• The Journal of the Korea institute of electronic communication sciences
• /
• v.14 no.5
• /
• pp.811-816
• /
• 2019

This paper first proposes a technique to separate clean speech signals and mixed noise signals by using an independent vector analysis algorithm of frequency band for 4 channel speech source signals with a noise. An improved output speech signal from the proposed independent vector analysis algorithm is obtained by using the cross-correlation between the signal outputs from the frequency domain delay-sum beamforming and the output signals separated from the proposed independent vector analysis algorithm. In the experiments, the proposed algorithm improves the maximum SNRs of 10.90dB and the segmental SNRs of 10.02dB compared with the frequency domain delay-sum beamforming algorithm for the input mixed noise speeches with 0dB and -5dB SNRs including white noise, respectively. Therefore, it can be seen from this experiment and consideration that the speech quality of this proposed algorithm is improved compared to the frequency domain delay-sum beamforming algorithm.

• The Journal of the Acoustical Society of Korea
• /
• v.38 no.1
• /
• pp.39-46
• /
• 2019

In this paper, combining features is proposed as a way to enhance the classification accuracy of sounds under noisy environments using the CNN (Convolutional Neural Network) structure. A robust log Mel-filter bank using Wiener filter and PNCCs (Power Normalized Cepstral Coefficients) are extracted to form a 2-dimensional feature that is used as input to the CNN structure. An ebird database is used to classify 43 types of bird species in their natural environment. To evaluate the performance of the combined features under noisy environments, the database is augmented with 3 types of noise under 4 different SNRs (Signal to Noise Ratios) (20 dB, 10 dB, 5 dB, 0 dB). The combined feature is compared to the log Mel-filter bank with and without incorporating the Wiener filter and the PNCCs. The combined feature is shown to outperform the other mentioned features under clean environments with a 1.34 % increase in overall average accuracy. Additionally, the accuracy under noisy environments at the 4 SNR levels is increased by 1.06 % and 0.65 % for shop and schoolyard noise backgrounds, respectively.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.11 no.1
• /
• pp.64-72
• /
• 1986

The error rate equation fo DS-BPSK(Direct Sequence Binary Phase Shift Keying) signal transmitted through the nonlinear satellite transponder has been derived in the cochannel interference and downlink Gaussian noise environment. The input to the satellite transponder is the superposition of DS-BPSK signal with one interfere which is a cochannel wide-band PN signal. The error rate performance of DS-BPSK system has been evaluated and shown in figures in terms of carrier to interference power ratio(CIR), downlink signal to noise power ratio(downlink SNR) and process gain. In the analysis, it has been shown that the use of a hard limiter in DS-BPSK satellite system leads to the generation of narrow-band intermodulation products which is independent of the process gain. Also it is known that the error rate performance can be improved in the low levels (below 10dB) of CIR as the CIR increase. As the process gain varies from 10 to 100 the curve gives the about 10 dB gain in downlink SNR to maintain a fixed error rate.

• Journal of the Korea society of information convergence
• /
• v.6 no.2
• /
• pp.45-49
• /
• 2013

Electrocardiogram(ECG) signal noise as well as conducting other bio-signal measurement were generated. It was intened to enhance the accuracy of cadiac disease diagnosis with removing signal white-noise. Sampling signal was made with generating white-noise. The noise were removed using wavelet transforms and thresholding. Removed noise were compared numerical using SNR(signal to noise ratio). The results compared SNR showed that SURE method was 5.931, 4.9301 in 3, 5dB noise, uninversal was 3.6590, 1.9698 in 7, 9dB noise. De-noising by Thresholding removed noise effectively. ECG signal is expected to improve the accuracy of cadiac desease dianosis.

• Proceedings of the IEEK Conference
• /
• 2006.06a
• /
• pp.507-508
• /
• 2006

High performance delta-sigma modulator is developed for audio-codec applications(i.e.. 16-bit resolution at a 20kHz signal bandwidth). The modulator is realized with fully-differential switched capacitor integrators. All stages employ a single-stage folded-cascode amplifier. The presented delta-sigma modulator when clocked at 3.2MHz achieves 85.2dB peak-SNDR and 94.8dB SNR. This modulator is designed in a SAMSUNG $0.18{\mu}m$ CMOS process. Finally, this paper shows the test setup and FFT result gained from delta-sigma modulator chip designed for audio applications.

• Journal of the Institute of Convergence Signal Processing
• /
• v.14 no.4
• /
• pp.249-254
• /
• 2013

A fast multiplier and ADC are essential to process the analog signals in real time. The double-base number system(DBNS) is known as an efficient method for this purpose. The DBNS uses the numbers 2 and 3 as the base numbers simultaneously. The system has an advantage of fast multiplication, less chip area, and low power consumption compared to the binary multiplier. However, the inherent errors of the log number's intrinsic tolerance in DBNS are accumulated in a FIR digital filter, so the signal-to-noise ratio(SNR) has a tendency to be degraded. In this paper, the nonlinear encoder of ADC is designed to compensate the accumulated errors of DBNS by analysing the error distributions of various filter coefficients. The new ADC does not sacrifice its own advantages because the encoder circuits are modified only. The experiments were done with an FIR filters those were designed to have -70dB of SNR in stop band. The proposed nonlinear ADC encoder could drop the SNR to -45dB in stop band, in contrast to -35dB with the linear encoder.

• The Journal of the Acoustical Society of Korea
• /
• v.38 no.6
• /
• pp.697-702
• /
• 2019

We propose a multi-band multi-scale DenseNet with dilated convolution that separates background music signals from broadcast content. Dilated convolution can learn the multi-scale context information represented by spectrogram. In computer simulation experiments, the proposed architecture is shown to improve Signal to Distortion Ratio (SDR) by 0.15 dB and 0.27 dB in 0dB and -10 dB Signal to Noise Ratio (SNR) environments, respectively.

• Journal of Digital Convergence
• /
• v.12 no.10
• /
• pp.265-270
• /
• 2014

Existing speech recognition algorithm have a problem with not distinguish the order of vocabulary, and the voice detection is not the accurate of noise in accordance with recognized environmental changes, and retrieval system, mismatches to user's request are problems because of the various meanings of keywords. In this article, we proposed to event based semantic ontology inference model, and proposed system have a model to extract the speech recognition feature extract using ERB filter. The proposed model was used to evaluate the performance of the train station, train noise. Noise environment of the SNR-10dB, -5dB in the signal was performed to remove the noise. Distortion measure results confirmed the improved performance of 2.17dB, 1.31dB.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.13 no.5
• /
• pp.400-410
• /
• 1988

In this paper, a fast synchronization system using SAW TDL matched filter and SAW recirculation loop not only for acquisition but also tracking in the direct sequence spread spectrum communication receiver in case of low SNR was designed and its characteristics were investigated. When signal of 16dB SNR was inputed at the receiver, the PN code of the receiver could be synchronized from the extracted signal for synchronization through SAW TDL matched filter and SAM recirculation loop for 30 recirculations. And the average synchronization time of this system was calculated. From the results, we found that this synchronization system could achieve faster synchronization of PN codes in the receiver under the circumstances of low SNR than that of using only matched filter.

• Journal of the Korean Institute of Telematics and Electronics D
• /
• v.36D no.3
• /
• pp.1-11
• /
• 1999

It is now common practice to utilize the quadrature RF coils to improve the signal-to-noise ratio (SNR) in the Magnetic Resonance Imaging (MRI) System. In addition, to make such an available SNR improvement, it is mandatory to use a well-designed quadrature coupler, which facilitates a perfect 3-dB coupling and quadrature-phase shift. However, the four ports matching condition has to be well considered during the RF excitation and the signal detection period. This work investigates the effects of such a mismatching condition (especially, due to patient) from the analysis, simulation, and real implementation and firstly proposes dynamic loading technique for a quadrature coupler and transmitting/receiving switch module to minimize a patient mismatching and enhance a system reliability. Also, we designed and implemented the quadrature coupler and transmitting/receiving switch module using microstrip. As a result, the SNR of our MRI system using the microstrip quadrature coupler and transmitting/receiving switch module with dynamic load increases 3 dB compared with the old one using USA quadrature switch. Also, the power capability of quadrature coupler and transmitting/receiving switch module is 5-kw peak power. Considering power loss and reduction of size, we used a RT/duroid 6010 substrate with high permittivity and for simulation we use Compact Software.