• The Journal of the Acoustical Society of Korea
• /
• v.38 no.5
• /
• pp.607-613
• /
• 2019

This paper proposes LF-MMI (Lattice Free Maximum Mutual Information)-based acoustic modeling using external knowledge for speech recognition. Note that an external knowledge refers to text data other than training data used in acoustic model. LF-MMI, objective function for optimization of training DNN (Deep Neural Network), has high performances in discriminative training. In LF-MMI, a phoneme probability as prior probability is used for predicting posterior probability of the DNN-based acoustic model. We propose using external knowledges for training the prior probability model to improve acoustic model based on DNN. It is measured to relative improvement 14 % as compared with the conventional LF-MMI-based model.

• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.20 no.1
• /
• pp.74-82
• /
• 2010

Power flow analysis(PFA) methods have shown many advantages in noise predictions and vibration analysis in medium-to-high frequency ranges. Applying the finite element technique to PFA has produced power flow finite element method(PFFEM) that can be effectively used for analysis of vibration of complicated structures. PFADS(power flow analysis design system) based on PFFEM as the vibration analysis program has been developed for vibration predictions and analysis of coupled structural systems. In this paper, to improve the function of vibro-acoustic coupled analysis in PFADS, the PFFEM has been extended for analysis of the interior noise problems in the vibro-acoustic fully coupled systems. The vibro-acoustic fully coupled PFFEM formulation based on energy coupled relations is extended to structural system model by using appropriate modifications to structural-structural, structural-acoustic and acoustic-acoustic joint matrices. It has been applied to prediction of the interior noise in two room model coupled with panels, and the PFFEM results are compared to those of statistical energy analysis(SEA).

• Proceedings of the Korean Society of Propulsion Engineers Conference
• /
• 2008.05a
• /
• pp.18-21
• /
• 2008

Acoustic design parameters of a half-wave resonator are studied experimentally for acoustic stability in a model combustor. According to standard acoustic-test procedures, acoustic-pressure signals are measured. Quantitative acoustic properties of damping factor and sound absorption coefficient are evaluated and thereby, the acoustic damping capacity of the resonator is characterized. The diameter and the number of a half-wave resonator, its distribution are selected as design parameters for optimal tuning of the resonator. Acoustic damping capacity increases as the resonators with diameter increases. The optimum number of resonators or the optimum open-area ratio decreases as boundary absorption decreases.

• Journal of the Society of Naval Architects of Korea
• /
• v.43 no.5 s.149
• /
• pp.586-596
• /
• 2006

A 3-dimensional FDLB model with additional internal degree of freedom is applied for diatomic gases such as air, in which an additional distribution function is introduced. Direct simulations of aero-acoustic by using the applied model and scheme are presented. Speed of sound is correctly recovered. As typical examples, the Aeolian tone emitted by a circular column is successfully simulated even very low Mach number flow. Acoustic pressure fluctuations with the same frequency of the Karman vortex street compared with the pressure fluctuation around a circular column is captured. Full three-dimensional acoustic wave past a compact block like pentagon, furthermore, is also emitted in y direction as dipole like sound.

• Journal of the Society of Naval Architects of Korea
• /
• v.49 no.5
• /
• pp.384-390
• /
• 2012

Power Flow Analysis (PFA) is introduced for solving the noise and vibration analysis of structures in medium-to-high frequency ranges. The vibration analysis software, $PFADS_{C{+}{+}}$ R4 based on Power Flow Finite Element Method (PFFEM) and the noise prediction software, $NASPFA_{C{+}{+}}$ R1 based on Power Flow Boundary Element Method (PFBEM) are developed. In this paper, the coupling equation which represents relation between structural energy and acoustic energy is developed for vibro-acoustic coupling analysis. And vibro-acoustic coupling analysis software based on PFA and coupling equation is developed. Developed software is composed of translator, cavity-finder, solver and post-processor over all. Translator can translate FE model into PFADS FE model and cavity-finder can automatically make NASPFA BE model from PFADS FE model for noise analysis. The solver module calculates the structural energy density, intensity of structures, the fictitious source on the boundary and the acoustic energy density at the field in acoustic cavities. Some applications of vibro-acoustic coupling analysis software to various structures and cruise ship are shown with reliable results.

• Journal of Industrial Technology
• /
• v.17
• /
• pp.87-93
• /
• 1997

A Special purpose program, based on the dimensional analysis, was developed to estimate the wide band turbulent noise in the blower of vehicle airhandling system. Acoustic power level was measured at 4 rating points around the operating condition. The experiment was performed on the reference blower model using international standard chamber, which could measure acoustic power according to the air-handling performance. Analytical model of the blower noise was determined by the measured data. Using the analytical acoustic model, it was possible to estimated the effect by the change of the operating condition, such as flow rate, static pressure and wheel rotating speed, furthermore, the diameter and the width of blower.

• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.12 no.1
• /
• pp.12-20
• /
• 2002

To control the noise of an enclosed sound field, we built a state space model using the acoustic modal parameter description. Using the state space model, we can investigate the controllability and observability, and find an appropriate position of control speaker and microphone to control sound field of the enclosed space. We implemented LQR(linear quadratic regulator) controller and reduced order observer to reduce the first acoustic mode. Experiments showed satisfactory results of 4∼10 dB reduction of magnitude of the first acoustic mode, and support the feasibility of the proposed scheme to lightly damped acoustic field.

• Journal of KSNVE
• /
• v.9 no.3
• /
• pp.493-501
• /
• 1999

The vibration modes of resilient mounting system and foundation structure which support diesel engine/generator set and acoustic enclosure walls play an important role in the vibration transmission process. So, it is necessary to perform vibration mode analysis of resilient mounting system and foundation structure. For some reasons, if the vibration modal analysis of resilient mounting system and foundation structure of acoustic enclosure could be simultaneously done by finite element method, it would be very efficient approach. In this paper, vibration modal analysis method using finite element method for multi stage mounting system having n d.o.f model was proposed. Vibration analysis of single and double stage resilient mounting system was performed to verify the validity of the proposed method. Also frequency response results were compared in case of rigid foundation model and finite element foundation model which was compared with experimental modal analysis results.

• MALSORI
• /
• no.66
• /
• pp.117-130
• /
• 2008

This paper presents a method for automatically detecting Korean prosodic boundaries using both acoustic and grammatical information for the performance improvement of speech information processing systems. While most of previous works are solely based on grammatical information, our method utilizes not only grammatical information constructed by a Maximum-Entropy-based grammar model using 10 grammatical features, but also acoustical information constructed by a GMM-based acoustic model using 14 acoustic features. Given that Korean prosodic structure has two intonationally defined prosodic units, intonation phrase (IP) and accentual phrase (AP), experimental results show that the detection rate of AP boundaries is 82.6%, which is higher than the labeler agreement rate in hand transcribing, and that the detection rate of IP boundaries is 88.7%, which is slightly lower than the labeler agreement rate.

• Proceedings of the Korean Society of Fisheries Technology Conference
• /
• 2009.05a
• /
• pp.38-41
• /
• 2009

The jellyfish Nemopilema nomurai has occurred in large numbers in Japan and Korea and has had on a negative effect on coastal fisheries in this region. Data on the abundance and distribution of jellyfish are needed to forecast when and where they will occur in coastal areas. Acoustic techniques are commonly used to study the distribution and abundance of fish and zooplankton. However before such surveys can be conducted, the acoustic characteristics of the target species must be known. In this study, the density of and speed-of-sound in jellyfish were measured to clarify their acoustic characteristics using a theoretical scattering model. The acoustic characteristics were estimated with the distorted-wave Born approximation (DWBA) model using these material properties and the shape of free-swimming jellyfish.