Design of ready-made acoustic model library by two-dimensional visualization of acoustic space

2004 ◽  
Author(s):  
Goshu Nagino ◽  
Makoto Shozakai
Keyword(s):  
Acoustic Model ◽  
Two Dimensional ◽  
Model Library ◽  
Acoustic Space

Prediction of Rail and Bridge Noise in Near- and Far-Field: A Combined 2.5-Dimensional and Two-Dimensional Method

2016 ◽  
Vol 139 (1) ◽  
Author(s):  
X. D. Song ◽  
Q. Li ◽  
D. J. Wu
Keyword(s):  
Interaction Analysis ◽  
Near Field ◽  
Three Dimensional ◽  
Shielding Effect ◽  
Far Field ◽  
Acoustic Model ◽  
Two Dimensional ◽  
Box Girder ◽  
Total Noise ◽  
Field Noise

Bridge noise and rail noise induced by passing trains should be included while estimating low- and medium-frequency (20–1000 Hz) noise in railway viaducts. However, the prediction of bridge noise and rail noise using a three-dimensional (3D) acoustic model is not efficient, especially for far-field points. In this study, a combined 2.5-dimensional (2.5D) and two-dimensional (2D) method is proposed to predict bridge noise and rail noise in both the near- and far-field. First, the near-field noise is obtained by combining the 2.5D acoustic model and a 3D vehicle–track–bridge interaction analysis. Then, the 2D method is used to estimate the attenuation of bridge noise and rail noise in the far-field, and the accuracy is validated through comparison with the 2.5D method. Third, the near-field points are treated as reference sources, and the noise at far-field points is predicted by combining the 2.5D and 2D methods. Finally, the proposed method is used to predict the bridge noise and rail noise for a box girder and a U-shaped girder. The spatial distribution of the bridge noise and rail noise is investigated. Generally, the rail noise is dominant above the bridge, and the bridge noise has a larger contribution to the total noise beneath the bridge. The rail noise from the U-shaped girder is much smaller than that from the box girder due to the shielding effect of the webs.

scholarly journals Optimal generative and discriminative acoustic model training for speech recognition

2021 ◽  
Author(s):  
Neil Joshi
Keyword(s):  
Time Series ◽  
Speech Recognition ◽  
Objective Function ◽  
Stochastic Models ◽  
Process Models ◽  
Classification Error ◽  
Support Vector ◽  
Acoustic Model ◽  
Single Process ◽  
Acoustic Space

The focus of this dissertation is to derive and demonstrate effective stochastic models for the speech recognition problem. Acoustic modeling for speech recognition typically involves representing the speech process within stochastic models. Modeling this high frequency time series effectively is a fundamental problem. This dissertation devised an objective function that relates the true speech distribution to its estimate. It is shown that through optimizing this function the speech process time series can be modeled without loss of information. The thesis proposes two such models that are developed to optimize the devised objective function. The first an acoustic model formulated for the speech with noise problem. The second a discriminately trained model consisting of optimal discriminant ML estimators. The first, a combination of recognizers that through a simple system fusion, combines multiple speech processes at the decision level. This is a stochastic modeling method devised to combine a parameterized spectral missing data, MD, theory based and a cepstral based speech process using a coupled hidden variable topology. In using a fused coupled hidden Markov model, HMM, topology, an optimal acoustic model is proposed that is inherently more robust than single process models under noisy conditions. The theoretical capability of this model is tested under both stationary and non stationary noise conditions. Under these test conditions the fused model has greater recognition accuracies than those of single process models. The second, formulated with a methodology that segments the acoustic space appropriately for discriminately trained models that optimize the devised objective function. This acoustic space is modeled with discriminant ML estimators formed with optimal decision boundaries using the large margin, support vector machine, SVM, learning method. These discriminately trained models maximize the entropy of the observation space and thereby are capable to model the speech process without loss. This is demonstrated experimentally with frame level classification error rates that are ∼ ≤ 3%.

Post-Wildfire Numerical Modeling for Flood Risk Management

2020 ◽  
Author(s):  
Ian Floyd ◽  
Stanford Gibson ◽  
Gaurav Savant ◽  
Alejandro Sanchez ◽  
Ronald Heath
Keyword(s):  
Shallow Water ◽  
Debris Flows ◽  
Large Scale ◽  
Numerical Models ◽  
The United States ◽  
Water Model ◽  
Sediment Runoff ◽  
Two Dimensional ◽  
Model Framework ◽  
Model Library

<p>The number and intensity of large wildfires in is a growing concern in the United States.  Over the past decade, the National Interagency Fire Centre (NSTC, 2015) reported increases of large fires in every western state in the arid and semi-arid western U.S.  Wildfires, remove vegetation, reduce organic soil horizons to ash, extirpate microbial communities, alters soil structure, and potential development of hydrophobic soils.  These processes all increase water and sediment runoff. Post-wildfire environments can cause a spectrum of hydrologic and sedimentation responses ranging from no response to catastrophic floods and deadly debris flows. Numerical modellers have developed a variety of Newtonian and non-Newtonian shallow-water algorithms to simulate each of these physical processes – making it difficult to model the range of post-wildfire flood conditions and understand model assumption and limitations. This makes a modular non-Newtonian computation library advantageous. This work presents a flexible, numerical model, library framework ‘DebrisLib’ to simulate large-scale, post-wildfire non-Newtonian flows using diverse shallow-water parents code architecture. This work presents the non-Newtonian model framework effectiveness by linking it with two different modelling frameworks, specifically the diffusive-wave one-dimensional and two-dimensional Hydrologic Engineering Center River Analysis System (HEC-RAS), and shallow-water two-dimensional Adaptive Hydraulics (AdH) numerical models. The model library was verified and validated using three flume experiments for mud flows, hyperconcentrated flows, and debris flows under steady and unsteady flow conditions. Additionally the shallow-water model library framework linked with the 1D Hydrologic Engineering Centre Hydrologic Modelling System (HEC-HMS) successfully predicted the 2018 post-wildfire flooding and debris flows following the 2017 Thomas Fire near Santa Barbara, California.</p>

scholarly journals Active noise control for two dimensional acoustic space by concentrated mass model

2017 ◽  
Vol 83 (854) ◽  
pp. 17-00305-17-00305
Author(s):  
Shotaro HISANO ◽  
Satoshi ISHIKAWA ◽  
Shinya KIJIMOTO ◽  
Yosuke KOBA
Keyword(s):  
Noise Control ◽  
Active Noise Control ◽  
Two Dimensional ◽  
Concentrated Mass ◽  
Mass Model ◽  
Active Noise ◽  
Acoustic Space

scholarly journals Optimal generative and discriminative acoustic model training for speech recognition

2021 ◽  
Author(s):  
Neil Joshi
Keyword(s):  
Time Series ◽  
Speech Recognition ◽  
Objective Function ◽  
Stochastic Models ◽  
Process Models ◽  
Classification Error ◽  
Support Vector ◽  
Acoustic Model ◽  
Single Process ◽  
Acoustic Space

The focus of this dissertation is to derive and demonstrate effective stochastic models for the speech recognition problem. Acoustic modeling for speech recognition typically involves representing the speech process within stochastic models. Modeling this high frequency time series effectively is a fundamental problem. This dissertation devised an objective function that relates the true speech distribution to its estimate. It is shown that through optimizing this function the speech process time series can be modeled without loss of information. The thesis proposes two such models that are developed to optimize the devised objective function. The first an acoustic model formulated for the speech with noise problem. The second a discriminately trained model consisting of optimal discriminant ML estimators. The first, a combination of recognizers that through a simple system fusion, combines multiple speech processes at the decision level. This is a stochastic modeling method devised to combine a parameterized spectral missing data, MD, theory based and a cepstral based speech process using a coupled hidden variable topology. In using a fused coupled hidden Markov model, HMM, topology, an optimal acoustic model is proposed that is inherently more robust than single process models under noisy conditions. The theoretical capability of this model is tested under both stationary and non stationary noise conditions. Under these test conditions the fused model has greater recognition accuracies than those of single process models. The second, formulated with a methodology that segments the acoustic space appropriately for discriminately trained models that optimize the devised objective function. This acoustic space is modeled with discriminant ML estimators formed with optimal decision boundaries using the large margin, support vector machine, SVM, learning method. These discriminately trained models maximize the entropy of the observation space and thereby are capable to model the speech process without loss. This is demonstrated experimentally with frame level classification error rates that are ∼ ≤ 3%.

Spectrophotometric measurements of objective-prism spectra

1966 ◽  
Vol 24 ◽  
pp. 118-119
Author(s):  
Th. Schmidt-Kaler
Keyword(s):  
Progress Report ◽  
Two Dimensional ◽  
Objective Prism ◽  
Made In

I should like to give you a very condensed progress report on some spectrophotometric measurements of objective-prism spectra made in collaboration with H. Leicher at Bonn. The procedure used is almost completely automatic. The measurements are made with the help of a semi-automatic fully digitized registering microphotometer constructed by Hög-Hamburg. The reductions are carried out with the aid of a number of interconnected programmes written for the computer IBM 7090, beginning with the output of the photometer in the form of punched cards and ending with the printing-out of the final two-dimensional classifications.

Introductory paper

1966 ◽  
Vol 24 ◽  
pp. 3-5
Author(s):  
W. W. Morgan
Keyword(s):  
Line Intensity ◽  
Classification Scheme ◽  
Two Dimensional ◽  
Spectral Classification ◽  
Dimensional Array ◽  
Rr Lyrae ◽  
Metallic Line ◽  
Introductory Paper ◽  
Parameter Varying ◽  
Definition Of

1. The definition of “normal” stars in spectral classification changes with time; at the time of the publication of theYerkes Spectral Atlasthe term “normal” was applied to stars whose spectra could be fitted smoothly into a two-dimensional array. Thus, at that time, weak-lined spectra (RR Lyrae and HD 140283) would have been considered peculiar. At the present time we would tend to classify such spectra as “normal”—in a more complicated classification scheme which would have a parameter varying with metallic-line intensity within a specific spectral subdivision.

Sign in / Sign up

Export Citation Format

Share Document