论文部分内容阅读
语音增强的目的就是要尽可能地消除语音信号中的噪声干扰以获得纯净的语音信号。但噪声都是随机产生的,其特性非常复杂,不利于研究总结,所以,要完全消除噪声是不可能的,我们只能尽可能地提高语音质量,而实际上语音增强的目标就是要提高语音质量。本文在高斯模型的基础上,给出了一种比较新的方法来对其实现语音增强——WDCT算法。WDCT的性能要比传统的DCT更优,但是会多一些计算的复杂度。最后,我们利用Matlab实现了基于WDCT算法的仿真,用结果验证了它的增强效果。
The purpose of speech enhancement is to eliminate the noise interference in the speech signal as much as possible to obtain the pure speech signal. But the noise is randomly generated, its characteristics are very complex, is not conducive to the study summary, so to completely eliminate the noise is impossible, we can only improve voice quality as possible, but in fact the goal of voice enhancement is to improve speech quality. Based on the Gaussian model, this paper presents a relatively new method to achieve its speech enhancement - WDCT algorithm. The performance of the WDCT is better than the traditional DCT, but it will be more computational complexity. Finally, we use Matlab to realize the simulation based on WDCT algorithm, and verify its enhancement effect with the result.