• The Journal of the Acoustical Society of Korea
• /
• v.16 no.2
• /
• pp.31-35
• /
• 1997

In this paper, we propose a variable vocabulary word recognizer that is able to recognize new words not exist in training data. For the variable vocabulary word recognizer, we must have an on-line lexicon generator to transform new candidate words to the corresponding pronunciation sequences of phones without any large lexicon table. And, we also must make outputs. In order to model the phones and allophones reliably, we define Korean allophones by triphone clustering based on phonetic knowledge of preceding and succeeding phones of each phone. Using the clustering method, we generated 1,548 allophones with POW (Phonetically Optimized Words) 3,848 word DB. We evaluated the proposed word recognizer with POW 3,848 DB, PBW (Phonetically Balanced Words) 445 DB, and 244 word DB in hotel reservation task. Experimental results showed word recognition accuracy of 79.6% for the POW DB corresponding to vocabulary-dependent case, 79.4% in case of 445 word lexicon and 88.9% in case of 100 word lexicon for the PBW DB, and 71.4% for the hotel reservation DB corresponding to vocabulary-independent case.

• MALSORI
• /
• no.52
• /
• pp.1-14
• /
• 2004

In this study, we investigated acoustic characteristics of imitated utterances by a professional voice imitator, focusing on prosodic properties such as vowel formants and f0 distribution. To see the patterns of a voice imitation by a professional voice imitator, we compared the imitator's voice data with target speakers' voice data. The professional imitator, Mr. Bae produced utterances imitating the former President Kim's, the comedian Choi's, and the singer Bae's voices. Auditorily, the imitator was judged to imitate all the target speakers' voices successfully. However, acoustic examination showed that the imitator was better at imitating the singer Bae's voice in that the imitator's and the singer Bae's voices are more alike with respect to vowel formants and f0 distribution. We infer this is because the imitator's normal voice is very similar to the singer Bae's voice. On the other hand, the imitator's voice data showed that the patterns of vowel formants and f0 distribution found in the imitator's imitation voices of the other two target speakers were different from those of target speakers' voices.

• The Journal of the Acoustical Society of Korea
• /
• v.22 no.8
• /
• pp.727-739
• /
• 2003

In this paper, we propose a generation method of regression classes for adaptation in the HM-Net (Hidden Markov Network) system. The MLLR (Maximum Likelihood Linear Regression) adaptation approach is applied to the HM-Net speech recognition system for expressing the characteristics of speaker effectively and the use of HM-Net in various tasks. For the state level sharing, the context domain state splitting of PDT-SSS (Phonetic Decision Tree-based Successive State Splitting) algorithm, which has the contextual and time domain clustering, is adopted. In each state of contextual domain, the desired phoneme classes are determined by splitting the context information (classes) including target speaker's speech data. The number of adaptation parameters, such as means and variances, is autonomously controlled by contextual domain state splitting of PDT-SSS, depending on the context information and the amount of adaptation utterances from a new speaker. The experiments are performed to verify the effectiveness of the proposed method on the KLE (The center for Korean Language Engineering) 452 data and YNU (Yeungnam Dniv) 200 data. The experimental results show that the accuracies of phone, word, and sentence recognition system increased by 34∼37%, 9%, and 20%, respectively, Compared with performance according to the length of adaptation utterances, the performance are also significantly improved even in short adaptation utterances. Therefore, we can argue that the proposed regression class method is well applied to HM-Net speech recognition system employing MLLR speaker adaptation.

• MALSORI
• /
• no.43
• /
• pp.113-125
• /
• 2002

As there are needs to search the Web information via wire or wireless telephones, VoiceXML forum was established to develop and promote the Voice eXtensible Markup Language (VoiceXML). VoiceXML simplifies the creation of personalized interactive voice response services on the Web, and allows voice and phone access to information on Web sites, call center databases. Also, it can utilize the Web-based technologies, such as CGI(Common Gateway Interface) scripts. In this paper, we have developed the voice portal service platform based on VoiceXML called TeleGateway. It enables integration of voice services with data services using the Automatic Speech Recognition (ASR) and Text-To-Speech (TTS) engines. Also, we have showed the various services on voice portal services.

• MALSORI
• /
• no.52
• /
• pp.161-169
• /
• 2004

This paper describes the automated call routing for train ticket reservation aid system based on speech recognition. We focus on the task of automatically routing telephone calls based on user's fluently spoken response instead of touch tone menus in an interactive voice response system. Vector-based call routing algorithm is investigated and mapping table for key term is suggested. Korail database collected by KT is used for call routing experiment. We evaluate call-classification experiments for transcribed text from Korail database. In case of small training data, an average call routing error reduction rate of 14% is observed when mapping table is used.

• MALSORI
• /
• no.62
• /
• pp.97-112
• /
• 2007

The speech recognizer in general operates in noisy acoustical environments. Many research works have been done to cope with the acoustical variations. Among them, the multiple-HMM model approach seems to be quite effective compared with the conventional methods. In this paper, we consider a multiple-model approach combined with the model compensation method and investigate the necessary number of the HMM model sets through noisy speech recognition experiments. By using the data-driven Jacobian adaptation for the model compensation, the multiple-model approach with only a few model sets for each noise type could achieve comparable results with the re-training method.

• MALSORI
• /
• v.55
• /
• pp.61-76
• /
• 2005

This paper aims to analyze the acoustic effects of vowels produced in a distant-talking environment. The analysis was performed using a statistical method. The influence of gender and speakers on the variation was also examined. The speech data used in this study consist of 500 distant-talking words and 500 normal words of 10 speakers (5 males and 5 females). Acoustic features selected for the analysis were the duration, the formants (Fl and F2), the fundamental frequency and the total energy. The results showed that the duration, F0, F1 and the total energy increased in the distant-talking speech compared to normal speech; female speakers showed higher increase in all features except for the total energy and the fundamental frequency. In addition, speaker differences were observed.

• Speech Sciences
• /
• v.9 no.1
• /
• pp.111-135
• /
• 2002

An annotation model of a non-native speech database has been devised, wherein English is the target language and Korean is the native language. The proposed annotation model features overt transcription of predictable linguistic information in native speech by the dictionary entry and several predefined types of error specification found in native language transfer. The proposed model is, in that sense, different from other previously explored annotation models in the literature, most of which are based on native speech. The validity of the newly proposed model is revealed in its consistent annotation of 1) salient linguistic features of English, 2) contrastive linguistic features of English and Korean, 3) actual errors reported in the literature, and 4) the newly collected data in this study. The annotation method in this model adopts the widely accepted conventions, Speech Assessment Methods Phonetic Alphabet (SAMPA) and the TOnes and Break Indices (ToBI). In the proposed annotation model, SAMPA is exclusively employed for segmental transcription and ToBI for prosodic transcription. The annotation of non-native speech is used to assess speaking ability for English as Foreign Language (EFL) learners.

• Journal of the Korean Society of Laryngology, Phoniatrics and Logopedics
• /
• v.11 no.2
• /
• pp.141-145
• /
• 2000

Background and Objectives : The nasalance score measured by Nasometer is a supplementary data for the perceptually rated nasality by a trained speech pathologist. The nasalance score varies with subjects. The objective of the present study was to examine whether there are differences in nasalance scores as a function of age and sex. Materials and Method : This study used 20 normal chidren aged from 3 to 8 and 40 normal adults aged from 21 to 37(male : female= 1 : 1) as subjects. The nasalance scores were analyzed in 3 different phonetic contests(nasal, /i/ vowel,/a/ vowel) and 4 different sentence lengths(1, 2, 4, 8 syllable). Results : The children had significantly higher nasalance scores in short sentences an the adults. The female subjects had significantly higher nasalance scores in nasal sentences and in short sentences than the male subjects. Conclusion : These results may indicate that sex and age differences should be considered in the interpretation of the nasalance score in nasal sentences or in short sentences.

• MALSORI
• /
• no.46
• /
• pp.117-126
• /
• 2003

This paper presents our language model adaptation for Korean spontaneous speech recognition. Korean spontaneous speech is observed various characteristics of content and style such as filled pauses, word omission, and contraction as compared with the written text corpus. Our approaches focus on improving the estimation of domain-dependent n-gram models by relevance weighting out-of-domain text data, where style is represented by n-gram based tf/sup */idf similarity. In addition to relevance weighting, we use disfluencies as Predictor to the neighboring words. The best result reduces 9.7% word error rate relatively and shows that n-gram based relevance weighting reflects style difference greatly and disfluencies are good predictor also.