ID 原文 译文
39236 为了处理脉冲噪声，传统的滤波x最大相关熵准则(filtered-x maximum correntropy criterion，FxMCC)自适应算法具有良好的降噪性能在主动噪声控制(active noise control，ANC)系统中。 In order to deal with impulsive noise, the traditional filtered-x maximum correntropy criterion(FxMCC) adaptive algorithm has good noise reduction in active noise control(ANC) systems.
39237 然而，FxMCC算法具有单一的高斯核，其降噪性能和收敛速度易受到核宽度取值的影响。 However, the FxMCC algorithm has a single Gaussian kernel, and its noise reduction performance and convergence speed are easily affected by the kernel width.
39238 因此，在ANC系统中，本文采用最大混合相关熵准则(maximum mixture correntropy criterion，MMCC)，以两个高斯核的混合作为代价函数来推导滤波x最大混合相关熵准则(filtered-x maximum mixture correntropy criterion，FxMMCC)算法。 Therefore, in the ANC system, this paper adopts the maximum mixture correntropy criterion(MMCC), and uses the mixture of two Gaussian kernels as the cost function to derive the filtered-x maximum mixture correntropy criterion(FxMMCC) algorithm.
39239 仿真结果表明，在标准对称α稳定(symmetricα-stable， SαS)分布的脉冲噪声的主动噪声控制中，本文提出的FxMMCC算法比其他ANC算法具有更好的降噪性能。 Simulation results show that the proposed FxMMCC algorithm in this paper has better noise reduction performance than other ANC algorithms in active noise control of impulsive noise with standard symmetric α-stable(SαS) distribution.
39240 在现代通信系统中，回波与混响常损害通信语音的质量和可懂度。 In modern telecommunications, both echo and reverberation can significantly disturb people's communication and degrade the speech intelligibility and quality.
39241 为克服回波与混响的负面影响，本文提出了一种基于深度学习的两阶段联合声学回波和混响抑制系统。 In order to overcome the negative impact of the echo and reverberation simultaneously, we proposed a two-stage joint-training system based on deep learning to enhance the speech signal, where echo cancellation and speech dereverberation were conducted sequentially.
39242 系统先用基于理想比值掩蔽的模型去除与目标信号不相关的声学回波; The system is composed of two stages, echo cancellation stage and dereverberation stage.
39243 然后用一个基于“隐掩蔽”的谱映射模型去除与目标信号强相关的混响干扰; The system firstly employed a model based on ideal ratio mask(IRM) to cancel the acoustic echo, which was uncorrelated with the target signal.
39244 最后联合训练两阶段模型以获得更好的系统性能。 Then the reverberation strongly correlated with the target signal was removed using a spectrum mapping model combined with a hidden mask. Then the two-stage model was jointly trained to obtain a better performance.
39245 一系列不同声学环境下的实验结果表明，本文所提出的系统可显著地消除回波与混响干扰，从而极大地增强了目标语音的语音质量与可懂度。 A series of systematic experiments were conducted in different conditions and the results indicated that the proposed system significantly improves the performance on echo cancellation and dereverberation and achieves better speech intelligibility and quality over other methods.