Title :
Speech enhancement with zero replacement signal at low SNR
Author :
Yu Iwai;Tetsuya Shimamura
Author_Institution :
Graduate School of Science and Engineering, Saitama University, Shimo-Okubo 255, Sakura-ku, 338-8570, Japan
Abstract :
In this paper, we propose a speech enhancement method with zero replacement signal of noisy speech signal. The proposed method utilizes the Fourier transform of the cross-correlation function between the noisy speech signal and zero replacement signal. Minimum values of the absolute value of the resulting spectrum become the power spectrum of original clean speech signal. For the proposed method, noise estimation is not required to emphasize the speech signal. By subjective evaluation in the presence of both stationary and non-stationary noises, it is shown that the proposed method provides better performance than the conventional method at lower signal-to-noise ratio.
Keywords :
"Speech","Speech enhancement","Noise measurement","Fourier transforms","Signal to noise ratio","Correlation"
Conference_Titel :
Intelligent Signal Processing and Communication Systems (ISPACS), 2015 International Symposium on
DOI :
10.1109/ISPACS.2015.7432754