• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2018.10a
• /
• pp.369-372
• /
• 2018

This paper suggests non-parallel-voice-conversion network conversing voice between unmapped voice pair as source voice and target voice. Conventional voice conversion researches used learning methods that minimize spectrogram's distance error. Not only these researches have some problem that is lost spectrogram resolution by methods averaging pixels. But also have used parallel data that is hard to collect. This research uses PPGs that is input voice's phonetic data and a GAN learning method to generate more clear voices. To evaluate the suggested method, we conduct MOS test with GMM based Model. We found that the performance is improved compared to the conventional methods.

• The Journal of the Acoustical Society of Korea
• /
• v.38 no.6
• /
• pp.687-696
• /
• 2019

Conventional feature of seismic classification focuses on strong seismic classification, while it is not suitable for classifying micro-seismic waves. We propose a feature extraction method based on histogram and Principal Component Analysis (PCA) in frequency-time space suitable for classifying seismic waves including strong, micro, and artificial seismic waves, as well as noise classification. The proposed method essentially employs histogram and PCA based features by concatenating the frequency and time information for binary classification which consist strong-micro-artificial/noise and micro/noise and micro/artificial seismic waves. Based on the recent earthquake data from 2017 to 2018, effectiveness of the proposed feature extraction method is demonstrated by comparing it with existing methods.

• The Journal of the Acoustical Society of Korea
• /
• v.34 no.4
• /
• pp.328-334
• /
• 2015

This paper describes an acoustic analysis of a Jing, Korean percussion instrument, according to different drive point and blow strength, and this analysis is focused on the softening and beat phenomena. Three kinds of blow strength (very strong, strong, and weak) and three locations of drive point (center, up, and right) are applied, and the spectrogram function built in Matlab is utilized to analyzing the softening and beat of target sounds. The stronger blow you drive to the center of the Jing, the more clearly softening is observed. Frequency shifting is increased proportionally to the blow strength and frequency and it is stand out on the harmonics in contrast with that of other partials. Beat of the Jing can be classified into the early beat and late beat. The beats by the outside driven Jing are distributed in wider frequency band than the beats by the center driven Jing. In addition, it is observed that the early beat is affected by few specific partials developed around harmonics for the center driven Jing.

• The Korean Journal of Applied Statistics
• /
• v.34 no.6
• /
• pp.979-993
• /
• 2021

Acoustic scene classification (ASC) categorizes an audio file based on the environment in which it has been recorded. This has long been studied in the detection and classification of acoustic scenes and events (DCASE). In this study, we considered the problem that ASC faces in real-world applications that the model used should have low-complexity. We compared several models that apply light-weight techniques. First, a base CNN model was proposed using log mel-spectrogram, deltas, and delta-deltas features. Second, depthwise separable convolution, linear bottleneck inverted residual block was applied to the convolutional layer, and Quantization was applied to the models to develop a low-complexity model. The model considering low-complexity was similar or slightly inferior to the performance of the base model, but the model size was significantly reduced from 503 KB to 42.76 KB.

• Journal of the Korea Convergence Society
• /
• v.10 no.9
• /
• pp.193-198
• /
• 2019

This study analyzed the changes of vowel formant, voice intensity, and fundamental frequency of vowels for 11 months using acoustochemical spectrogram analysis of women diagnosed with amyotrophic lateral sclerosis (ALS). The test word was a vowel /a, i, u/ and a diphthong /h + ja + da/, /h + wi + da/, and /h +ɰi+ da/. Speech data were collected through the word reading task presented on the monitor using 'Alvin' program, and the recording environment was set to 5,500 Hz for the nyquist frequency and 11,000 Hz for the sampling rate. The records were analyzed by using spectrograms to vowel formants, voice intensity, and fundamental frequency. As a result of analysis, the fundamental frequency and intensity of the ALS process were decreased and the formant slope of the diphthong was decreased rather than the formant change in the vowel. This result suggests that the vowel distortion of ALS due to disease progression is due to the decrease of tongue and jaw co morbidity.

• Journal of Broadcast Engineering
• /
• v.27 no.6
• /
• pp.885-896
• /
• 2022

Motor failure in manufacturing plays an important role in future A/S and reliability. Motor failure is detected by measuring sound, current, and vibration. For the data used in this paper, the sound of the car's side mirror motor gear box was used. Motor sound consists of three classes. Sound data is input to the network model through a conversion process through MelSpectrogram. In this paper, various methods were applied, such as data augmentation to improve the performance of classifying fault motors and various methods according to class imbalance were applied resampling, reweighting adjustment, change of loss function and representation learning and classification into two stages. In addition, the curriculum learning method and self-space learning method were compared through a total of five network models such as Bidirectional LSTM Attention, Convolutional Recurrent Neural Network, Multi-Head Attention, Bidirectional Temporal Convolution Network, and Convolution Neural Network, and the optimal configuration was found for motor sound classification.

• The Journal of the Acoustical Society of Korea
• /
• v.41 no.3
• /
• pp.312-318
• /
• 2022

This paper proposes a model which classifies the type of guns and information about sound source location using deep neural network. The proposed classification model is composed of convolutional neural networks (CNN) and long short-term memory (LSTM). For training and test the model, we use the Gunshot Audio Forensic Dataset generated by the project supported by the National Institute of Justice (NIJ). The acoustic signals are transformed to Mel-Spectrogram and they are provided as learning and test data for the proposed model. The model is compared with the control model consisting of convolutional neural networks only. The proposed model shows high accuracy more than 90 %.

• The Journal of the Acoustical Society of Korea
• /
• v.42 no.3
• /
• pp.214-220
• /
• 2023

In this paper, we propose a method to improve texture recognition performance from haptic acceleration signals representing the texture characteristics of object surface materials by using a Conformer model that combines the advantages of a convolutional neural network and a transformer. In the proposed method, three-axis acceleration signals generated by impact sound and vibration are combined into one-dimensional acceleration data while a person contacts the surface of the object materials using a tool such as a stylus , and the logarithmic Mel-spectrogram is extracted from the haptic acceleration signal similar to the audio signal. Then, Conformer is applied to the extracted the logarithmic Mel-spectrogram to learn main local and global frequency features in recognizing the texture of various object materials. Experiments on the Lehrstuhl für Medientechnik (LMT) haptic texture dataset consisting of 60 materials to evaluate the performance of the proposed model showed that the proposed method can effectively recognize the texture of the object surface material better than the existing methods.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.6
• /
• pp.1516-1529
• /
• 2023

In this paper, a motor gearbox fault-detection system based on a hierarchical flow-based model is proposed. The proposed system is used for the anomaly detection of a motion sound-based actuator module. The proposed flow-based model, which is a generative model, learns by directly modeling a data distribution function. As the objective function is the maximum likelihood value of the input data, the training is stable and simple to use for anomaly detection. The operation sound of a car's side-view mirror motor is converted into a Mel-spectrogram image, consisting of a folding signal and an unfolding signal, and used as training data in this experiment. The proposed system is composed of an encoder and a decoder. The data extracted from the layer of the pretrained feature extractor are used as the decoder input data in the encoder. This information is used in the decoder by performing an interlayer cross-scale convolution operation. The experimental results indicate that the context information of various dimensions extracted from the interlayer hierarchical data improves the defect detection accuracy. This paper is notable because it uses acoustic data and a normalizing flow model to detect outliers based on the features of experimental data.

• The Journal of the Convergence on Culture Technology
• /
• v.9 no.2
• /
• pp.545-551
• /
• 2023

As digital information is readily available to everyone today, the adoption of digital evidence is increasing. However, it is virtually impossible to determine the authenticity of forgery in the case of a voice recording file that has gone through a sophisticated editing process along with the spread of various voice file editing tools. This study aims to prove that forgery, which is difficult to distinguish from the original file, is possible by using insertion, deletion, linking, and synthetic editing technologies in voice recording files. This study presents the difficulty of detecting forgery by encoding a forged voice file with the same extension as the original. In addition, it was shown that forgery detection is impossible if additional transition band deletion and secondary encoding are performed only for experiments in which features occurred. Through this, this study is expected to contribute to the establishment of more stringent evidence admissibility criteria for adopting voice recording files as digital evidence.