Sergio Suárez Guerra

Noise pattern recognition of airplanes taking off: task for a monitoring system

This paper presents an original work for aircraft noise monitoring systems and it analyzes the airplanes noise signals and a method to identify them. The method uses processed spectral patterns and a neuronal network feed-forward, programmed by means of virtual instruments. The obtained results, very useful in portable systems, make possible to introduce redundancy to permanent monitoring systems. The noise level in a city has fluctuations between 50 dB (A) and 100 dB (A). It depends on the population density and its activity, commerce and services in the public thoroughfare, terrestrial and aerial urban traffic, of the typical activities of labor facilities and used machinery, which give varied conditions that must be faced of diverse ways within the corresponding normalization. The sounds or noises that exceed the permissible limits, whichever the activities or causes that originate them, are considered events susceptible to degrade the environment and the health.

Reconocimiento automático de voz emotiva con memorias asociativas Alfa-Beta SVM

Resumen —Una de las de investigacion de mayor interes y con mas crecimiento en la actualidad, dentro del area de procesamiento de voz, es el reconocimiento automatico de emociones, el cual consta de 2 etapas; la primera es la extraccion de parametros a partir de la senal de voz y la segunda es la eleccion del modelo para hacer la tarea de clasificacion. La problematica que actualmente existe es que no se han identificado aun los parametros mas representativos del problema ni tampoco se ha encontrado al mejor clasificador para hacer la tarea. En este articulo se introduce un nuevo modelo asociativo de reconocimiento automatico de voz emotiva basado en las maquinas asociativas Alfa-Beta SVM, cuyas entradas se han codificado como representaciones bidimensionales de la energia de las senales de voz. Los resultados experimentales muestran que este modelo es competitivo en la tarea de clasificacion automatica de emociones a partir de senales de voz [1].

Mexico City urban noise control network.

This paper presents an IP network model for monitoring urban noise emitted from mobile and fixed sources. The model is applying to build a monitoring network in the historic center of Mexico City. Design of network system includes measuring station design, network design, central of monitoring design, measuring politics and functionality, event detection modules, central database of historic information, report module design, and administration system. Mexico City Urban Noise Control Network is going to be an important tool to enforce application of local noise regulations, detection of infractions for excessive noise, detection of security events like gun shots, and monitoring of noise pollution levels. Measuring stations are developed to detect average standard levels and individual events, to generate warnings and alarms, and to detect some specific events. Network is developed over a wireless Network.

Using adaptive filter and wavelets to increase automatic speech recognition rate in noisy environment

This paper shows results obtained in the Automatic Speech Recognition (ASR) task for a corpus of digits speech files with a determinate noise level immerse. In the experiments, we used several speech files that contained Gaussian noise. We used HTK (Hidden Markov Model Toolkit) software of Cambridge University in the experiments. The noise level added to the speech signals was varying from fifteen to forty dB increased by a step of 5 units. We used an adaptive filtering to reduce the level noise (it was based in the Least Measure Square -LMS- algorithm) and two different wavelets (Haar and Daubechies). With LMS we obtained an error rate lower than if it was not present and it was better than wavelets employed for this experiment of Automatic Speech Recognition. For decreasing the error rate we trained with 50% of contaminated and originals signals to the ASR system. The results showed in this paper are focused to try analyses the ASR performance in a noisy environment and to demonstrate that if we are controlling the noise level and if we know the application where it is going to work, then we can obtain a better response in the ASR tasks. Is very interesting to count with these results because speech signal that we can find in a real experiment (extracted from an environment work, i.e.), could be treated with these technique and we can decrease the error rate obtained. Finally, we report a recognition rate of 99%, 97.5% 96%, 90.5%, 81% and 78.5% obtained from 15, 20, 25, 30, 35 and 40 noise levels, respectively when the corpus mentioned before was employed and LMS algorithm was used. Haar wavelet level 1 reached up the most important results as an alternative to LMS algorithm, but only when the noise level was 40 dB and using original corpus.

Using adaptive filter to increase automatic speech recognition rate in a digit corpus

This paper shows results obtained in the Automatic Speech Recognition (ASR) task for a corpus of digits speech files with a determinate noise level immerse. The experiments realized treated with several speech files that contained Gaussian noise. We used HTK (Hidden Markov Model Toolkit) software of Cambridge University in the experiments. The noise level added to the speech signals was varying from fifteen to forty dB increased by a step of 5 units. We used an adaptive filtering to reduce the level noise (it was based in the Least Measure Square -LMS- algorithm). With LMS we obtained an error rate lower than if it was not present. It was obtained because of we trained with 50% of contaminated and originals signals to the ASR. The results showed in this paper to analyze the ASR performance in a noisy environment and to demonstrate that if we have controlling the noise level and if we know the application where it is going to work, then we can obtain a better response in the ASR tasks. Is very interesting to count with these results because speech signal that we can find in a real experiment (extracted from an environment work, i.e.), could be treated with these technique and decrease the error rate obtained. Finally, we report a recognition rate of 99%, 97.5% 96%, 90.5%, 81% and 78.5% obtained from 15, 20, 25, 30, 35 and 40 noise levels, respectively when the corpus that we mentioned above was employed. Finally, we made experiments with a total of 2600 sentences (between noisy and filtered sentences) of speech signal.

Implementation of the Macro and Micro Mechanical Cochlea Model in a GPU

For a long time, cochlea models have been an interesting area of study for scientists in different fields such as medicine, especially in otorhinolaryngology, physics and acoustic engineering, among others. That is because, in mammals, this organ is the most important element in the transduction of the sound pressure that is received by the outer and middle ear.

Using Values of the Human Cochlea in the Macro and Micro Mechanical Model for Automatic Speech Recognition

Recently the parametric representation using cochlea behavior has been used in different studies related with Automatic Speech Recognition (ASR). That is because this hearing organ in mammalians is the most important element used to make a transduction of the sound pressure that is received by the outer ear. This paper shows how the macro and micro mechanical model is used in ASR tasks. The values that Neely, Elliot and Ku founded in their works, related with the macro and micro mechanical model such as Neely were used to set the central frequencies of a bank filter to obtain parameters from the speech in a similar form as MFCC (Mel Frequency Cepstrum Coefficients) has been constructed.

Simulation of a model of the basilar membrane in two dimensions for Spanish vowels.

The inner ear has the cochlea as the principal element; this is a biological element in the form of a snail, within which the mechanical energy is converted into electrical energy. This process is realized by the inner ear cells on the basilar membrane. This membrane response is different for different frequencies of excitation; the result of this process is the human audition. This paper shows a simulation of a model in two dimensions of the basilar membrane and its characteristic response when excited by the tow first formants of Spanish vowels; these formants are obtained by an analysis by mixtures of Gaussians.

Gaussian wavelet functions modulated for audible response of the human ear.

One of the most important aspects related with automatic speech recognition (ASRs) systems is to find a set of characteristics that represent speech signal; nowadays, LPC, CLPC, cepstrum, MFCC, LFCC, and Melspec, among others, have been used to solve that problem. Likewise, analyzing the audible behavior of the ear has shown to have good performance in comparison with those methods that consider the pronunciation as element to obtain speech signal representation. In this paper an analysis related with Gaussian wavelets representation based on Bark and Mel audible representations is presented. For these wavelets, a Gaussian function represents the basis decomposition function; further, we need a modulator to displace the function associated with each of the scales of analysis to the central frequency into Bark and Mel audible representations. At the same time, we use the variance parameter of the Gaussian function to adapt its bandwidth to the Bark and Mel audible representations. Finally, we show a comparison between two alternatives in time and frequency representations.

Automatic speech recognition technical for Náhuatl language isolated words.

Automatic speech recognition (ASR) represents a great expectative in communications systems where interaction between computers and humans is present. In the last three decades it has been grounded. The use of computer techniques helps people who speak in a native language or with physical limitation to make transactions, leave messages, obtain information, or control some device using voice expressions. This work shows results obtained when using different techniques of ASR into the Nahuatl language. Nahuatl language is a native language spoken in Mexico and is the most important because there are many people speaking it. Nahuatl language has around 49 variants. Preserving language is important because language is the most efficient means of transmitting a culture and it is the owners of that culture who lose the most when the language dies. Mel cepstrum coefficients, vector quantization, and hidden Markov models of continuous density were used. The experiments reported 99% accuracy in ASR on Nahuatl num...