The problems of cosine phase unwrapping method are: 1. the residual noise in the phase cosine function after phase-shifting filter can be eliminated greatly, but the position of main wave crests, main wave troughs and zero points need judged correctly;

The residual noise and the spectrum distortion after enhancement are adverse factors for speech recognition, and their effects will be compensated by Parallel Model Combination (PMC) in recognition stage or by Cepstral Mean Normalization (CMN) in feature extraction stage.

When the PMT is simply replaced by PDA,a difference between lamp spectrum and atmosphere spectrum,as will as unwanted standard deviation of residual noise as much as 1.4×10-3 appears after corrected for the pixel response. A fiber mode mixer,based on fiber disturb mode principle was developed and tested. This new device reduces the standard deviation to 3.4×10-4 in field experiments.

H owever, the signal still contains residual noise as well as speech distortion wh ich were compensated for by Parallel Model Combination in the recognition stage or by Cepstral Mean Normalization in the feature extraction stage.

The third part is to suppress the background clutter and enhance the target by the methods of frame subtraction and background prediction on the basis of the analysis of the spatial and temporal property of the target and the background, to eliminate the remanent Gaussian noises using higher-order cumulant, and to accomplish the detection and recognition of pulse spectral signal.

Based on examining of the noise pattern, two kinds of adaptive median filter algorithms were proposed in this paper. (1)To remove impulse noise that have high occurrence probability, we have adaptive median filter algorithm based on detecting of residual impulses (note as AMF1), through detecting the presence of residual impulse in the median filter output to decide the window size of filter, therefore effectively remove impulse noise;

The use of photodiode array(PDA)detectors enhances DOAS system considerably owing to PDA's higher sensitivity and advantage in multi-channel application.

The standard error of the residual noise at this wavelength is about 1 mJy.

The simulation results show that the proposed algorithm yields less residual noise and better speech quality than the Gaussian based speech enhancement algorithms proposed in recent years.

With this algorithm, the image can be decomposed into constituents of different dimensionality, i.e., point-like, thread-like, and area-like structures, and residual noise.

The measurement precision and the source noise level of data at 21h in this region were obtained in the analysis of residual noise.

We use mixed model results to obtain estimates of unit-specific random effects, state disturbance terms and residual noise terms.

The basic principle and method of measuring the antenna excess noise temperature Tas by using the radio stars as calibration are presented in this paper. In this method, the antenna noise temperature TAN is used as a standard of calibration when the antenna points at sky. It was found that the results of measurement using this method agree with that obtained by using liquid nitrogen at low and normal ambient temperature as standard of calibration. The method has advantage of more simplicity of the measurement...

The basic principle and method of measuring the antenna excess noise temperature Tas by using the radio stars as calibration are presented in this paper. In this method, the antenna noise temperature TAN is used as a standard of calibration when the antenna points at sky. It was found that the results of measurement using this method agree with that obtained by using liquid nitrogen at low and normal ambient temperature as standard of calibration. The method has advantage of more simplicity of the measurement and equipment. The total error of antenna excess noise temperature obtained by this method could be estimated to be ±3%. The error of antenna gain measured is about±0.3 dB.

This paper proposes a linearization conception of the statistical quantization characteristic of roundoff quantizer. According to this conception the method of limit cycle suppression in the first-and second-order digital filters is found. All the limit cycles in digital filters can be suppressed by the use of this method and no remaining noise exists in the filter output. The calculation method of finding the limit cycle suppression time is proposed also and the calculation results are consistent with those...

This paper proposes a linearization conception of the statistical quantization characteristic of roundoff quantizer. According to this conception the method of limit cycle suppression in the first-and second-order digital filters is found. All the limit cycles in digital filters can be suppressed by the use of this method and no remaining noise exists in the filter output. The calculation method of finding the limit cycle suppression time is proposed also and the calculation results are consistent with those obtained from experiments.

In this paper, we study three types of noisy speech enhancement algorithms based on Minimum Mean-Square Error Short-Time Spectral Estimation: Spectral Amplitude Estimation (MMSE-STSA), LOG-Spectral Amplitude Estimation (LOG-MMSE-STSA). and Relative Spectral Amplitude Estimation (MMSE-REL-STSA). On the basis of theoretical analysis, experimental studies are conducted. Computer simulations show that SNR improvement of 3.4-12dB can be obtained after enhancement processing when the original SNR of noisy speech degraded...

In this paper, we study three types of noisy speech enhancement algorithms based on Minimum Mean-Square Error Short-Time Spectral Estimation: Spectral Amplitude Estimation (MMSE-STSA), LOG-Spectral Amplitude Estimation (LOG-MMSE-STSA). and Relative Spectral Amplitude Estimation (MMSE-REL-STSA). On the basis of theoretical analysis, experimental studies are conducted. Computer simulations show that SNR improvement of 3.4-12dB can be obtained after enhancement processing when the original SNR of noisy speech degraded by Additional White Gaussian Noise ranges from 5dB to-10dB. Among the three algorithms, MMSE-LOG-STSA provides the best SNR improvement, and the SNR results coincide with informal listening.We also study the reduction of residual noise accompanied by the enhanced speech. By using the center clipping and gain function modification of estimator, the residual noise can be reduced significantly.