• Journal of the Korea Society of Computer and Information
• /
• v.20 no.12
• /
• pp.145-152
• /
• 2015

In this paper, a standardized phonological rule should be embodied for all people who want to use Hangul beyond nationality. This researcher suggested a phonological process protocol in unicode for the purpose of developing a computer system-based program as one of methods to pronounce Hangul correctly. First of all, to solve a problem of standardizing Hangul pronunciation, it is necessary to develop a phonological process system. In other words, when a mobile phone or a keyboard is used, inputted strings should be changed exactly by an automatic device made according to the phonological rule. For this, a phonological process table is the most important and unicode which digitalizes Hangul is the most effective. Therefore, the purpose of this study is to complete a phonological process table, suggest an algorithm by the table, and develop a computer program pursuing correct Hangul pronunciation. For the methodology of this study, the features of unicode was used to solve a problem of coding each syllable by phonemes after strings were inputted from the phonological process program. This study suggested a protocol that sounds of inputted strings were automatically changed via recombination using a phonological process table. The ultimate purpose of the study was to help pronounce Hangul correctly according to the pronunciation rule.

• Proceedings of the KSPS conference
• /
• 2004.05a
• /
• pp.239-242
• /
• 2004

An automatic pronunciation correction system provides users with correction guidelines for each pronunciation error. For this purpose, we propose a speech recognition system which automatically classifies pronunciation errors when Koreans speak a foreign language. In this paper, we also propose machine scoring methods for automatic assessment of pronunciation quality by the speech recognizer. Scores obtained from an expert human listener are used as the reference to evaluate the different machine scores and to provide targets when training some of algorithms. We use a log-likelihood score and a normalized log-likelihood score as machine scoring methods. Experimental results show that the normalized log-likelihood score had higher correlation with human scores than that obtained using the log-likelihood score.

• Speech Sciences
• /
• v.7 no.2
• /
• pp.245-253
• /
• 2000

In this paper, we implement an elementary system to correct accents and pronunciations in English words spoken by non-native English speakers. In case of the accent evaluation, energy and pitch information are used to find stressed syllables, and then we extract the segment information of input patterns using a dynamic time warping method to discriminate and evaluate accent position. For the pronunciation evaluation, we utilize the segment information using the same algorithm as in accent evaluation, and perform the spectral distance measure for each phoneme between input patterns and reference patterns. Based on these spectral distances, we decide whether to recommend the pronunciation correction or not. Our results show that 98 percent of accent and 71 percent of pronunciation evaluation agree with the perceptual measure.

• Proceedings of the KSPS conference
• /
• 2000.07a
• /
• pp.241-254
• /
• 2000

The past decade has witnessed the abundant use of computer in testing language skills such as listening and reading. Compared with these language skills, we have experienced little use of computer in testing a speaking skill including pronunciation. This is largely due to limitations of the current computer technology. One of such limitations for testing pronunciation is to store and automatically evaluate what the learner utters. Due to this limitation, the computer simply stores what the learner utters and raters evaluate it afterward on a certain rating continuum. With the advent of voice recognition technology, however, the computer has been able to test pronunciation in a systematic way. This technology enables the computer to identify, visually show, and evaluate the learner's intonation pattern by means of autocorrection. The evaluation is expressed in terms of the degree in which the learner's intonation pattern overlaps with that of the native speaker of the target language. In particular, the degree is numerically displayed on the screen, and this numeral is considered as the score of the learner's utterance under our testing framework.

• Journal of KIISE:Software and Applications
• /
• v.34 no.2
• /
• pp.104-111
• /
• 2007

This paper examines how the performance of an automatic speech recognizer was improved for Korean Point-of-Interest (POI) data by modeling pronunciation variation using structural prosodic information such as prosodic words and syllable length. First, multiple pronunciation variants are generated using prosodic words given that each POI word can be broken down into prosodic words. And the cross-prosodic-word variations were modeled considering the syllable length of word. A total of 81 experiments were conducted using 9 test sets (3 baseline and 6 proposed) on 9 trained sets (3 baseline, 6 proposed). The results show: (i) the performance was improved when the pronunciation lexica were generated using prosodic words; (ii) the best performance was achieved when the maximum number of variants was constrained to 3 based on the syllable length; and (iii) compared to the baseline word error rate (WER) of 4.63%, a maximum of 8.4% in WER reduction was achieved when both prosodic words and syllable length were considered.

• MALSORI
• /
• no.49
• /
• pp.95-105
• /
• 2004

An automatic pronunciation correction system provides learners with correction guidelines for each mispronunciation. For this purpose we develope a speech recognizer which automatically classifies pronunciation errors when Koreans speak a foreign language. In order to develope the methods for automatic assessment of pronunciation quality, we propose a language model based score as a machine score in the speech recognizer. Experimental results show that the language model based score had higher correlation with human scores than that obtained using the conventional log-likelihood based score.

• Speech Sciences
• /
• v.11 no.2
• /
• pp.217-226
• /
• 2004

An automatic pronunciation correction system provides users with correction guidelines for each pronunciation error. For this purpose, we develop a speech recognition system which automatically classifies pronunciation errors when Koreans speak a foreign language. In this paper, we propose a machine scoring method for automatic assessment of pronunciation quality by the speech recognizer. Scores obtained from an expert human listener are used as the reference to evaluate the different machine scores and to provide targets when training some of algorithms. We use a log-likelihood score and a normalized log-likelihood score as machine scoring methods. Experimental results show that the normalized log-likelihood score had higher correlation with human scores than that obtained using the log-likelihood score.

• Journal of Korean Library and Information Science Society
• /
• v.25
• /
• pp.167-206
• /
• 1996

Some conclusions can be derived form the study : 1) The script of Japan author's name for 3 years(1960-1962, not published by the National Assembly Library but by Korea Library Association)and that of 1963's was arranged by their mother tongue although they had not the "author index". 2) "Periodical's Index" which the publication of National Assembly Library was not accept the principle that the person's name should be pronounce and script by one's mother tongue. It means that the Library was not accept the uniqueness of personal name. 3) Because the arrangement of the same person's name is mixed with one's mother tongue pronunciation and Korean one that they are scattered each another. 4) The same surname and the same Chinese character has different arrangement because of pronunciation rule of Korean language. 5) The same person's name was regarded as a different one because of nonaccurate name transcription. 6) A Japanese name was transcribed as Hangul with Korean pronunciation. 7) A Japanese name was transcribed as Hangul with Korean pronunciation and added Chinese Character in parenthesis. 8) A same Japanese name was regarded as a different one when it was transcribed with Chinese character and Hangul. 9) The arrangement of a same person's name was different when between the surname and forename has one space and has not. 10) "Author Index" is not playing as a role of name authority file.a role of name authority file.

• Journal of the Korean BIBLIA Society for library and Information Science
• /
• v.4 no.1
• /
• pp.1-22
• /
• 1980

It was analysed and considered about entry methods of headings in Japanese book cataloging which using methods of major universities and public libraries in our nation at present. I was suggested and clarified one's position that authors heading must make an entry to our pronunciation after translating from sound of chines letters, title headings must make an entry translation from Japanese title or Chinese title, to Korean. On the cataloging of authors heading or title heading, if cataloging make an entry to Japanese pronunciation it have many faults as follow. (1) In the title and authors heading of Japanese books, if it make an entry as Kana letter, not only non skilled librarian at Japanese language will have many faults but also non skilled users to Japanese language or pronunciation will very difficult to reference in title and authors catalog card. (2) If it Romanize the catalog card of title and authors heading in the Japanese books, librarian of the catalog part will make efforts and take pains to the catalog work. In addition, catalog users which is skilled in Japanese language but can not understand or unknown to the Romanized entry, and which if not understand to the, methods of entry will very difficult to card investigation or can not refer to catalog card. (3) If it make an entry to Korean in the authors and title heading from Japanese sound or pronunciation, librairans and library users which is unskilful as Japanese language will be inconvenience to works and investigation of library materials.

• Phonetics and Speech Sciences
• /
• v.2 no.4
• /
• pp.51-58
• /
• 2010

English vowels play an important role in the daily communication between Korean students and international visitors. However, many Korean students still have difficulty producing them distinctively. Vowels vary according to shapes of oral and pharyngeal cavities, which are mainly determined by the degree of jaw opening and tongue position. Yang (2008a) proposed a simplified chart of English and Korean vowels for an educational purpose. He also suggested to use wooden chopsticks to secure distinguishable jaw openings. The purpose of this study is to tap whether wooden chopsticks can be applicable to a method for correcting English vowel pronunciation. Twelve male and female students participated in the recordings of eight /hVd/ words followed by additional recordings with wooden chopsticks between upper and lower teeth. The first and second formant trajectories of both natural and controlled vowel productions were obtained and compared at six equidistant measurement points using Praat. Results showed that the formant values of natural vowel productions were comparable to those of controlled productions. Vowels with similar formant trajectories of male students were separated with the aid of chopsticks. The width of each chopstick could be controlled similarly in the experiment. The author concludes that wooden chopsticks can be useful to correct vowel pronunciation. Further studies are desirable for native speakers to make perceptual evaluations of controlled vowel productions by nonnative speakers.