基于cs理论的语音增强算法的研究-电路与系统专业论文.docxVIP

基于 基于 CS 理论的语音增强算法的研究 西北师 西北师范大学硕士学位论文 PAGE PAGE IV III III Abstract Speech signal is a kind of non-stationary and time-varying signal. Voice is the most important and most common method for information exchange of human being. Speech signal processing, which mainly includes speech coding, speech synthesis, speech recognition and speech enhancement, has a very wide range of application. Usually, the researchers conduct various speech signals processing in relatively pure conditions. But speech signal always contaminated by various noises in real life environment. The noise will seriously affect the quality and the intelligibility of speech signal. This thesis reviews the advantages and disadvantages of speech enhancement algorithms by focusing on the classical spectral subtraction algorithm and sub-space based speech enhancement algorithm. A novel speech enhancement algorithm is proposed by introducing the compressed sensing (CS) theory. The main works and novelties are as follows: Firstly, this thesis proposes a novel speech endpoint detection method by calculating the quotient of short-time zero-crossing and short-time energy of speech signal. A threshold is obtained according to the quotient of the start segment of speech signal. Then each frame of speech signal is classified into non-speech segment or speech segment. The speech segments are used for speech enhancement. We conduct an endpoint detection experiment on the pure speech signal and the noisy speech signal contaminated by train noise with 0dB of Signal-to-Noise Rate (SNR). The short-time energy, the short-time zero-crossing, the short-time entropy of spectrum, the product of short-time energy and short-time zero-crossing and the quotient of short-time zero-crossing and short-time energy. Experimental results demonstrate that the quotient of short-time zero-crossing and short-time energy is more robust for endpoint detection on noisy speech. Secondly, the thesis proposes a novel approach for speech e