基于子带DCT和EMD的混合软阈值语音增强

Q2 Physics and Astronomy

Advances in Acoustics and Vibration Pub Date : 2014-05-20 DOI:10.1155/2014/765454

Erhan Deger, M. I. Molla, K. Hirose, N. Minematsu, Md. Kamrul Hasan

{"title":"基于子带DCT和EMD的混合软阈值语音增强","authors":"Erhan Deger, M. I. Molla, K. Hirose, N. Minematsu, Md. Kamrul Hasan","doi":"10.1155/2014/765454","DOIUrl":null,"url":null,"abstract":"This paper presents a two-stage soft thresholding algorithm based on discrete cosine transform (DCT) and empirical mode decomposition (EMD). In the first stage, noisy speech is decomposed into eight frequency bands and a specific noise variance is calculated for each one. Based on this variance, each band is denoised using soft thresholding in DCT domain. The remaining noise is eliminated in the second stage through a time domain soft thresholding strategy adapted to the intrinsic mode functions (IMFs) derived by applying EMD on the signal obtained from the first stage processing. Significantly better SNR improvement and perceptual speech quality results for different noise types prove the superiority of the proposed algorithm over recently reported techniques.","PeriodicalId":44068,"journal":{"name":"Advances in Acoustics and Vibration","volume":"2014 1","pages":"1-11"},"PeriodicalIF":0.0000,"publicationDate":"2014-05-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1155/2014/765454","citationCount":"5","resultStr":"{\"title\":\"Subband DCT and EMD Based Hybrid Soft Thresholding for Speech Enhancement\",\"authors\":\"Erhan Deger, M. I. Molla, K. Hirose, N. Minematsu, Md. Kamrul Hasan\",\"doi\":\"10.1155/2014/765454\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"This paper presents a two-stage soft thresholding algorithm based on discrete cosine transform (DCT) and empirical mode decomposition (EMD). In the first stage, noisy speech is decomposed into eight frequency bands and a specific noise variance is calculated for each one. Based on this variance, each band is denoised using soft thresholding in DCT domain. The remaining noise is eliminated in the second stage through a time domain soft thresholding strategy adapted to the intrinsic mode functions (IMFs) derived by applying EMD on the signal obtained from the first stage processing. Significantly better SNR improvement and perceptual speech quality results for different noise types prove the superiority of the proposed algorithm over recently reported techniques.\",\"PeriodicalId\":44068,\"journal\":{\"name\":\"Advances in Acoustics and Vibration\",\"volume\":\"2014 1\",\"pages\":\"1-11\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2014-05-20\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://sci-hub-pdf.com/10.1155/2014/765454\",\"citationCount\":\"5\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Advances in Acoustics and Vibration\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1155/2014/765454\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q2\",\"JCRName\":\"Physics and Astronomy\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Advances in Acoustics and Vibration","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1155/2014/765454","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"Physics and Astronomy","Score":null,"Total":0}

引用次数: 5

摘要

提出了一种基于离散余弦变换(DCT)和经验模态分解(EMD)的两阶段软阈值算法。在第一阶段，将含噪语音分解为8个频带，并计算每个频带的特定噪声方差。基于该方差，在DCT域中使用软阈值对每个波段进行去噪。在第二阶段，通过时域软阈值策略来消除剩余的噪声，该策略适应于对第一阶段处理得到的信号应用EMD得到的本征模态函数(IMFs)。在不同噪声类型下显著提高信噪比和感知语音质量，证明了该算法优于最近报道的技术。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

查看原文本刊更多论文

Subband DCT and EMD Based Hybrid Soft Thresholding for Speech Enhancement

This paper presents a two-stage soft thresholding algorithm based on discrete cosine transform (DCT) and empirical mode decomposition (EMD). In the first stage, noisy speech is decomposed into eight frequency bands and a specific noise variance is calculated for each one. Based on this variance, each band is denoised using soft thresholding in DCT domain. The remaining noise is eliminated in the second stage through a time domain soft thresholding strategy adapted to the intrinsic mode functions (IMFs) derived by applying EMD on the signal obtained from the first stage processing. Significantly better SNR improvement and perceptual speech quality results for different noise types prove the superiority of the proposed algorithm over recently reported techniques.

求助全文

通过发布文献求助，成功后即可免费获取论文全文。去求助

来源期刊

Advances in Acoustics and Vibration ACOUSTICS-

自引率

0.00%

发文量

期刊介绍： The aim of Advances in Acoustics and Vibration is to act as a platform for dissemination of innovative and original research and development work in the area of acoustics and vibration. The target audience of the journal comprises both researchers and practitioners. Articles with innovative works of theoretical and/or experimental nature with research and/or application focus can be considered for publication in the journal. Articles submitted for publication in Advances in Acoustics and Vibration must neither have been published previously nor be under consideration elsewhere. Subject areas include (but are not limited to): Active, semi-active, passive and combined active-passive noise and vibration control Acoustic signal processing Aero-acoustics and aviation noise Architectural acoustics Audio acoustics, mechanisms of human hearing, musical acoustics Community and environmental acoustics and vibration Computational acoustics, numerical techniques Condition monitoring, health diagnostics, vibration testing, non-destructive testing Human response to sound and vibration, Occupational noise exposure and control Industrial, machinery, transportation noise and vibration Low, mid, and high frequency noise and vibration Materials for noise and vibration control Measurement and actuation techniques, sensors, actuators Modal analysis, statistical energy analysis, wavelet analysis, inverse methods Non-linear acoustics and vibration Sound and vibration sources, source localisation, sound propagation Underwater and ship acoustics Vibro-acoustics and shock.