Speech enhancement is fundamental for various real time speech applications and it is a challenging task in the case of a single channel because practically only one data channel is available. We have proposed a supervised single channel speech enhancement algorithm in this paper based on a deep neural network (DNN) and less aggressive Wiener filtering as additional DNN layer. During the training stage the network learns and predicts the magnitude spectrums of the clean and noise signals from input noisy speech acoustic features. Relative spectral transform-perceptual linear prediction (RASTA-PLP) is used in the proposed method to extract the acoustic features at the frame level. Autoregressive moving average (ARMA) filter is applied to smooth the temporal curves of extracted features. The trained network predicts the coefficients to construct a ratio mask based on mean square error (MSE) objective cost function. The less aggressive Wiener filter is placed as an additional layer on the top of a DNN to produce an enhanced magnitude spectrum. Finally, the noisy speech phase is used to reconstruct the enhanced speech. The experimental results demonstrate that the proposed DNN framework with less aggressive Wiener filtering outperforms the competing speech enhancement methods in terms of the speech quality and intelligibility.
Gabor Wigner Transform (GWT) is a composition of two time-frequency planes (Gabor Transform (GT) and Wigner Distribution (WD)), and hence GWT takes the advantages of both transforms (high resolution of WD and cross-terms free GT). In multi-component signal analysis where GWT fails to extract auto-components, the marriage of signal processing and image processing techniques proved their potential to extract autocomponents. The proposed algorithm maintained the resolution of auto-components. This work also shows that the Fractional Fourier Transform (FRFT) domain is a powerful tool for signal analysis. Performance analysis of modified fractional GWT reveals that it provides a solution of cross-terms of WD and blurring of GT.
The current study was aimed to evaluate the industrial efﬂ uents biodegradation potential of an indigenous microorganism which reduced water pollution caused by these efﬂ uents. In the present study biodegradation of three textile industrial efﬂ uents was performed with locally isolated brown rot fungi named Coniophora puteana IEBL-1. Response Surface Methodology (RSM) was employed under Box Bhenken Design (BBD) for the optimization of physical and nutritional parameters for maximum biodegradation. Quality of treated efﬂ uents was checked by study of BOD, COD and analysis through HPLC. Three ligninolytic enzymes named lignin peroxidase, manganese peroxidase and laccase were also studied during the biodegradation process. The results showed that there was more than 85% biodegradation achieved for all three efﬂ uents with decrease in Biological Oxygen Demand (BOD) and Chemical Oxygen Demand (COD) below the recommended values for industrial efﬂ uent i.e. 80 mg/L for BOD and 220 mg/L for COD after optimization of nutritional parameters in the second stage. Analysis of samples through HPLC revealed the formation of less toxic diphenylamine, 3-methyldiphenylamine and N-methylaniline after treatment. The ligninolytic enzymes assays conﬁ rmed the role of lignin peroxidase (LiP), manganese peroxidase (MnP) and laccase in biodegradation process. Lignin peroxidase with higher activity has more contribution in biodegradation of efﬂ uents under study. It can be concluded through the results that Coniophora buteana IEBL-1 is a potential fungus for the treatment of industrial efﬂuents.