• The Journal of the Acoustical Society of Korea
• /
• v.31 no.4
• /
• pp.266-272
• /
• 2012

Speaker segmentation is a process of automatically detecting the speaker boundary points in the audio data. Speaker segmentation methods are divided into two categories depending on whether they use a prior knowledge or not: One is the model-based segmentation and the other is the metric-based segmentation. In this paper, we introduce the eigenvoice-based speaker weight distance method and compare it with the representative metric-based methods. Also, we employ and compare the Euclidean and cosine similarity functions to calculate the distance between speaker weight vectors. And we verify that the speaker weight distance method is computationally very efficient compared with the method directly using the distance between the speaker adapted models constructed by the eigenvoice technique.

• Proceedings of the Acoustical Society of Korea Conference
• /
• spring
• /
• pp.91-94
• /
• 2000

최근 운율 구조와 문장구조 및 음운규칙과 관련 된 많은 언어학적 연구가 이루어져, 언어 이해 차원에서 의미 정보, 문장 구조 정보, discourse structure 등을 위한 운율 정보의 유용성이 입증되었으나, 이러한 결과가 최근의 음성인식 시스템에는 거의 적용되지 못하고 있다. 본 연구에서는 계층적인 방법을 기초로 하여 한국어의 연속음성으로부터 운율구를 검출하는 세그멘테이션법을 제안하였다. 우선, 입력된 음성으로부터 문장단위의 경계를 검출하기 위하여 휴지기를 이용하였으며 에너지, 휴지기의 지속시간 및 피치궤적을 참조하여 강세구의 경계를 검출하였다. 실험음성의 텍스트는 "만물상"이며, 남녀 각 2명의 표준어 화자가 빠른 속도와 보통 속도로 낭독한 음성데이터를 대상으로 비교하였다.

• The Journal of the Acoustical Society of Korea
• /
• v.22 no.2
• /
• pp.135-144
• /
• 2003

A method for the automatic segmentation of speech signals is described. The method is dedicated to the construction of a large database for a Text-To-Speech (TTS) synthesis system. The main issue of the work involves the refinement of an initial estimation of phone boundaries which are provided by an alignment, based on a Hidden Market Model(HMM). Multi-layer perceptron (MLP) was used as a phone boundary detector. To increase the performance of segmentation, a technique which individually trains an MLP according to phonetic transition is proposed. The optimum partitioning of the entire phonetic transition space is constructed from the standpoint of minimizing the overall deviation from hand labelling positions. With single speaker stimuli, the experimental results showed that more than 95% of all phone boundaries have a boundary deviation from the reference position smaller than 20 ms, and the refinement of the boundaries reduces the root mean square error by about 25%.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• v.9 no.2
• /
• pp.1019-1022
• /
• 2005

As recent information communication technology is rapidly developed, a lot of researches related to speech signal processing have been processed. So pitch period is applied as important factor to many application fields such as speech recognition, speaker identification, speech analysis and synthesis. Therefore, many algorithms related to pitch detection have been proposed in time domain and frequency domain and AMDF(average magnitude difference function) which is one of pitch detection algorithms in time domain chooses time interval from valley to valley as pitch period. But, in selection of valley point to detect pitch period, complexity of the algorithm is increased. So in this paper we proposed pitch detection algorithm using rotation transform of AMDF, that taking the global minimum valley point as pitch period and established a threshold about the phoneme in beginning portion, to exclude pitch period selection. and compared existing methods with proposed method through simulation.