Search - dct speech

Search - dct speech - List

[Speech/Voice recognition/combine] 1004 DL : 0: 介绍基于SVM模型的语音识别方法的处理与研究，通过把SVM 与DCT相结合来进行语音的识别。-Introduced the SVM model based on speech recognition method of treatment and research, through the combination of SVM with the DCT for speech recognition.
Date : 2026-01-01 Size : 612kb User : kellan
[Speech/Voice recognition/combine] The-research-of-anti-niose-speech DL : 0: 论文首先介绍了传统的语音特征参数MFCC，它是基于人耳听觉特性设计的一种特征参数，在静音环境下能得到较高的识别率，但在信噪比较低时识别率急剧下降，不利于实用化。本文通过对MFCC算法的分析和研究，发现其中的FFT和DCT在整个时频空间使用固定的。分析窗，这不符合语音信号特性，而小波变换具有多分辨率特性，更符合人耳的听觉特性。因此，本文将小波变换和MFCC算法相结合，提出了三种新的语音识别特-Speech recognition has wide use in the field of communication and so on·Speech feature parameter extraction is an important part of the speech recognition system．The performance of the feature parameter haftuences the system’S performance directly．And the environmental noise is a kev factor of restricting the performance of the feature parameter．This paper，s research object is extracting the speech feature parameter under the noise environment·Then it analyzes the auditory model of human，deeply analyzes and researches the traditional speech feature parameter MFCC， and proposes three improved MFCC parameters based on theⅥ，avelet transformation and the auditory characteristic ofhuman．Besides，the paper also gives all improved method about the feature parameter ZCPA．
Date : 2026-01-01 Size : 1.91mb User : 周卓然
[Speech/Voice recognition/combine] Laplace DL : 0: 传统的短时谱估计语音增强算法通常假设语音谱分量相互独立，没有考虑语音谱分量间的相关性。针对这一问题，该文提出一种新的基于多元Laplace分布模型的短时谱估计算法。首先，假设语音的离散余弦变换(DCT) 系数服从多元Laplace分布，以此利用谱分量间的相关性；在此基础上，利用多元随机矢量的高斯尺度混合模型表示，推导得到语音DCT系数矢量的最小均方误差(MMSE)估计的解析表达式；并进一步推导了基于该分布模型的语音存在概率，对最小均方误差估计子进行修正。实验结果表明，该算法在抑制背景噪声和减少语音失真等方面优于传统的语音增强方法。-The spectral components of speech are usually assumed to be independent in traditional short-time spectrum estimation, which is not the case in practice. Tosolve this problem, a new speech enhancement algorithm with multivariate Laplace speech model is proposed in this paper. Firstly, the speech Discrete Cosine Transform (DCT) coefficients are modeled by a multivariate Laplace distribution, so the correlations between speech spectral components can be exploited. And then a Minimum-Mean-Square-Error (MMSE) estimator based on the proposed model is derived using a Gaussian scale mixture representation of random vectors. Furthermore, the speech presence uncertainty with the new model is derived to modify the MMSE estimator. Experimental results show that the developed method has better noise suppression performance and lower speech distortion compared to the traditional speech enhancement method.
Date : 2026-01-01 Size : 1.01mb User : 立枣酒