首页> 外国专利> A CODEBOOK-BASED SPEECH ENHANCEMENT METHOD USING SPEECH ABSENCE PROBABILITY AND APPARATUS THEREOF

A CODEBOOK-BASED SPEECH ENHANCEMENT METHOD USING SPEECH ABSENCE PROBABILITY AND APPARATUS THEREOF

机译:利用语音缺席概率及其装置的基于码本的语音增强方法

摘要

PURPOSE: A codebook-based speech enhancement method using a speech presence probability and a device thereof are provided to be robust in a fast-changing noise environment by applying the speech presence probability to a codebook-based speech enhancement algorithm. CONSTITUTION: A voice recognition enhancement device converts an inputted voice signal from a time domain signal into a frequency domain signal (S11). The device calculates a binary estimation model for a speech absence probability (SAP) (S16). The device selects a noise codebook and clean voice codebook of the voice signal (S13,S15). The device calculates the gain of a noise without the presence of a clean voice and the gain of the noise with the presence of a clean voice through the SAP and determines a parameter set (S17). The device removes the noise by using Wiener filtering (S18). [Reference numerals] (AA) Input signal y = x + d; (BB) Input signal spectrum; (CC) Noise code book; (DD) Clean voice code book; (S12) Estimate noise; (S13) Select a noise code vector (IS-D); (S15) Select a voice code vector (IS-D); (S16) SAP modeling; (S17) Estimate a gain (_x^2, _d^2) and calculate an optimal parameter set
机译:目的:通过将语音存在概率应用于基于码本的语音增强算法,提供了一种使用语音存在概率的基于码本的语音增强方法及其设备,使其在快速变化的噪声环境中具有鲁棒性。组成:语音识别增强设备将输入的语音信号从时域信号转换为频域信号(S11)。设备为语音缺席概率(SAP)计算二进制估计模型(S16)。设备选择语音信号的噪声码本和纯净语音码本(S13,S15)。该设备通过SAP计算不存在纯净语音的噪声的增益和存在纯净语音的噪声的增益,并确定参数集(S17)。设备通过使用维纳滤波(S18)消除噪声。 [附图标记](AA)输入信号y = x + d; (BB)输入信号频谱; (CC)噪声代码本; (DD)干净的语音密码本; (S12)估计噪声; (S13)选择噪声代码向量(IS-D); (S15)选择语音编码向量(IS-D); (S16)SAP建模; (S17)估计增益(_x ^ 2,_d ^ 2)并计算最佳参数集

著录项

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号