基于谱减法的语音增强技术讨论 摘 要: 本文主要讨论目的是增强语音质量，减少语音失真和提高其可懂度。首先介绍了语音增强的进展状况及讨论意义，详细对比分析了各种语音增强算法的优缺点，着重讨论了语音增强的相关基本理论。在以上讨论分析的基础上本文采纳了谱减法进行语音增强，给出谱减法的基本原理和实现过程。此外，本文采纳语音激活检测(VAD)方法进行噪声估量。整个算法用 MATLAB 进行模拟仿真，仿真结果表明，所提出的算法在去除背景噪声的同时，保证了较小的语音失真，主观测听效果也比较理想。关键词:语音增强；谱减法；噪声估量；语音激活检测Study of Speech Enhancement Technology Based on Spectral SubtractionAbstract: The purpose of this paper is to improve speech quality, to reduce speech distortion, and improve speech intelligibility. We first introduce the development and significance of speech enhancement, and give a detailed comparison of various speech enhancement algorithms. Especially, we study the basic theories of speech enhancement. Then, on the analysis above, we use spectral subtraction to improve speech quality, and also introduce basic principle of spectral subtraction. Furthermore, we adopt the noise estimator based on Voice Activity Detection (VAD) to estimate noise. In the experiment of this paper, the speech enhancement algorithm is implemented on computer with MATLAB. The simulation result indicates that the algorithm can eliminate noise very well, ensure the smaller voice distortion and also more pleasant to a human listener.Key words: speech enhancement; spectral subtraction; noise estimation; VAD目 录第 1 章 绪论...............................................................................................................................11.1 课题讨论背景.................................................................................................................11.2 语音增强的历史和进展现状..................................
