Speech Enhancement using Spectral Subtraction-type Algorithms: A Comparison and Simulation Study

Upadhyay, Navneet; Karmakar, Arindam

doi:10.1016/j.procs.2015.06.066

Cited by 98 publications

(37 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The attenuation factor is generally obtained by averaging the time-frequency components of trained background noise. Some variants of spectral subtraction methods proposed over-estimation of the noise spectrum and spectral flooring for non-stationary noises [22,24]. Here, the empirical Wiener attenuation rule is applied to set the attenuation factor because it minimizes the sum of the distortion of signals and energy [25].…”

Section: Noise Reductionmentioning

confidence: 99%

See 1 more Smart Citation

An in-Process Inspection System to Detect Noise Originating from within the Interior Trim Panels of Car Doors

Baek

Kim

2020

Sensors

View full text Add to dashboard Cite

Car body parts are sometimes responsible for irritating noise caused by assembly defects. Typically, various types of noise are known to originate from within the interior trim panels of car doors. This noise is considered to be an important factor that degrades the emotional satisfaction of the driver of the car. This research suggests an in-process inspection system consisting of an inspection workstation and a noise detection method. The inspection workstation presses down the car door trim panel by using a pneumatic pusher while microphones record the acoustic signals directly above the door trim panel and on the four sides of the workstation. The collected signals are analyzed by the proposed noise detection method after applying noise reduction. The noise detection method determines the presence of irritating noise by using noise source localization in combination with the time difference of arrival method and the relative signal strengths. The performance of the in-process noise detection system was evaluated by conducting experiments on faulty and healthy car door trim panels.

show abstract

Section: Noise Reductionmentioning

confidence: 99%

“…where β 1 , β 2 ≥ 0 and λ ≥ 1 are the over-subtraction factors and ξ es i [w, m] is the SNR of the power spectrum. The soft-thresholding method is applied to determine the over-subtraction factors [22,26]; hence, the following values are specified: β 1 = 0.5, β 2 = 1, and λ = 10. ξ es i [w, m] is then represented as follows:…”

Section: Noise Reductionmentioning

confidence: 99%

An in-Process Inspection System to Detect Noise Originating from within the Interior Trim Panels of Car Doors

Baek

Kim

2020

Sensors

View full text Add to dashboard Cite

show abstract

“…Because the speech signal contains noise caused by the acoustic background or the acoustic signals produced by the instrument used for recording the voice, as well as due to the variability of recording scenarios and speaking styles of the subjects, the first step is to normalize the speech signal which is done in the normalization block. For this, we initially apply a noise reduction filter based on spectral subtraction [52], being an efficient method with a low computational cost. We continue with energy normalization in order to compensate for different recording scenarios from different databases.…”

Section: Base Layermentioning

confidence: 99%

Feedforward Neural Network-Based Architecture for Predicting Emotions from Speech

Gavrilescu

Vizireanu

2019

Data

View full text Add to dashboard Cite

We propose a novel feedforward neural network (FFNN)-based speech emotion recognition system built on three layers: A base layer where a set of speech features are evaluated and classified; a middle layer where a speech matrix is built based on the classification scores computed in the base layer; a top layer where an FFNN- and a rule-based classifier are used to analyze the speech matrix and output the predicted emotion. The system offers 80.75% accuracy for predicting the six basic emotions and surpasses other state-of-the-art methods when tested on emotion-stimulated utterances. The method is robust and the fastest in the literature, computing a stable prediction in less than 78 s and proving attractive for replacing questionnaire-based methods and for real-time use. A set of correlations between several speech features (intensity contour, speech rate, pause rate, and short-time energy) and the evaluated emotions is determined, which enhances previous similar studies that have not analyzed these speech features. Using these correlations to improve the system leads to a 6% increase in accuracy. The proposed system can be used to improve human–computer interfaces, in computer-mediated education systems, for accident prevention, and for predicting mental disorders and physical diseases.

show abstract

“…This situation demonstrates the need for improving speech quality in such noisy recordings. Popular approaches for speech enhancement includes beamforming [5,6], spectral subtraction [7], and source separation [8][9][10].…”

Section: Introductionmentioning

confidence: 99%

Exploiting Nonnegative Matrix Factorization with Mixed Group Sparsity Constraint to Separate Speech Signal from Single-channel Mixture with Unknown Ambient Noise

Duong¹,

Nguyen²,

Nguyen³

2018

EAI Endorsed Transactions on Context-aware Systems and Applicat

View full text Add to dashboard Cite

This paper focuses on solving a challenging speech enhancement problem: improving the desired speech from a singlechannel audio signal containing high-level unspecified noise (possibly environmental noise, music, other sounds, etc.). Using source separation technique, we investigate a solution combining nonnegative matrix factorization (NMF) with mixed group sparsity constraint that allows exploiting generic noise spectral model to guide the separation process. The experiment performed on a set of benchmarked audio signals with different types of real-world noise shows that the proposed algorithm yields better quantitative results in term of the signal-to-distortion ratio than the previously published algorithms.

show abstract

Speech Enhancement using Spectral Subtraction-type Algorithms: A Comparison and Simulation Study

Cited by 98 publications

References 12 publications

An in-Process Inspection System to Detect Noise Originating from within the Interior Trim Panels of Car Doors

An in-Process Inspection System to Detect Noise Originating from within the Interior Trim Panels of Car Doors

Feedforward Neural Network-Based Architecture for Predicting Emotions from Speech

Exploiting Nonnegative Matrix Factorization with Mixed Group Sparsity Constraint to Separate Speech Signal from Single-channel Mixture with Unknown Ambient Noise

Contact Info

Product

Resources

About