• The Journal of the Acoustical Society of Korea
• /
• v.43 no.1
• /
• pp.9-18
• /
• 2024

The importance of active sonar systems is emerging due to the quietness of underwater targets and the increase in ambient noise due to the increase in maritime traffic. However, the low signal-to-noise ratio of the echo signal due to multipath propagation of the signal, various clutter, ambient noise and reverberation makes it difficult to identify underwater targets using active sonar. Attempts have been made to apply data-based methods such as machine learning or deep learning to improve the performance of underwater target recognition systems, but it is difficult to collect enough data for training due to the nature of sonar datasets. Methods based on mathematical modeling have been mainly used to compensate for insufficient active sonar data. However, methodologies based on mathematical modeling have limitations in accurately simulating complex underwater phenomena. Therefore, in this paper, we propose a sonar signal synthesis method based on a deep neural network. In order to apply the neural network model to the field of sonar signal synthesis, the proposed method appropriately corrects the attention-based encoder and decoder to the sonar signal, which is the main module of the Tacotron model mainly used in the field of speech synthesis. It is possible to synthesize a signal more similar to the actual signal by training the proposed model using the dataset collected by arranging a simulated target in an actual marine environment. In order to verify the performance of the proposed method, Perceptual evaluation of audio quality test was conducted and within score difference -2.3 was shown compared to actual signal in a total of four different environments. These results prove that the active sonar signal generated by the proposed method approximates the actual signal.

• The Journal of the Acoustical Society of Korea
• /
• v.24 no.2
• /
• pp.68-77
• /
• 2005

The purpose of a spatial audio system is to give a listener an impression as if he were present in a recorded environment when its sound is reproduced. For this purpose a dummy head microphone is generally used. Because of its human-like shape, dummy head microphone can reproduce spatial images through headphone reproduction. However, its shape and size are restriction to public use and it is difficult to convert the output signal of dummy head microphone into a multi-channel signal for multi-channel environment. So, in this paper, we propose a multi-channel 3D microphone technology. The multi-channel 3D microphone acquire a spatial audio using five microphones around a horizontal plane of a rigid sphere and through post processing, it can reproduce various reproduction signals for headphone, stereo, stereo dipole, 4ch and 5ch reproduction environments. Because of complex computation, we implemented H/W based post processing system. To verily the Performance of the multi-channel 3D microphone, localization experiments were Performed. The result shows that a front/back confusion, which is the one of common limitations of conventional dummy head technology, can be reduced dramatically.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2014.06a
• /
• pp.157-159
• /
• 2014

In this paper, we introduce SlimSRP, an ultra-low-power digital signal processor (DSP) solution for mobile audio and voice applications. So far, application processors (APs) have taken charge of all the tasks in mobile devices. However, they have suffered from short battery life problems to deal with complex usage scenarios, such as always-on voice trigger with continuous audio playback. From extensive analysis of audio and voice application characteristics, SlimSRP is designed to relive the performance and power burden of APs. It employs three-issue VLIW architecture, and the major low-power and high-performance techniques include: (1) an optimized register-file architecture friendly for constants generation, (2) a powerful instruction set to reduce the number of register file accesses and (3) a unique instruction compression scheme that contributes to saved memory size and reduced cache miss. An implementation of SlimSRP runs at up to 200MHz and the logic occupies 95K NAND2 gates in Samsung 28LPP process. The experimental results demonstrate that a MP3 decoder application with a 128kbps 44.1kHz input can run at 5.1MHz and the logic consumes only 22uW/MHz.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.37 no.8A
• /
• pp.712-719
• /
• 2012

This paper presents Single-bit Third order Delta-Sigma Modulator, which can be applied to the Low speed High resolution ADC in Audio signal Process System. Whereas the Operational Amplifier in modulator consumed static power dissipation in operating, this modulator used Switching on/off techniques, which makes the Power dissipation of the modulator reduced. Also proposed modulator minimizes frequency characteristic variation by optimizing switch position. And this modulator chooses Single-bit type to guarantee stability. The designed ADC went through 0.35um CMOS n-well 1-poly 4-metal process to be a final product, and the final product has shown 17.1mW of power dissipation with 3.3V of Supply Voltage, 6.4MHz of conversion rate. And 84.3dB SNDR and 13.5bit ENOB with 20KHz of input frequency.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.38 no.7
• /
• pp.455-463
• /
• 2001

In this Paper, an approximate processing method is proposed and tested. The proposed method uses variable CSD (VCSD) coefficients which approximate filter stopband attenuation by controlling the precision of the CSD coefficient sets. A decimation filter for Audio Codec '97 specifications has been designed having processor architecture that consists of program/data memory, arithmetic unit, energy/level decision, and sinc filter blocks, and fabricated with 0.6${\mu}{\textrm}{m}$ CMOS sea-of-gate technology. For the combined two halfband FIR filters in decimation filter, the number of addition operations were reduced to 63.5%, 35.7%, and 13.9%, compared to worst-case which is not an adaptive one. Experimental results show that the total power reduction rate of the filter is varying from 3.8 % to 9.0 % with respect to worst-case. The proposed approximate processing method using variable CSD coefficients is readily applicable to various kinds of filters and suitable, especially, for the speech and audio applications, like oversampling ADCs and DACs, filter banks, voice/audio codecs, etc.

• The Journal of the Acoustical Society of Korea
• /
• v.29 no.7
• /
• pp.458-467
• /
• 2010

In this paper, a new audio reproduction system was developed in which the cross-talk signals would be reasonably cancelled at an arbitrary listener position. To adaptively remove the cross-talk signals according to the listener's position, a method of tracking the listener position was employed. This was achieved using the two microphones, where the listener direction was estimated using the time-delay between the two signals from the two microphones, respectively. Moreover, room reverberation effects were taken into consideration where linear prediction analysis was involved. To remove the cross-talk signals at the left-and right-ears, the paths between the sources and the ears were represented using the KEMAR head-related transfer functions (HRTFs) which were measured from the artificial dummy head. To evaluate the usefulness of the proposed listener tracking system, the performance of cross-talk cancellation was evaluated at the estimated listener positions. The performance was evaluated in terms of the channel separation ration (CSR), a -10 dB of CSR was experimentally achieved although the listener positions were more or less deviated. A real-time system was implemented using a floating-point digital signal processor (DSP). It was confirmed that the average errors of the listener direction was 5 degree and the subjects indicated that 80 % of the stimuli was perceived as the correct directions.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.42 no.4 s.304
• /
• pp.167-176
• /
• 2005

This paper proposes a new power minimization technique for the frame-based multimedia signal processing. The derivation of the technique is based on the newly proposed microscopic DVS(Dynamic Voltage Scaling) method, where, the operating frequency and the supply voltage levels are dynamically controlled according to the processing requirement for each frame of multimedia data. The multimedia signal processing algorithms are also redesigned and optimized to maximize the power saving efficiency of the microscopic DVS technology. The characterization of the mean/variance distribution of the processing load in the frame-based multimedia signal processing provides the major basis not only for the optimized application of the microscopic DVS technology but also for the optimization of the multimedia algorithms. The power saying efficiency of the proposed DVS approach is experimentally tested with the algorithms of MPEG-2 video decoder and MPEG-2 AAC audio encoder on the ARM9 RISC processor. The experimental results with the diverse MPEG-2 video and audio files show The average power saving efficiencies of 50$\%$ and 30$\%$, respectively. The results also agree very well with those of the analytic derivations.

• The Journal of the Acoustical Society of Korea
• /
• v.42 no.4
• /
• pp.329-344
• /
• 2023

In this study, we comprehensively analyze the generalization performance of various deep learning-based active sonar target classifiers when applied to small and imbalanced active sonar datasets. To generate the active sonar datasets, we use data from two different oceanic experiments conducted at different times and ocean. Each sample in the active sonar datasets is a time-frequency domain image, which is extracted from audio signal of contact after the detection process. For the comprehensive analysis, we utilize 22 Convolutional Neural Networks (CNN) models. Two datasets are used as train/validation datasets and test datasets, alternatively. To calculate the variance in the output of the target classifiers, the train/validation/test datasets are repeated 10 times. Hyperparameters for training are optimized using Bayesian optimization. The results demonstrate that shallow CNN models show superior robustness and generalization performance compared to most of deep CNN models. The results from this paper can serve as a valuable reference for future research directions in deep learning-based active sonar target classification.