文章基本信息

标题：De-noising Speech Signal Using Composite Wavelets and Grazing Estimation Methods
本地全文：下载
作者：Mohammed Anwer ; Nashid Kalam ; Rezwan al-Islam Khan 等
期刊名称：Journal of Emerging Trends in Computing and Information Sciences
电子版ISSN：2079-8407
出版年度：2013
卷号：4
期号：11
页码：827-831
出版社：ARPN Publishers
摘要：The traditional audio de-noising technique using Daubechies wavelet is improved by using a preprocessor using statistical grazing estimation. ‘Noisy’ signals were synthetically generated by introducing ‘noise’ to a ‘clean’ signal. The signal was sampled at 8 kHz and digitized at 16 bits per second. The noise from the audio signal was removed in the first step using statistical grazing estimation with specified threshold level. In the second step Daubechies-D4 wavelet transform was used to remove the noise further. The results were compared with results obtained with grazing estimation only, Daubechies-D4 wavelet transformation and Haar wavelet transformations separately. The results show that the composite denoising technique results in consistently better signal-to-noise ratio than any other technique.
关键词：Speech processing; denoising; Daubechies wavelets; grazing estimation; signal processing.