Subband Noise Reduction Methods for Speech Enhancement

2006 ◽  
pp. 91-115 ◽  
Author(s):  
Eric J. Diethorn
Keyword(s):  
Noise Reduction ◽  
Speech Enhancement ◽  
Reduction Methods

scholarly journals Dual-Channel Speech Enhancement Based on Speech Presence Probability

2020 ◽  
Vol 9 (3) ◽  
pp. 2198-2200
Keyword(s):  
Noise Reduction ◽  
Speech Enhancement ◽  
Hearing Aids ◽  
Speech Processing ◽  
Speech Signal ◽  
Reduction Process ◽  
Coherent Noise ◽  
Residual Noise ◽  
Speech Distortion ◽  
Reduction Methods

This paper introduces technology to improve sound quality, which serves the needs of media and entertainment. Major challenging problem in the speech processing applications like mobile phones, hands-free phones, car communication, teleconference systems, hearing aids, voice coders, automatic speech recognition and forensics etc., is to eliminate the background noise. Speech enhancement algorithms are widely used for these applications in order to remove the noise from degraded speech in the noisy environment. Hence, the conventional noise reduction methods introduce more residual noise and speech distortion. So, it has been found that the noise reduction process is more effective to improve the speech quality but it affects the intelligibility of the clean speech signal. In this paper, we introduce a new model of coherence-based noise reduction method for the complex noise environment in which a target speech coexists with a coherent noise around. From the coherence model, the information of speech presence probability is added to better track noise variation accurately; and during the speech presence and speech absent period, adaptive coherence-based method is adjusted. The performance of suggested method is evaluated in condition of diffuse and real street noise, and it improves the speech signal quality less speech distortion and residual noise.

scholarly journals On the Running Noise of Toothed Belt Drive : 2nd Report, Influence of Running Condition and Some Noise Reduction Methods

1971 ◽  
Vol 37 (293) ◽  
pp. 203-211
Author(s):  
Aizoh KUBO ◽  
Toshiaki ANDO ◽  
Susumu SATO ◽  
Toshio AIDA ◽  
Takeshi HOSHIRO
Keyword(s):  
Noise Reduction ◽  
Belt Drive ◽  
Toothed Belt ◽  
Reduction Methods

COMPARISON OF MONOPOLE AND DIPOLE NOISE REDUCTION METHODS

2020 ◽  
Vol 4 (67) ◽  
pp. 153-160
Author(s):  
Oleg I. Polivaev ◽  
◽  
Alexey N. Kuznetsov ◽  
Dmitriy Yu. Terekhov ◽  
Viktor V. Trufanov ◽  
...  
Keyword(s):  
Noise Reduction ◽  
Reduction Methods

Noise Reduction During Acoustic Monitoring of Laser Welding of High-Strength Steels

2009 ◽  
Author(s):  
Wei Huang ◽  
Radovan Kovacevic
Keyword(s):  
Laser Welding ◽  
Noise Reduction ◽  
Acoustic Signal ◽  
Welding Process ◽  
High Strength Steels ◽  
High Strength ◽  
Weld Penetration ◽  
Acoustic Signatures ◽  
Reduction Methods ◽  
Laser Welding Process

During the laser welding process of high-strength steels, different defects, such as a partial weld penetration, spatters, and blow-through holes could be present. In order to detect the presence of defects and achieve a quality control, acoustic monitoring based on microphones is applied to the welding process. As an effective sensor to monitor the laser welding process, however, the microphone is greatly limited by intensive noise existing in the complex industrial environment. In this paper, in order to acquire a clean acoustic signal from the laser welding process, two noise reduction methods are proposed: one is the spectral subtraction method based on one microphone and the other one is the beamforming based on a microphone array. By applying these two noise reduction methods, the quality of the acoustic signal is enhanced, and the acoustic signatures are extracted both in the time domain and frequency domain. The analysis results show that the extracted acoustic signatures can well indicate the different weld penetration states and they can also be used to study the internal mechanisms of the laser-material interaction.

scholarly journals Improved Turbulence Profiling with Field-Adapted Acoustic Doppler Velocimeters Using a Bifrequency Doppler Noise Suppression Method

2008 ◽  
Vol 25 (3) ◽  
pp. 452-463 ◽  
Author(s):  
D. Hurther ◽  
U. Lemmin
Keyword(s):  
Noise Reduction ◽  
Carrier Frequency ◽  
Noise Suppression ◽  
Isotropic Turbulence ◽  
Field Studies ◽  
Sample Volume ◽  
Grid Turbulence ◽  
Electronic Circuitry ◽  
Reduction Methods ◽  
Velocity Spectra

Abstract A novel noise reduction method and corresponding technique are presented for improving turbulence measurements with acoustic Doppler velocimeters (ADVs) commonly used in field studies of coastal and nearshore regions, rivers, lakes, and estuaries. This bifrequency method is based on the decorrelation of the random and statistically independent Doppler noise terms contained in the Doppler signals at two frequencies. It is shown through experiments in an oscillating grid turbulence (OGT) tank producing diffusive isotropic turbulence that a shift in carrier frequency of less than 10% is sufficient to increase the resolved frequency range by a decade in the turbulent velocity spectra. Over this spectral range, the slope of the velocity spectra agrees well with the universal inertial range value of −5/3. The limit due to spatial averaging effects over the sample volume can be determined from the abrupt deviation of the spectral slope from the −5/3 value. As a result, the relative error of the turbulent intensity estimate and the turbulent kinetic energy (TKE) dissipation rate, measured by two different methods, does not exceed 10% in the case of isotropic turbulence. Furthermore, the bifrequency method allows accurate estimates of the turbulent microscales as shown by the good agreement of the ratio between the Taylor and Kolmogorov microscales and an Re1/4t power law. Compared to previous Doppler noise reduction methods (Garbini et al.), an increase in time resolution by a factor of 4 is achieved. The proposed method also avoids the loss of TKE energy contained in isotropic flow structures of size equal to and smaller than the sample volume. Different from Doppler noise methods proposed by Hurther and Lemmin and Blanckaert and Lemmin, this method does not require additional hardware components, electronic circuitry, or sensors because the redundant instantaneous velocity field information is captured with the same transducer. The required shift in carrier frequency is small enough for the bifrequency method to be easily implemented in commercial ADVs.

Sign in / Sign up

Export Citation Format

Share Document