Language Acquisition 1
HomeFull List of Titles
1: ICSLP'98 Proceedings
Keynote SpeechesText-To-Speech Synthesis 1Spoken Language Models and Dialog 1Prosody and Emotion 1Hidden Markov Model Techniques 1Speaker and Language Recognition 1Multimodal Spoken Language Processing 1Isolated Word RecognitionRobust Speech Processing in Adverse Environments 1Spoken Language Models and Dialog 2Articulatory Modelling 1Talking to Infants, Pets and LoversRobust Speech Processing in Adverse Environments 2Spoken Language Models and Dialog 3Speech Coding 1Articulatory Modelling 2Prosody and Emotion 2Neural Networks, Fuzzy and Evolutionary Methods 1Utterance Verification and Word Spotting 1 / Speaker Adaptation 1Text-To-Speech Synthesis 2Spoken Language Models and Dialog 4Human Speech Perception 1Robust Speech Processing in Adverse Environments 3Speech and Hearing Disorders 1Prosody and Emotion 3Spoken Language Understanding Systems 1Signal Processing and Speech Analysis 1Spoken Language Generation and Translation 1Spoken Language Models and Dialog 5Segmentation, Labelling and Speech Corpora 1Multimodal Spoken Language Processing 2Prosody and Emotion 4Neural Networks, Fuzzy and Evolutionary Methods 2Large Vocabulary Continuous Speech Recognition 1Speaker and Language Recognition 2Signal Processing and Speech Analysis 2Prosody and Emotion 5Robust Speech Processing in Adverse Environments 4Segmentation, Labelling and Speech Corpora 2Speech Technology Applications and Human-Machine Interface 1Large Vocabulary Continuous Speech Recognition 2Text-To-Speech Synthesis 3Language Acquisition 1Acoustic Phonetics 1Speaker Adaptation 2Speech Coding 2Hidden Markov Model Techniques 2Multilingual Perception and Recognition 1Large Vocabulary Continuous Speech Recognition 3Articulatory Modelling 3Language Acquisition 2Speaker and Language Recognition 3Text-To-Speech Synthesis 4Spoken Language Understanding Systems 4Human Speech Perception 2Large Vocabulary Continuous Speech Recognition 4Spoken Language Understanding Systems 2Signal Processing and Speech Analysis 3Human Speech Perception 3Speaker Adaptation 3Spoken Language Understanding Systems 3Multimodal Spoken Language Processing 3Acoustic Phonetics 2Large Vocabulary Continuous Speech Recognition 5Speech Coding 3Language Acquisition 3 / Multilingual Perception and Recognition 2Segmentation, Labelling and Speech Corpora 3Text-To-Speech Synthesis 5Spoken Language Generation and Translation 2Human Speech Perception 4Robust Speech Processing in Adverse Environments 5Text-To-Speech Synthesis 6Speech Technology Applications and Human-Machine Interface 2Prosody and Emotion 6Hidden Markov Model Techniques 3Speech and Hearing Disorders 2 / Speech Processing for the Speech and Hearing Impaired 1Human Speech ProductionSegmentation, Labelling and Speech Corpora 4Speaker and Language Recognition 4Speech Technology Applications and Human-Machine Interface 3Utterance Verification and Word Spotting 2Large Vocabulary Continuous Speech Recognition 6Neural Networks, Fuzzy and Evolutionary Methods 3Speech Processing for the Speech-Impaired and Hearing-Impaired 2Prosody and Emotion 72: SST Student Day
SST Student Day - Poster Session 1SST Student Day - Poster Session 2Author Index
A B C D E F G H I
J K L M N O P Q R
S T U V W X Y Z
Multimedia Files
Non-Native Productions Of Japanese Single Stops That Are Too Long For One Mora Unit
Authors:
Yasuyo Minagawa-Kawai, University of Tokyo (Japan)
Shigeru Kiritani, University of Tokyo (Japan)
Page (NA) Paper number 1025
Abstract:
The difficulty for non-native speakers in producing Japanese geminate stops with a long enough closure has been pointed out and extensively studied. However, the reverse problem exists for particular language speakers such as Chinese. For these speakers, production of Japanese intervocalic single stops sound like geminate stops. This study aimed to show acoustic evidence of this problem in the production of Japanese voiceless stops by learners of Japanese, and to compare speakers from different language groups, English, Chinese and French speakers. The results of the experiment indicated that since some of the stop productions by Americans and Chinese were too long for one mora unit, the Japanese heard a geminate (2 mora) instead of a single. Because French subjects had a small difference in the stop durations from the Japanese value, only a few of their productions of single stops were perceived as geminates. The language differences in timing pattern for Japanese stops were interpreted to reflect differences in timing implementation of native languages that is stress-timed and syllable-timed, at least concerning the difference between English and French speakers.
SL981025.PDF (From Author)
SL981025.PDF (Rasterized)
TOP
The Process Of Generation And Development Of Second Language Japanese Accentuation
Authors:
Nobuko Yamada, Faculty of Humanities, Ibaraki University (Japan)
Page (NA) Paper number 1056
Abstract:
This study will investigate how non-native speakers of Japanese acquire Japanese accentuation from the viewpoint of the location of the accent nucleus. Hypothetical models for the process of generation and for developmental sequence of interlanguage Japanese accentuation, which is interim accentual system created by learners, will be proposed. The subjects appear to generate their interlanguage as the results of application of strategies or examples of accentuation. Those seem to be discovered from L2 input, or chosen and fetch from their memory. The subjects' competence of accentuation appear to be developed by L2 input, starting with L1 and universal property. They seem to discover and apply 5 types of strategies toward acquisition of target accentual rules of Japanese.
SL981056.PDF (From Author)
SL981056.PDF (Scanned)
TOP
Perceptual Properties of Russians with Japanese Fricatives
Authors:
Seiya Funatsu, Hiroshima Women's University (Japan)
Shigeru Kiritani, The University of Tokyo (Japan)
Page (NA) Paper number 809
Abstract:
This study investigated the perceptual properties of second language learners in acquiring second language phonemes. The case where the relation between two phonemes of a second language and those of a native language changes according to following vowels was studied. The perceptual properties of Russians with regards to Japanese fricatives were examined. In the perception test, the confusion of [(omitted)o] with [so] was very large. This phenomenon could be caused by the difference between the transition onset time from [s'] to vowels and that from the other consonants to vowels. It is considered that, in the case of following vowel [a] and [o], Russians equated Japanese [s] and [(omitted)] with Russian [s] and [s'] respectively. However, in the case of [u], they did not equate them in such a manner. This is probably because the acoustic properties of Japanese [(omitted)] are very different from those of Russian [u].
SL980809.PDF (From Author)
SL980809.PDF (Rasterized)
TOP
Assessment of Dutch Pronunciation by Means of Automatic Speech Recognition Technology
Authors:
Catia Cucchiarini, A2RT, University of Nijmegen (The Netherlands)
Febe De Wet, A2RT, University of Nijmegen (The Netherlands)
Helmer Strik, A2RT, University of Nijmegen (The Netherlands)
Louis Boves, A2RT, University of Nijmegen (The Netherlands)
Page (NA) Paper number 751
Abstract:
Experiments were carried out to determine whether log-likelihood ratios (LRs) can be employed to improve automatic assessment of Dutch pronunciation. Read speech of natives and non-natives was judged by three groups of expert raters and was then analyzed by means of a continuous speech recognizer. Three automatic measures were calculated, two LRs and rate of speech (ros), and then compared with the expert ratings. It appears that expert ratings of pronunciation quality can accurately be predicted on the basis of ros alone and that LRs do not contribute to better prediction. However, LRs can be useful to automatic pronunciation assessment because they can help detect fast speakers who produce totally wrong sentences.
SL980751.PDF (From Author)
SL980751.PDF (Rasterized)
TOP
Phonetic-Level Mispronunciation Detection in Non-Native Swedish Speech
Authors:
Philippe Langlais, CTT - KTH (Sweden)
Anne-Marie Öster, TMH-KTH (Sweden)
Björn Granström, CTT- KTH (Sweden)
Page (NA) Paper number 311
Abstract:
This contribution presents part of the work initiated at the CTT for the development of speech technology to assist non-native speakers learn Swedish. This study focuses mainly on the automatic location of mispronunciations at a phonetic level. We first describe the database we created for this work and then report on the reliability of several phonetic scores to automatically locate segmental problems in student utterances.
SL980311.PDF (From Author)
SL980311.PDF (Rasterized)
TOP
Computer-Based Second Language Production Training By Using Spectrographic Representation And HMM-Based Speech Recognition Scores
Authors:
Reiko Akahane-Yamada, ATR Human Information Processing Research Laboratories (Japan)
Erik McDermott, ATR Human Information Processing Research Laboratories (Japan)
Takahiro Adachi, ATR Human Information Processing Research Laboratories (Japan)
Hideki Kawahara, ATR Human Information Processing Research Laboratories (Japan)
John S. Pruitt, ATR Human Information Processing Research Laboratories (Japan)
Page (NA) Paper number 429
Abstract:
How can we provide feedback to second language (L2) learners about the goodness of their productions in an automatic way? In this paper, we introduce our attempts to provide effective feedback when we train native speakers of Japanese to produce English /r/ and /l/. First, we adopted spectrographic representation overlayed with formant frequencies as feedback. Second, we investigated the correlation between human judgments of L2 production quality and acoustic scores produced by an HMM-based speech recognition system. We also adopted the HMM-based scores as feedback in the production training. Evaluation of the pre- and post-training productions by human judges showed that production abilities of the trainees improved in both training groups, suggesting that both spectrographic representation and HMM-based scores were useful and meaningful as feedback. These results are discussed in the context of optimizing L2 speech training.