Academic literature on the topic 'Mel Frequency Cepstral Coefficients (MFCC)'

Sistema de reconhecimento de fala tem amplo emprego no universo industrial, no aperfeiçoamento de operações e procedimentos humanos e no setor do entretenimento e recreação. O objetivo específico do trabalho foi conceber e desenvolver um sistema de reconhecimento de voz, capaz de identificar comandos de voz, independentemente do locutor. A finalidade precípua do sistema é controlar movimentos de robôs, com aplicações na indústria e no auxílio de deficientes físicos. Utilizou-se a abordagem da tomada de decisão por meio de uma rede neural treinada com as características distintivas do sinal de fala de 16 locutores. As amostras dos comandos foram coletadas segundo o critério de conveniência (em idade e sexo), a fim de garantir uma maior discriminação entre as características de voz, e assim alcançar a generalização da rede neural utilizada. O préprocessamento consistiu na determinação dos pontos extremos da locução do comando e na filtragem adaptativa de Wiener. Cada comando de fala foi segmentado em 200 janelas, com superposição de 25% . As features utilizadas foram a taxa de cruzamento de zeros, a energia de curto prazo e os coeficientes ceptrais na escala de frequência mel. Os dois primeiros coeficientes da codificação linear preditiva e o seu erro também foram testados. A rede neural empregada como classificador foi um perceptron multicamadas, treinado pelo algoritmo backpropagation. Várias experimentações foram realizadas para a escolha de limiares, valores práticos, features e configurações da rede neural. Os resultados foram considerados muito bons, alcançando uma taxa de acertos de 89,16%, sob as condições de pior caso da amostragem dos comandos.<br>Systems for speech recognition have widespread use in the industrial universe, in the improvement of human operations and procedures and in the area of entertainment and recreation. The specific objective of this study was to design and develop a voice recognition system, capable of identifying voice commands, regardless of the speaker. The main purpose of the system is to control movement of robots, with applications in industry and in aid of disabled people. We used the approach of decision making, by means of a neural network trained with the distinctive features of the speech of 16 speakers. The samples of the voice commands were collected under the criterion of convenience (age and sex), to ensure a greater discrimination between the voice characteristics and to reach the generalization of the neural network. Preprocessing consisted in the determination of the endpoints of each command signal and in the adaptive Wiener filtering. Each speech command was segmented into 200 windows with overlapping of 25%. The features used were the zero crossing rate, the short-term energy and the mel-frequency ceptral coefficients. The first two coefficients of the linear predictive coding and its error were also tested. The neural network classifier was a multilayer perceptron, trained by the backpropagation algorithm. Several experiments were performed for the choice of thresholds, practical values, features and neural network configurations. Results were considered very good, reaching an acceptance rate of 89,16%, under the `worst case conditions for the sampling of the commands.

Speech as a medium for communication has always been important in its ability to convey our ideas, personality and emotions. It is therefore not strange that Quality of Experience (QoE) becomes central to any business relying on voice communication. Using Unified Communication (UC) systems, users can communicate with each other in several ways using many different devices, making QoE an important aspect for such systems. For this thesis, automatic methods for assessing speech quality of the voice calls in Briteback’s UC application is studied, including a comparison of the researched methods. Three methods all using a Gaussian Mixture Model (GMM) as a regressor, paired with extraction of Human Factor Cepstral Coefficients (HFCC), Gammatone Frequency Cepstral Coefficients (GFCC) and Modified Mel Frequency Cepstrum Coefficients (MMFCC) features respectively is studied. The method based on HFCC feature extraction shows better performance in general compared to the two other methods, but all methods show comparatively low performance compared to literature. This most likely stems from implementation errors, showing the difference between theory and practice in the literature, together with the lack of reference implementations. Further work with practical aspects in mind, such as reference implementations or verification tools can make the field more popular and increase its use in the real world.

In this paper a novel speaker recognition system is introduced. Automated speaker recognition has become increasingly popular to aid in crime investigations and authorization processes with the advances in computer science. Here, a recurrent neural network approach is used to learn to identify ten speakers within a set of 21 audio books. Audio signals are processed via spectral analysis into Mel Frequency Cepstral Coefficients that serve as speaker specific features, which are input to the neural network. The Long Short-Term Memory algorithm is examined for the first time within this area, with interesting results. Experiments are made as to find the optimum network model for the problem. These show that the network learns to identify the speakers well, text-independently, when the recording situation is the same. However the system has problems to recognize speakers from different recordings, which is probably due to noise sensitivity of the speech processing algorithm in use.

Cette thèse présente une investigation acoustico-phonétique des détails phonétiques des fricatives russes.L'objectif principal était de détecter des corrélats acoustiques porteurs d'infor- mations linguistiques et idiosyncrasiques. Les questions abordées étaient de savoir si le lieu d'articulation, le sexe du locuteur ou son identité peuvent être prédits par des indices acoustiques et quelles mesures acoustiques représentent les indicateurs les plus fiables. En outre, la distribution des caractéristiques spécifiques au locuteur et à la variation inter et intra locuteur à travers les indices acoustiques a été étudiée plus en détail. Le projet a commencé par la création d'une grande base de données audio des fricatives russes. Des enregistrements acoustiques ont été obtenus auprès de 59 locuteurs russes natifs. Le jeu de données résultant est composé de 22 561 occurrences comprenant les fricatives [f], [s], [ʃ], [x], [v], [z], [ʒ], [sj], [ɕ], [vʲ], [zʲ]. Deux analyses ont été menées à partir de cette base de données. Dans la première étude, un échantillon de données de 6320 occurrences (40 locuteurs) a été utilisé. Trois techniques d'extraction acoustisque (à partir du son complet, de la durée du bruit et des fenêtres centrales de 30 ms) ont été sollicitées pour extraire des mesures temporelles et spectrales. En outre, 13 coefficients cepstraux (Mel-Frequency Cepstral Coefficients, MFCC) ont été calculés à partir de la fenêtre centrale de 30 ms. Des classificateurs fondés sur des arbres de décision simples, des forêts aléatoires, des machines à vecteurs de support (Support-vector machine, SVM) et des réseaux neuronaux ont été entraînés et testés pour distinguer trois fricatives non palatalisées [f], [s] et<br>This thesis represents an acoustic-phonetic investigation of phonetic details in Russian fricatives. The main aim was to detect acoustic correlates that carry linguistic and idiosyncratic information. The questions addressed were whether the place of articulation, speakers' gender and ID can be predicted by a set of acoustic cues and which acoustic measures represent the most reliable indicators. Furthermore, the distribution of speaker-specific characteristics and inter- and intra-speaker variation across acoustic cues were studied in more detail.The project started with the generation of a large audio database of Russian fricatives. Then, two follow-up analyses were conducted. Acoustic recordings were collected from 59 native Russian speakers. The resulting dataset consists of 22,561 tokens including the fricatives [f], [s], [ʃ], [x], [v], [z], [ʒ], [sj], [ɕ], [vʲ], [zʲ].The first study employed a data sample of 6320 tokens (from 40 speakers). Temporal and spectral measurements were extracted using three acoustic cue extraction techniques (full sound, the noise part, and the middle 30ms windows). Furthermore, 13 Mel Frequency Cepstral Coefficients were computed from the middle 30ms window.Classifiers based on single decision trees, random forests, support vector machines, and neural networks were trained and tested to distinguish between the three non-palatalized fricatives [f], [s] and [ʃ].The results demonstrate that machine learning techniques are very successful at classifying the Russian voiceless non-palatalized fricatives [f], [s] and [ʃ] by using the centre of gravity and the spectral spread irrespective of contextual and speaker variation. The three acoustic cue extraction techniques performed similarly in terms of classification accuracy (93% and 99%), but the spectral measurements extracted from the noise parts resulted in slightly better accuracy. Furthermore, Mel Frequency Cepstral Coefficients show marginally higher predictive power over spectral cues (&lt; 2%).This suggests that both spectral measures and Mel Frequency Cepstral provide sufficient information for the classification of these fricatives and their choice depends on the particular research question or application. The second study's dataset consists of 15812 tokens (59 speakers) that contain [f], [s], [ʃ], [x], [v], [z], [ʒ], [sj], [ɕ]. As in the first study, two types of acoustic cues were extracted including 11 acoustic speech features (spectral cues, duration and HNR measures) and 13 Mel Frequency Cepstral Coefficients. Classifiers based on single decision trees and random forests were trained and tested to predict speakers' gender and ID

The growth of multimedia content in recent years motivated the research on audio classification and content retrieval area. In this thesis, a general environmental audio classification and retrieval approach is proposed in which higher level semantic classes (outdoor, nature, meeting and violence) are obtained from lower level acoustic classes (emergency alarm, car horn, gun-shot, explosion, automobile, motorcycle, helicopter, wind, water, rain, applause, crowd and laughter). In order to classify an audio sample into acoustic classes, MPEG-7 audio features, Mel Frequency Cepstral Coefficients (MFCC) feature and Zero Crossing Rate (ZCR) feature are used with Hidden Markov Model (HMM) and Support Vector Machine (SVM) classifiers. Additionally, a new classification method is proposed using Genetic Algorithm (GA) for classification of semantic classes. Query by Example (QBE) and keyword-based query capabilities are implemented for content retrieval.

In radar applications, there are often times when one does not only want to know that there is a target that reflecting the out sent signals but also what kind of target that reflecting these signals. This project investigates the possibilities to from raw radar data transform reflected signals and take use of human perception, in particular our hearing, and by a machine learning approach where patterns and characteristics in data are used to answer the earlier mentioned question. More specific the investigation treats two kinds of targets that are fairly comparable namely smaller Unmanned Aerial Vehicles (UAV) and Birds. By extracting complex valued radar video so called I/Q data generated by these targets using signal processing techniques and transform this data to a real signals and after this transform the signals to audible signals. A feature set commonly used in speech recognition namely Mel Frequency Cepstral Coefficients are used two describe these signals together with two Support Vector Machine classification models. The two models where tested with an independent test set and the linear model achieved a overall prediction accuracy 93.33 %. Individually the prediction resulted in 93.33 % correct classification on the UAV and 93.33 % on the birds. Secondly a radial basis model with a overall prediction accuracy of 98.33 % where achieved. Individually the prediction resulted in 100% correct classification on the UAV and 96.76 % on the birds. The project is partly done in collaboration with J. Clemedson [2] where the focus is, as mentioned earlier, to transform the signals to audible signals.<br>I radar applikationer räcker det ibland inte med att veta att systemet observerat ett mål när en reflekted signal dekekteras, det är ofta också utav stort intresse att veta vilket typ av föremål som signalen reflekterades mot. Detta projekt undersöker möjligheterna att utifrån rå radardata transformera de reflekterade signalerna och använda sina mänskliga sinnen, mer specifikt våran hörsel, för att skilja på olika mål och också genom en maskininlärnings approach där med hjälp av mönster och karaktärsdrag för dessa signaler används för att besvara frågeställningen. Mer ingående avgränsas denna undersökning till två typer av mål, mindre obemannade flygande farkoster (UAV) och fåglar. Genom att extrahera komplexvärd radar video även känt som I/Q data från tidigare nämnda typer av mål via signalbehandlingsmetoder transformera denna data till reella signaler, därefter transformeras dessa signaler till hörbara signaler. För att klassificera dessa typer av signaler används typiska särdrag som också används inom taligenkänning, nämligen, Mel Frequency Cepstral Coefficients tillsammans med två modeller av en Support Vector Machine klassificerings metod. Med den linjära modellen uppnåddes en prediktions noggrannhet på 93.33%. Individuellt var noggrannheten 93.33 % korrekt klassificering utav UAV:n och 93.33 % på fåglar. Med radial bas modellen uppnåddes en prediktions noggrannhet på 98.33%. Individuellt var noggrannheten 100 % korrekt klassificering utav UAV:n och 96.76% på fåglar. Projektet är delvis utfört med J. Clemedson [2] vars fokus är att, som tidigare nämnt, transformera dessa signaler till hörbara signaler.

We evaluate the security of human voice password databases from an information theoretical point of view. More specifically, we provide a theoretical estimation on the amount of entropy in human voice when processed using the conventional GMM-UBM technologies and the MFCCs as the acoustic features. The theoretical estimation gives rise to a methodology for analyzing the security level in a corpus of human voice. That is, given a database containing speech signals, we provide a method for estimating the relative entropy (Kullback-Leibler divergence) of the database thereby establishing the security level of the speaker verification system. To demonstrate this, we analyze the YOHO database, a corpus of voice samples collected from 138 speakers and show that the amount of entropy extracted is less than 14-bits. We also present a practical attack that succeeds in impersonating the voice of any speaker within the corpus with a 98% success probability with as little as 9 trials. The attack will still succeed with a rate of 62.50% if 4 attempts are permitted. Further, based on the same attack rationale, we mount an attack on the ALIZE speaker verification system. We show through experimentation that the attacker can impersonate any user in the database of 69 people with about 25% success rate with only 5 trials. The success rate can achieve more than 50% by increasing the allowed authentication attempts to 20. Finally, when the practical attack is cast in terms of an entropy metric, we find that the theoretical entropy estimate almost perfectly predicts the success rate of the practical attack, giving further credence to the theoretical model and the associated entropy estimation technique.

The thesis deals with a design and an implementation of software for a detection of logopaedia defects of speech. Due to the need of early logopaedia defects detecting, this software is aimed at a child’s age speaker. The introductory part describes the theory of speech realization, simulation of speech realization for numerical processing, phonetics, logopaedia and basic logopaedia defects of speech. There are also described used methods for feature extraction, for segmentation of words to speech sounds and for features classification into either correct or incorrect pronunciation class. In the next part of the thesis there are results of testing of selected methods presented. For logopaedia speech defects recognition algorithms are used in order to extract the features MFCC and PLP. The segmentation of words to speech sounds is performed on the base of Differential Function method. The extracted features of a sound are classified into either a correct or an incorrect pronunciation class with one of tested methods of pattern recognition. To classify the features, the k-NN, SVN, ANN, and GMM methods are tested.