Speech enhancement method based on speech presence probability and phase estimation

Abstract

The invention belongs to the technical field of signal processing, and relates to a speech enhancement method based on speech presence probability and phase estimation. The method comprises the following steps: (1) estimating the speech presence probability; (2) estimating the pure voiced phase; (3) estimating the pure speech magnitude spectrum; and (4) estimating pure speech signals. The Q value is estimated through a multivariable linear regression technology, and the accuracy of speech presence probability estimation is improved. The pure voiced phase is estimated between adjacent bands using a harmonic model of short-time Fourier transform domain. The pure speech magnitude spectrum is estimated based on the phase difference, and the compensation effect of phase on the pure speech magnitude spectrum is fully utilized. The speech enhancement method of the method is a single-channel speech enhancement method. Speech signals with noise are collected using one microphone. The method is easy to implement. The problem that estimation of speech presence probability is inaccurate can be well solved. The compensation effect of the voiced phase and the phase difference on the pure speech magnitude spectrum is fully utilized. Thus, the performance of the speech enhancement method is improved.

Description

technical field [0001] The present invention relates to a speech enhancement method, in particular to a speech enhancement method based on speech existence probability and phase estimation, which is used in smart phones, hearing aids or smart home audio equipment and human-computer interaction to improve the quality of speech signals, It belongs to the technical field of signal processing. Background technique [0002] Environmental noise tends to degrade the performance of speech processing systems, and to avoid this, speech enhancement techniques are often used. The purpose of speech enhancement is to extract the original speech signal as pure as possible from the noisy speech signal, improve the quality or intelligibility of the speech signal, and then improve the robustness of the speech processing system. Speech enhancement is an important research direction in the field of signal processing, and it is playing an increasingly important role in the fields of speech reco...

AI Technical Summary

Problems solved by technology

[0005] The technical problem to be solved by the present invention is to overcome the inaccurate estimation of the speech presence probability in the existing speech enhancement method based on the speech presence probability, and the poor quality of the speech signal after the enhancement due to the lack of full use of the phase information

Images