This paper presents a deep neural network(DNN)-based speech enhancement algorithm based on the soft audible noise masking for the single-channel wind noise reduction. To reduce the low-frequency residual noise, the ps...This paper presents a deep neural network(DNN)-based speech enhancement algorithm based on the soft audible noise masking for the single-channel wind noise reduction. To reduce the low-frequency residual noise, the psychoacoustic model is adopted to calculate the masking threshold from the estimated clean speech spectrum. The gain for noise suppression is obtained based on soft audible noise masking by comparing the estimated wind noise spectrum with the masking threshold. To deal with the abruptly time-varying noisy signals, two separate DNN models are utilized to estimate the spectra of clean speech and wind noise components. Experimental results on the subjective and objective quality tests show that the proposed algorithm achieves the better performance compared with the conventional DNN-based wind noise reduction method.展开更多
High purity （99.999% or 5N, mass fraction） indium （In） was obtained through vacuum distillation using a 2N （99%） In as input material under a dynamic vacuum of 5 Pa. The glow discharge mass spectrometry （GDMS）...High purity （99.999% or 5N, mass fraction） indium （In） was obtained through vacuum distillation using a 2N （99%） In as input material under a dynamic vacuum of 5 Pa. The glow discharge mass spectrometry （GDMS） was applied for the analysis of input material and the distilled indium. The results indicate that high-volatile impurities namely Cd, Zn, T1 and Pb can be removed from the indium matrix at the low fraction stage of 1 223 K for 120 min; Low-volatile impurities such as Fe, Ni, Cu, Sn can be reduced at the high fraction stage of 1 323 K for 120 min. The separation coefficient ,8i and activity coefficient Yi of impurities are calculated according to the experiments to fill the inadequate data of the thermodynamics.展开更多
基金partially supported by the National Natural Science Foundation of China (Nos.11590772, 11590770)the Pre-research Project for Equipment of General Information System (No.JZX2017-0994/Y306)
文摘This paper presents a deep neural network(DNN)-based speech enhancement algorithm based on the soft audible noise masking for the single-channel wind noise reduction. To reduce the low-frequency residual noise, the psychoacoustic model is adopted to calculate the masking threshold from the estimated clean speech spectrum. The gain for noise suppression is obtained based on soft audible noise masking by comparing the estimated wind noise spectrum with the masking threshold. To deal with the abruptly time-varying noisy signals, two separate DNN models are utilized to estimate the spectra of clean speech and wind noise components. Experimental results on the subjective and objective quality tests show that the proposed algorithm achieves the better performance compared with the conventional DNN-based wind noise reduction method.
基金Project(2009AA003) supported by Science and Technology Innovation Plan of Yunnan Province, China
文摘High purity （99.999% or 5N, mass fraction） indium （In） was obtained through vacuum distillation using a 2N （99%） In as input material under a dynamic vacuum of 5 Pa. The glow discharge mass spectrometry （GDMS） was applied for the analysis of input material and the distilled indium. The results indicate that high-volatile impurities namely Cd, Zn, T1 and Pb can be removed from the indium matrix at the low fraction stage of 1 223 K for 120 min; Low-volatile impurities such as Fe, Ni, Cu, Sn can be reduced at the high fraction stage of 1 323 K for 120 min. The separation coefficient ,8i and activity coefficient Yi of impurities are calculated according to the experiments to fill the inadequate data of the thermodynamics.
