• Phonetics and Speech Sciences
• /
• v.10 no.4
• /
• pp.111-119
• /
• 2018

This paper proposes a method to improve speech recognition performance by extracting various pronunciations of the pseudo-morpheme unit from an eojeol unit corpus and generating a new recognition unit considering pronunciation variations. In the proposed method, we first align the pronunciation of the eojeol units and the pseudo-morpheme units, and then expand the pronunciation dictionary by extracting the new pronunciations of the pseudo-morpheme units at the pronunciation of the eojeol units. Then, we propose a new recognition unit that relies on pronunciation by tagging the obtained phoneme symbols according to the pseudo-morpheme units. The proposed units and their extended pronunciations are incorporated into the lexicon and language model of the speech recognizer. Experiments for performance evaluation are performed using the Korean speech recognizer with a trigram language model obtained by a 100 million pseudo-morpheme corpus and an acoustic model trained by a multi-genre broadcast speech data of 445 hours. The proposed method is shown to reduce the word error rate relatively by 13.8% in the news-genre evaluation data and by 4.5% in the total evaluation data.

• The Journal of the Acoustical Society of Korea
• /
• v.26 no.4
• /
• pp.159-165
• /
• 2007

In this paper, we propose a new pre-selection of candidate units that is suitable for the unit selection based Japanese TTS system. General pre-selection method performed by calculating a context-dependent cost within IP (Intonation Phrase). Different from other languages, however. Japanese has an accent represented as the height of a relative pitch, and several words form a single accentual phrase. Also. the prosody in Japanese changes in accentual phrase units. By reflecting such prosodic change in pre-selection. the qualify of synthesized speech can be improved. Furthermore, by calculating a context-dependent cost within accentual phrase, synthesis speed can be improved than calculating within intonation phrase. The proposed method defines AP. analyzes AP in context and performs pre-selection using accentual phrase matching which calculates CCL (connected context length) of the Phoneme's candidates that should be synthesized in each accentual phrase. The baseline system used in the proposed method is VoiceText, which is a synthesizer of Voiceware. Evaluations were made on perceptual error (intonation error, concatenation mismatch error) and synthesis time. Experimental result showed that the proposed method improved the qualify of synthesized speech. as well as shortened the synthesis time.

• Journal of Broadcast Engineering
• /
• v.1 no.2
• /
• pp.85-95
• /
• 1996

The purpose of this paper Is to propose a dubbed signal time-synchroniztion technique based on the SOLA(Synchronized Over-Lap and Add) method which has been widely used to modify the time scale of speech signal. In broadcasting audio recording environments, the high degree of background noise requires dubbing process. Since the time difference between the original and the dubbed signal ranges about 200mili seconds, process is required to make the dubbed signal synchronize to the corresponding image. The proposed method finds he starting point of the dubbing signal using the short-time energy of the two signals. Thereafter, LPC cepstrum analysis and DTW(Dynamic Time Warping) process are applied to synchronize phoneme positions of the two signals. After determining the matched point by the minimum mean square error between orignal and dubbed LPC cepstrums, the SOLA method is applied to the dubbed signal, to maintain the consistency of the corresponding phase. Effectiveness of proposed method is verified by comparing the waveforms and the spectrograms of the original and the time synchronized dubbing signal.

• Proceedings of the Acoustical Society of Korea Conference
• /
• spring
• /
• pp.337-340
• /
• 2004

The exactly reproduced prosody of a TTS system is one of the key factors that affect the naturalness of synthesized speech. In general, rules about prosody had been gathered either from linguistic knowledge or by analyzing the prosodic information from natural speech. But these could not be perfect and some of them could be incorrect. So we proposed artificial neural network(ANN)s that can be trained to team the prosody of natural speech and generate it. In learning phase, let ANNs learn the pitch and energy contour of center phoneme by applying a string of phonemes in a sentence to ANNs and comparing the output pattern with target pattern and making adjustment in weighting values to get the least mean square error between them. In test phase, the estimation rates were computed. We saw that ANNs could generate the prosody of a sentence.

• 재활복지
• /
• v.20 no.4
• /
• pp.229-246
• /
• 2016

The aim of this study is to conduct superordinate word selection task to compare their performance and reaction time, and superordinate word writing task to compare the differences in their performance and error pattern in 40 elderly adults and 43 young adults. As a result, first, in both tasks, elderly adults had a smaller number of correct responses. Second, elderly adults showed slower reaction time than young adults. Third, in superordinate word writing task, elderly adults showed more relevant errors than irrelevant errors. The reason elderly adults had a smaller number of correct responses in both tasks was that the links among the pieces of information in the semantic lexicon weakened or deteriorated due to normal aging. Slower reaction time was based on neurophysiological changes of the brain and cognitive processing speed. In addition, the relevant errors showed that they could access the lexicon for target words and produce explanation the relevant characteristics, even though they could not retrieve the target words.

• The Journal of the Acoustical Society of Korea
• /
• v.25 no.2
• /
• pp.88-95
• /
• 2006

In this Paper. a method for improving the performance of automatic speech recognition (ASR) system for conversational speech is proposed. which mainly focuses on increasing the robustness against the rapidly speaking utterances. The proposed method doesn't require an additional speech recognition task to represent speaking rate quantitatively. Energy distribution for special bands is employed to detect the vowel regions, the number of vowels Per unit second is then computed as speaking rate. To improve the Performance for fast speech. in the pervious methods. a sequence of the feature vectors is expanded by a given scaling factor, which is computed by a ratio between the standard phoneme duration and the measured one. However, in the method proposed herein. utterances are classified by their speaking rates. and the scaling factor is determined individually for each class. In this procedure, a maximum likelihood criterion is employed. By the results from the ASR experiments devised for the 10-digits mobile phone number. it is confirmed that the overall error rate was reduced by $17.8\%$ when the proposed method is employed

• Journal of KIISE:Software and Applications
• /
• v.37 no.10
• /
• pp.788-801
• /
• 2010

Advancing of mobile device is remarkable, so the research on mobile input device is getting more important issue. There are lots of input devices such as keypad, QWERTY keypad, touch and speech recognizer, but they are not as convenient as typical keyboard-based desktop input devices so input strings usually contain many typing errors. These input errors are not trouble with communication among person, but it has very critical problem with searching in database, such as dictionary and address book, we can not obtain correct results. Especially, Hangeul has more than 10,000 different characters because one Hangeul character is made by combination of consonants and vowels, frequency of error is higher than English. Generally, suffix tree is the most widely used data structure to deal with errors of query, but it is not enough for variety errors. In this paper, we propose fast approximate Korean word searching system, which allows variety typing errors. This system includes several algorithms for applying general approximate string searching to Hangeul. And we present profanity filters by using proposed system. This system filters over than 90% of coined profanities.

• Journal of KIISE:Software and Applications
• /
• v.29 no.3
• /
• pp.167-179
• /
• 2002

In this paper, we propose a new acoustic model for characterizing segmental features and an algorithm based upon a general framework of hidden Markov models (HMMs) in order to compensate the weakness of HMM assumptions. The segmental features are represented as a trajectory of observed vector sequences by a polynomial regression function because the single frame feature cannot represent the temporal dynamics of speech signals effectively. To apply the segmental features to pattern classification, we adopted segmental HMM(SHMM) which is known as the effective method to represent the trend of speech signals. SHMM separates observation probability of the given state into extra- and intra-segmental variations that show the long-term and short-term variabilities, respectively. To consider the segmental characteristics in acoustic model, we present segmental-feature HMM(SFHMM) by modifying the SHMM. The SFHMM therefore represents the external- and internal-variation as the observation probability of the trajectory in a given state and trajectory estimation error for the given segment, respectively. We conducted several experiments on the TIMIT database to establish the effectiveness of the proposed method and the characteristics of the segmental features. From the experimental results, we conclude that the proposed method is valuable, if its number of parameters is greater than that of conventional HMM, in the flexible and informative feature representation and the performance improvement.

• Journal of the Korea Society of Computer and Information
• /
• v.16 no.12
• /
• pp.83-92
• /
• 2011

Typographical errors by the author's mistyping occur frequently in a document being prepared with word processors contrary to usual publications. Preparing this online document, the most common orthographical errors are spelling errors resulting from incorrectly typing intent keys to near keys on keyboard. Typical spelling checkers detect and correct these errors by using morphological analyzer. In other words, the morphological analysis module of a speller tries to check well-formedness of input words, and then all words rejected by the analyzer are regarded as misspelled words. However, if morphological analyzer accepts even mistyped words, it treats them as correctly spelled words. In this paper, I propose a simple method capable of detecting and correcting errors that the previous methods can not detect. Proposed method is based on the characteristics that typographical errors are generally not repeated and so tend to have very low frequency. If words generated by operations of deletion, exchange, and transposition for each phoneme of a low frequency word are in the list of high frequency words, some of them are considered as correctly spelled words. Some heuristic rules are also presented to reduce the number of candidates. Proposed method is able to detect not syntactic errors but some semantic errors, and useful to scoring candidates.

• Phonetics and Speech Sciences
• /
• v.16 no.3
• /
• pp.71-78
• /
• 2024

This study aimed to establish evaluation methods for the speech processing stages of phonological encoding, phonological short-term memory, and articulation transcoding from a psycholinguistic perspective. A meta-analysis of 21 studies published between 2000 and 2024, involving 1,442 participants, was conducted. Participants were divided into six groups: general, dyslexia, speech sound disorder, language delay, apraxia+aphasia, and childhood apraxia of speech. The analysis revealed effect sizes of g=.46 for phonological encoding errors, g=.57 for phonological short-term memory errors, and g=.63 for articulation transition errors. These results suggest that substitution errors, order and repetition errors, and phoneme addition and voicing substitution errors are key indicators for assessing these abilities. This study contributes to a comprehensive understanding of speech and language disorders by providing a methodological framework for evaluating speech processing stages and a detailed analysis of error characteristics. Future research should involve non-word repetition tasks across various speech and language disorder groups to further validate these methods, offering valuable data for the assessment and treatment of these disorders.