vowel formants in large corpora. [Work supported by CNPq and Fapesp.]

Emotion Recognition in Speech Using Neural Network

Emotion recognition in speech studies has grown after the growing of speech recognition technologies. The aim of such studies is to make language interfaces in human-computer interaction applications more wide in use and to make it efficient. And it may help the studiers of the human sound areas. This study deals with four spoken sentences of sixteen short utterance expressing five emotions: a happiness, anger, sadness, fear, and normal (unemotional) state. The feature extraction techniques are used to capture the most important information of the signal, this process is applied in the time domain. The extracted feature vector contain 12 LPC (linear predictive coding) parameters, the pitch, the power, and the three first formant frequencies. While in the second part of the system (the emotion recognizer) the artificial neural network technology is used, the designing network is with feedforward backpropagation. This design used ten separated nets. each one make a decision whether the spoken sentence is belong to one of only tow emotions. So the computational time is increased with this system, while such a system has better performance with respect to other systems. The designed system could recognize all emotions with ratio 75% except the fear one which recognized in ratio of 50%.

Empirical Lower Bound on the Bitrate for the Transparent Memoryless Coding of Wideband LPC Parameters

In this letter, we determine empirical lower bounds on the bitrate required to transparently code linear predictive coding (LPC) parameters derived from wideband speech. This is achieved via extrapolation of the operating distortion-rate curve of an unconstrained vector quantizer that is trained using artificial vectors generated by a Gaussian mixture model. Memoryless coding is considered and two competing LPC parameter representations are investigated. Our results show a lower bound of 31 bits/frame when assuming high-rate linearity in the operating distortion-rate curve and 35 bits/frame for an exponential curve. We also evaluate a recent quantization scheme and compare its performance against this lower bound

Determination of the frequency vibrato extent based on the linear predictive coding parameters

Many time-domain signals can be modeled as a sum of sinusoids whose main parameters are amplitude and frequency. The frequency can be obtained from the linear predictive coding (LPC) parameters by the pole location in the Z-domain. Although intuition dictates that the amplitude could also be obtained from the LPC parameters, there appears to be no information available that would allow its computation. The purpose of this research is to investigate the relationship between the LPC parameters and the amplitude of a wave in the time domain. In order to accomplish this, the power spectrum obtained using the LPC method was plotted against the FFT spectrum of vibrato samples. They were then compared, and several parameters of the LPC spectrum were changed until both spectra matched. From these matched spectra, the wave amplitude was obtained. This new method was then used in an application to determine the time-varying amplitude of the vibrato wave. This served two purposes: first, to verify the accuracy of the method, and second, to show the usefulness of the time-domain parameters. The results show that there is a direct relationship between the LPC parameters and the amplitude of a wave in the time domain.

Voicing-specific LPC quantization for variable-rate speech coding

Phonetic classification of speech frames allows distinctive quantization and bit allocation schemes suited to the particular class. Separate quantization of the linear predictive coding (LPC) parameters for voiced and unvoiced speech frames is shown to offer useful gains for representing the synthesis filter commonly used in code-excited linear prediction (CELP) and other coders. Subjective test results are reported that determine the required bit rate and accuracy in the two classes of voiced and unvoiced LPC spectra for CELP coding with phonetic classification. It was found, in this context, that unvoiced spectra need 9 b/frame or more whereas voiced spectra need 25 b/frame or more with the quantization schemes used. New spectral distortion criteria needed to assure transparent LPC spectral quantization for each voicing class in CELP coders are presented. Similar subjective test results for speech synthesized from the true residual signal are also presented, leading to some interesting observations on the role of the analysis-by-synthesis structure of CELP. Objective performance assessments based on the spectral distortion measure are also presented. The theoretical distortion-rate function for the spectral distortion measure is estimated for voiced and unvoiced LPC parameters and compared with experimental results obtained with unstructured vector quantization (VQ). These results show a saving of at least 2 b/frame for unvoiced spectra compared to voiced spectra to achieve the same spectral distortion performance.

Vehicle classification by acoustic signature

The aim of this research is to investigate the feasibility of developing a traffic monitoring detector for the purpose of reliable on-line vehicle classification to aid traffic management systems. The detector used was a directional microphone connected to a DAT (Digital Audio Tape) recorder. The digital signal was preprocessed by LPC (Linear Predictive Coding) parameter conversion based on autocorrelation analysis. A Time Delay Neural Network (TDNN) was chosen to classify individual travelling vehicles based on their speed-independent acoustic signature. The paper provides a description of the TDNN architecture and training algorithm, and an overview of the LPC preprocessing and feature extraction technique as applied to audio monitoring of road traffic. The performance of TDNN vehicle classification, convergence, and accuracy for the training patterns are fully illustrated. To establish the viability of this classification approach, initially, recordings were carried out on a strip of airfield for four types of vehicles under controlled conditions. A TDNN network was successfully trained with 100% accuracy in classification for the training patterns, as well as the test patterns. The net was also robust to changes in the starting position of the acoustic waveforms with 86% accuracy for the same test data set. In the second phase of the experiment, roadside recordings were made at a two-way urban road site in the city of Leeds with no control over the environmental parameters such as background noise, interference from other travelling vehicles, or the speed of the recorded vehicle. A second TDNN network was also successfully trained with 96% accuracy for the training patterns and 84% accuracy for the test patterns.

A field trial of acoustic signature analysis for vehicle classification

The aim of this research is to investigate the feasibility of developing a traffic monitoring detector for the purpose of reliable on-line vehicle classification to aid traffic management systems. The detector used was a directional microphone connected to a DAT (Digital Audio Tape) recorder. The digital signal was pre-processed by LPC (Linear Predictive Coding) parameter conversion based on autocorrelation analysis. A Time Delay Neural Network (TDNN) was chosen to classify individual travelling vehicles based on their speed-independent acoustic signature. Locations for data acquisition included roadside recordings at a number of two-way urban road sites in the city of Leeds with no control over the environmental parameters such as background noise, interference from other travelling vehicles or the speed of the recorded vehicles. The results and performance analysis of TDNN vehicle classification, the convergence for training patterns and accuracy of test patterns are fully illustrated. The paper also provides a description of the TDNN architecture and training algorithm, and an overview of the LPC pre-processing and feature extraction technique as applied to audio monitoring of road traffic. In the final phase of the experiment, the four broad categorisations of vehicles for training the network consisted of: buses or lorries; small or large saloons; various types of motorcycles; and light goods vehicles or vans. A TDNN network was successfully trained with 94% accuracy for the training patterns and 82.4% accuracy for the test patterns.

Linear predictive coding parameter manipulation/synthesis of incorrectly produced /r/.

This study describes the manipulation and synthesis of LPC parameters to edit incorrectly produced utterances. In particular, it shows that formant frequencies can be manipulated to produce a consistent and reliable change in perception. It also demonstrates that this method can be used to produce quality synthesis of high-pitched voices.

Digital underwater acoustic voice communications

This paper describes the design of an underwater acoustic diver communication system controlled by a digital signal processor. The speech signal transmission rate is compressed by using linear predictive coding (LPC) and the extracted parameters are transmitted through the water to a synchronized receiver by employing digital pulse position modulation (DPPM). The pulse position in each time frame is estimated by an energy detection and decision algorithm which enables the received LPC parameters to be recovered and used to synthesize the speech signal.

Theoretical analysis of the high-rate vector quantization of LPC parameters

The paper presents a theoretical analysis of high-rate vector quantization (VQ) systems that use suboptimal, mismatched distortion measures, and describes the application of the analysis to the problem of quantizing the linear predictive coding (LPC) parameters in speech coding systems. First, it is shown that in many high-rate VQ systems the quantization distortion approaches a simple quadratically weighted error measure, where the weighting matrix is a "sensitivity matrix" that is an extension of the concept of the scalar sensitivity. The approximate performance of VQ systems that train and quantize using mismatched distortion measures is derived, and is used to construct better distortion measures. Second, these results are used to determine the performance of LPC vector quantizers, as measured by the log spectral distortion (LSD) measure, which have been trained using other error measures, such as mean-squared (MSE) or weighted mean-squared error (WMSE) measures of LEPC parameters, reflection coefficients and transforms thereof, and line spectral pair (LSP) frequencies. Computationally efficient algorithms for computing the sensitivity matrices of these parameters are described. In particular, it is shown that the sensitivity matrix for the LSP frequencies is diagonal, implying that a WMSE measured LSP frequencies converges to the LSD measure in high-rate VQ systems. Experimental results to support the theoretical performance estimates are provided.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Predictive coding of speech using microphone/speaker adaptation and vector quantization

A problem with speech coders based on trained quantizers is their lack of robustness against variations in the microphone and input filter response. In this paper a simple backward adaptive prefiltering technique is proposed as a means of improving the robustness and quality of a speech coder at no cost in bit rate. The technique is particularly useful in conjunction with vector quantization (VQ) of the linear predictive coding (LPC) parameters. The performance of the prefilter, denoted a microphone and speaker adaptation (MSA) filter, has been evaluated in terms of prediction gain and spectral distortion, together with objective and subjective quality of a 7.5 kbit/s CELP speech coder. In this coder a 10-bit direct VQ of the LPC parameters using the residual energy distortion measure has been applied. This is consistent with the covariance method of LPC analysis. Simulation results illustrate that the MSA filter significantly improves the performance and robustness of the LPC VQ against changes in the input response. The 7.5 kbit/s CELP with a trained excitation codebook and MSA was found to be clearly better (subjectively and objectively) than the one without MSA. The coder with MSA also showed to be practically indistinguishable from the same CELP with unquantized LPC coefficients and a stochastic excitation codebook. >

Interframe differential coding of line spectrum frequencies

Line spectrum frequencies (LSF's) uniquely represent the linear predictive coding (LPC) filter of a speech frame. In many vocoders LSF's are used to encode the LPC parameters. In this paper, an interframe differential coding scheme is presented for the LSF's. The LSF's of the current speech frame are predicted by using both the LSF's of the previous frame and some of the LSF's of the current frame. Then, the difference resulting from prediction is quantized.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Artificial conversational speech signal generation method for measuring characteristics of devices operated by speech signals

A method for generating an artificial conversational speech signal as the input test signal in measuring the characteristics of devices operated by speech signals, such as speech detectors, voice switches, and echo controllers is proposed. A state transition model among talkspurt, pause, double‐talk, and mutual silence in conversational speech signals was introduced to simulate the statistical characteristics of these states in real conversational speech signals. This model assumed that the cumulative distribution of the duration time in each state is exponential. During talkspurt intervals, an artificial voice was generated using a speech synthesizer that controlled the spectrum and source characteristics independently. This artificial signal generation method can generate a signal with given characteristics by changing the values of system parameters, such as the average duration time of each state, transition probability from one state to another, and LPC (linear predictive coding) parameters that expr...