ID 原文 译文
39216 在解码端，在接收到其中任一语音码流时，用G.722.1解码器进行解码，其语音质量不低于G.722.1编码器的解码结果，而在接收到两个语音码流时，用G.722.1解码器先分别对两个语音码流进行解码，然后对解码结果进行联合处理，其最终的语音质量有明显提升，即有一定编码增益。 When the decoder receives one of the encoding results, it is decoded by the G.722.1 decoder, the speech quality is higher than the decoding result of the G.722.1 encoder.At the same time, when both encoding results are received, the G.722.1 decoder is used to decode the two voice code streams respectively, and then the decoding results are weighted, the final voice quality is significantly improved.In other words, there is a certain coding gain.
39217 仿真实验结果表明，本文分布式语音编码方法的抗丢包效果明显，相对于原始编解码器其语音质量进一步提升。 The simulation experiment results show that the distributed speech coding algorithm proposed in this paper can improve the encoding quality significantly compared to the traditional coding techniques.
39218 针对现有深度神经网络语音增强方法对带噪语音的去噪能力有限、语音质量提升不高的问题，提出了一种基于奇异谱分析的深度神经网络语音增强方法。 In order to solve the problems of limited denoising ability and low speech quality improvement of existing deep neural network speech enhancement methods for noisy speech, a deep neural network speech enhancement method based on singular spectrum analysis is proposed.
39219 通过引入奇异谱分析算法对带噪语音进行预处理，以初步分离得到语音信号与噪声。 By introducing the singular spectrum analysis algorithm to preprocess the noisy speech, the speech signal and noise are preliminarily separated.
39220 接着将语音信号与噪声用于深度神经网络模型的训练，以得到性能更优的网络模型，从而使得本文方法具有更好的性能。 Then the speech signal and noise are used to train the depth neural network model to obtain a network model with better performance, so that the new method of deep neural network speech enhancement based on singular spectrum analysis has better performance.
39221 最后在重建干净语音的环节中，同时使用神经网络估计得到的对数功率谱和带噪语音的对数功率谱，并加入了权重系数，使得本文提出的方法可以适应不同信噪比的情形，有效的去除背景噪声，降低语音信号的失真。 Finally, both the logarithmic power spectrum estimated by the neural network and the logarithmic power spectrum of noisy speech are used to reconstruct clean speech. The method proposed in this paper can adapt to the situation of different signal-to-noise ratio, effectively remove the background noise and reduce the distortion of speech signal.
39222 本文通过仿真实验验证了该方法的有效性和鲁棒性。 In this paper, simulation experiments are carried out to verify the effectiveness and robustness of the method.
39223 提出了一种基于区域约束的双耳近场自适应波束形成算法，该算法将目标信号及其附近区域内的若干个导向矢量组成的矩阵进行特征值分解，并在最小化输出信号能量的同时对主要的特征向量进行约束，从而有效解决传统近场波束形成器对目标声源的方位估计误差和位置扰动过于敏感的问题。 To solve the problem that the traditional binaural near-field beamformer is too sensitive to the azimuth estimation error and the position disturbance of the desired sound source, a region-constrained binaural near-field adaptive beamforming algorithm is proposed, which applies the singular value decomposition method on a matrix composed of the steering vectors corresponding to the desired signal and several discrete points in its nearby area, and then obtains the enhanced speech signal by minimizing the output signal power subject to some constraints on these main eigenvectors.
39224 实验结果表明，所提方法可以有效抑制远场同向干扰声源，并且相对于传统近场波束形成器具有更高的鲁棒性。 Experimental results show that the proposed algorithm can suppress far-filed co-directional interference effectively and is more robust than the traditional binaural near-field beamforming algorithms.
39225 客观评价指标表明，该算法的降噪性能及语音质量提升方面均优于对比算法。 Additionally, objective evaluation results show that the noise reduction performance and speech quality improvement of the proposed algorithm are superior to the competing algorithms.