为了更好地分析噪声估计算法的优缺点和性能，本次论文主要利用噪声估计算法

In order to better analyze the advantages and disadvantages and performance of the noise estimation algorithm, this paper mainly uses the effect of the noise estimation algorithm in the speech enhancement technology for effective analysis. The improvement of the speech enhancement technology needs to be based on both noise estimation algorithms, mainly By suppressing background noise and other noises with a noise estimation algorithm, pure speech is extracted from noisy speech. <br>In recent years, social technology has achieved an unprecedented leap with the development of globalization. As a communication and daily communication in people's daily life, voice is an indispensable "necessary" for people. The noise interference received during communication is called background noise. It is precisely because of the presence of background noise that when people communicate, the voice is polluted, and the processing performance and characteristic parameters of the voice are affected. This contaminated speech is called noisy speech. To recover the original speech from the noisy speech, suppress background noise, and extract pure speech, a noise estimation algorithm is needed, which is the source of speech enhancement. When performing speech detection, it may be affected by background noise. At this time, it is necessary to suppress the background noise by the noise estimation algorithm. Therefore, it is of practical significance to study noise estimation algorithms in terms of speech enhancement technology processing and international speech noise research. <br>This thesis is mainly to search and consult several classic noise estimation algorithms to master its performance characteristics, and then select an algorithm with good noise estimation accuracy and robustness.

In order to better analyze the advantages and disadvantages of noise estimation algorithm and performance, this paper mainly uses the effect of noise estimation algorithm in voice enhancement technology to carry out effective analysis, the improvement of voice enhancement technology needs to be based on the noise estimation algorithm, mainly by using noise estimation algorithm to suppress background noise and other noise, from the sound-band voice extracted pure speech.<br>In recent years, social science and technology with the development of globalization has been an unprecedented leap. Voice as a daily communication, remote communication, is an indispensable "essential" people. The noise interference in the process of communication is called background noise, it is precisely because of the existence of background noise makes people communicate, so that the voice is polluted, the speech processing performance, characteristic parameters and so on are affected. This contaminated voice is called noise-banded speech, in order to recover the original voice from noisy speech, suppress background noise, extract pure speech needs to use the noise estimation algorithm, which is the source of voice enhancement. Voice detection may be affected by background noise, which requires the suppression of background noise by the noise estimation algorithm. Therefore, it is of practical significance to study noise estimation algorithm sither in the processing of speech enhancement technology or in the study of international speech noise.<br>This paper mainly through several classic noise estimation algorithms to search and grasp its performance characteristics, and then select a noise estimation accuracy and robust algorithm.

In order to better analyze the advantages, disadvantages and performance of noise estimation algorithm, this paper mainly uses the effect of noise estimation algorithm in speech enhancement technology for effective analysis. The improvement of speech enhancement technology needs to be based on the two noise estimation algorithms, mainly through the use of noise estimation algorithm to suppress background noise and other noise, to extract pure speech from noisy speech 。<br>In recent years, with the development of globalization, social science and technology has made an unprecedented leap. As the communication in people's daily life, voice communication is an indispensable "necessity". The noise interference in the process of communication is called background noise. It is because of the existence of background noise that when people communicate, the voice is polluted, and the speech processing performance and characteristic parameters are affected. This polluted speech is called noisy speech. In order to recover the original speech from noisy speech, suppress background noise and extract pure speech, noise estimation algorithm is needed, which is the source of speech enhancement. In the process of speech detection, it may be affected by the background noise, so the noise estimation algorithm is needed to suppress the background noise. Therefore, it is of practical significance to study the noise estimation algorithm in both speech enhancement technology processing and international speech noise research.<br>In this paper, several classical noise estimation algorithms are searched to master their performance characteristics, and then a noise estimation algorithm with good accuracy and robustness is selected.<br>