Estimation of Noise Magnitude for Speech Denoising Using Minima-Controlled-Recursive-Averaging Algorithm Adapted by Harmonic Properties

Estimation of Noise Magnitude for Speech Denoising Using Minima-Controlled-Recursive-Averaging Algorithm Adapted by Harmonic Properties

The accuracy of noise estimation is important for the performance of a speech denoising system. Most noise estimators suffer from either overestimation or underestimation on the noise level. An overestimate on noise magnitude will cause serious speech distortion for speech denoising. Conversely, a g...

Full description

Bibliographic Details
Main Authors: Ching-Ta Lu, Chung-Lin Lei, Jun-Hong Shen, Ling-Ling Wang, Kun-Fu Tseng
Format: Article
Language:English
Published: MDPI AG 2016-12-01
Series:Applied Sciences
Subjects:
noise estimation
variable segment length
speech denoising
harmonic adaptation
minimum-controlled-recursive-controlled averaging
Online Access:http://www.mdpi.com/2076-3417/7/1/9

Internet

http://www.mdpi.com/2076-3417/7/1/9

Similar Items