• Journal of KIISE:Computing Practices and Letters
• /
• v.15 no.12
• /
• pp.993-997
• /
• 2009

Clustering is well known method and that can be used in statistical machine translation. In this paper we propose a corpus clustering method using syntactic structure and POS information of dependency grammar. And using this cluster language model as additional feature to phrased-based statistical machine translation system to improve translation Quality.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1998.06e
• /
• pp.269-272
• /
• 1998

본 논문에서는 고품질 한국어 합성을 위한 합성단위에 대해서 연구한다. 합성단위는 합성음의 음질을 좌우할 뿐만 아니라 전체 시스템의 크기에도 영향을 미친다. 음소와 같이 단위의 수가 적은 경우 적은 메모리로 시스템의 구성이 가능하지만 음운천이구간의 처리가 어려우며, 복합음소단위의 경우 많은 메모리를 요구하지만 음운천이특성을 잘 표현할 수 있는 장점이 있다. 본 논문에서는 합성단위가 한국어 합성음질에 미치는 영향을 분석하기 위하여 반음절, CVC형, VCV형 복합음소를 대상으로 음성을 합성하였다. 실험에 사용된 합성시스템은 최근 제안된 코퍼스에 기반한 합성시스템이다. 실험 전에 파악된 각 단위들의 통계적인 특성과 합성음의 음질을 비교한 결과 CVC형 복합음소가 제안된 시스템에 가장 적합한 합성단위로 판정되었다.

• Proceedings of the IEEK Conference
• /
• 1998.06a
• /
• pp.19-22
• /
• 1998

In this study, the efficient channel coding method combined with CS-ACELP is proposed. The same convolutional coder and Viterbi decoder of COMA mobile communication system is used as channel coder. To make the best available use of limited channel coding redundancy, unequal error protection of punctured convolutional coder is used for variable reate allocation. But, the overall code rate is given by 2. The performance of proposed coder is analyzed and simulated in a Rayleigh fading channel. Experimental results show that the objective and subjective speech quality of variable rate channel coding methods are superior to those of non-variable channel coding method.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.25 no.12B
• /
• pp.2046-2051
• /
• 2000

본 논문에서는 기존의 음질 평가 방법들보다 우수할 뿐 아니라 다양한 채널 경로의 음성 신호에 대해서도 일관된 성능을 갖는 새로운 음질 평가 방법 PLP-CMS(Perceptual Linear Predictive-Cepstral Mean Subtraction)를 제안한다. CDMA PCS 이동 전화 환경에서 음성 신호의 주관적 음질을 효과적으로 예측할 수 있는 PLP-CMS는 심리 음향 선형 예측 분석(PLP Analysis: Perceptual Linear Predictive Analysis)을 이용하여 주관적 음질과의 상관 관계를 높였으며, 겝스트럼 평균 차감(CMS: Cepstral Mean Subtraction) 과정을 통하여 PSTN 경로에 무관하게 일관된 성능을 갖음을 확인하였다.

• Proceedings of the IEEK Conference
• /
• 2002.07b
• /
• pp.872-875
• /
• 2002

A digital watermarking technique applying psychoacoustic model for audio signal is proposed in this paper. In the watermarking scheme, the pseudo-random bit stream used as a watermark signal is embedded into the audio signal in both speech and music. The strength of the embedded signal is subject to the human auditory system in such a way that the disturbances on host audio signal are beyond the sensing of human ears. The experimental results show that the quality of the watermarked audio signal, in term of signal to noise ratio, can be improved up to 3.2 dB.

• Proceedings of the Korean Society of Marine Engineers Conference
• /
• 2005.11a
• /
• pp.57-58
• /
• 2005

Recently the telematics system is used widely. Users want to high quality communications. Since the primary advantage of using an array is to enhance a desired signal and reject jamming interferences, array signal processing is essential to satisfy unmet demand of user. In general, beamforming is a spatial filtering operation performed on the data received by an array of sensors. So we propose the beamformer design that use FPGA for real time processing. And we use zero-padding interpolation for high resolution data.

• MALSORI
• /
• no.53
• /
• pp.157-169
• /
• 2005

This paper proposes an efficiently representing method of spatial cues on multichannel spatial audio coding. The Binaural Cue Coding (BCC) method introduced recently represents multichannel audio signals by means of Inter Channel Level Difference (ICLD) or Source Index (SI). We tried to express more efficiently ICLD and SI information based on Inter Channel Correlation in this paper. We adopt different spatial cues according to ICC and propose a compensation method of empty spectrums created by using SI. We performed a MOS test and measuring spectral distortion. The results show that the proposed method can reduce the bitrate of side information without large degradation of the audio quality.

• MALSORI
• /
• no.67
• /
• pp.167-180
• /
• 2008

This paper describes a novel spectral envelope conversion method based on Gaussian mixture model (GMM). The core of this paper is rearranging source feature vectors in input space to the transformed feature vectors in feature space for the better modeling of GMM of source and target features. The quality of statistical modeling is dependent on the distribution and the dimension of data. The proposed method transforms both of the distribution and dimension of data and gives us the chance to model the same data with different configuration. Because the converted feature vectors should be on the input space, only source feature vectors are rearranged in the feature space and target feature vectors remain unchanged for the joint pdf of source and target features using KPCA. The experimental result shows that the proposed method outperforms the conventional GMM-based conversion method in various training environment.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.8
• /
• pp.3257-3269
• /
• 2020

Articulation disorders are characterized by an inability to achieve clear pronunciation due to misuse of the articulators. In this paper, a method of detecting such disorders by comparing to the standard pronunciations is proposed. This method defines the standard pronunciations from the speeches of normal children by clustering them with three features which are the Linear Predictive Cepstral Coefficient (LPCC), the Mel-Frequency Cepstral Coefficient (MFCC), and the Relative Spectral Analysis Perceptual Linear Prediction (RASTA-PLP). By calculating the distance between the centroid of the standard pronunciation and the inputted pronunciation, disordered speech whose features locates outside the cluster is detected. 89 children (58 of normal children and 31 of children with disorders) were recruited. 35 U-TAP test words were selected and each word's standard pronunciation is made from normal children and compared to each pronunciation of children with disorders. In the experiments, the pronunciations with disorders were successfully distinguished from the standard pronunciations.

• The Journal of the Acoustical Society of Korea
• /
• v.17 no.2
• /
• pp.43-49
• /
• 1998

본 논문에서는 고성능 CELP 음성 압축기를 위한 "Boaseline 코드벡터"와 "Implied 코드벡터"로 구성되는 새로운 구조의 코드북을 제안한다. Implied 코드벡터는 피치 주기 이 전의 합성음으로부터 구하여지며 여기(勵起)신호의 피치 구조를 강화하여 합성음의 음질을 향상시킨다. Implied 코드벡터는 전달되지 않고 인코더 및 디코더에서 각각 합성음을 이용 하여 독립적으로 구하여진다. 또한 펄스와 랜덤 성분을 모두 가지는 복합 여기방식을 이용 하여 음질을 더욱 향상시킨다. 제안된 코드북 구조를 이용하여 10msec프레임을 가지는 8kbps CELP 음성 압축기를 설계하여 하나의 DSP칩에 실시간 구현 하였고, 이것의 성능을 SNRseg와 MOS로 측정하였다. 평균 SNRseg는 12.14dB로 CS-ACELP의 SNRseg보다 6dB 높고, 조용한 환경에서의 MOS는 3.80으로 G.729 CS-ACELP의 MOS보다 0.02 높다.