Acoustic Phonetics 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
Assimilation of Place in Japanese and Dutch
Authors:
Anne Cutler, Max Planck Institute for Psycholinguistics (The Netherlands)
Takashi Otake, Dokkyo University (Japan)
Page (NA) Paper number 93
Abstract:
Assimilation of place of articulation across a nasal and a following stop consonant is obligatory in Japanese, but not in Dutch. In four experiments the processing of assimilated forms by speakers of Japanese and Dutch was compared, using a task in which listeners blended pseudo-word pairs such as ranga-serupa. An assimilated blend of this pair would be rampa, an unassimilated blend rangpa. Japanese listeners produced significantly more assimilated than unassimilated forms, both with pseudo-Japanese and pseudo-Dutch materials, while Dutch listeners produced significantly more unassimilated than assimilated forms in each materials set. This suggests that Japanese listeners, whose native-language phonology involves obligatory assimilation constraints, represent the assimilated nasals in nasal-stop sequences as unmarked for place of articulation, while Dutch listeners, who are accustomed to hearing unassimilated forms, represent the same nasal segments as marked for place of articulation.
SL980093.PDF (From Author)
SL980093.PDF (Rasterized)
TOP
Prosodic Constraint on V-to-V Coarticulation in Japanese
Authors:
Yuko Kondo, Musashino Art University (Japan)
Yumiko Arai, Y & Y Phonetic Laboratory (Japan)
Page (NA) Paper number 186
Abstract:
Coarticulation is considered to be constrained by language specific grammar. In particular, prosodic units, which have an important function in organizing speech, are likely to affect the extent of coarticulation. The present study addresses the question of the relationship between V-to-V coarticulation and bimoraic foot in Japanese. It was shown that the foot boundary constrains the extent of coarticulation. It was also shown that both within- and across-foot, anticipatory effects are stronger than carryover effects in Japanese.
SL980186.PDF (From Author)
SL980186.PDF (Rasterized)
TOP
Postvocalic /r/-deletion in Standard Dutch: How Experimental Phonology Can Profit From ASR Technology
Authors:
Catia Cucchiarini, A2RT, University of Nijmegen (The Netherlands)
Henk van den Heuvel, A2RT, University of Nijmegen (The Netherlands)
Page (NA) Paper number 753
Abstract:
In this study automatic speech recognition (ASR) techniques were used to substantiate the findings of phonological research on postvocalic /r/-deletion in Standard Dutch. A database containing spontaneous speech utterances stemming from man-machine interactions in an automatic train-table inquiry system was used for this purpose. Pronunciation variants with and without /r/ were automatically generated on the basis of our specification of the phonological rule of /r/-deletion and were then included in the lexicon of a continuous speech recognizer (CSR) which was used in forced recognition mode. The results show that in a corpus containing 214,102 words, in which /r/-deletion could be applied 16,865 times, it was actually applied in 47.6% of the cases. This is a high percentage of occurrence for a phenomenon that has so sporadically been described. Furthermore, the results substantiate our rule specification: /r/-deletion occurs more often after schwa than after any other vowel.
SL980753.PDF (From Author)
SL980753.PDF (Rasterized)
TOP
More Evidence For The Perceptual Basis Of Sound Change? Suprasegmental Effects In The Development Of Distinctive Nasalization
Authors:
John Hajek, University of Melbourne (Australia)
Ian Watson, University of Oxford (U.K.)
Page (NA) Paper number 254
Abstract:
Cross-linguistic studies of the development of distinctive nasalization show evidence of significant suprasegmental conditioning. Amongst conditioning factors uncovered so far are vowel length and stress. Across languages it is reported that in the related contexts /V:N/ and /VN/, identical except for vowel length, phonologization of nasalization and N-deletion always occur preferentially in the context of long vowels. There is also cross-linguistic evidence of stress-conditioning of distinctive nasalization: nasalization and N-deletion appear to occur preferentially in stressed syllables. In this study, we discuss the results of an experiment designed to measure the possible effects of vowel duration and prominence on the perception of vowel nasalization. Both are seen to have an effect, although in different ways. Results presented here also lend support to the hypothesis that some sound changes, such as those involved in distinctive nasalization, may have a primarily perceptual basis.
SL980254.PDF (From Author)
SL980254.PDF (Rasterized)
TOP
Speech Production Of Vowel Sequences Using A Physiological Articulatory Model
Authors:
Jianwu Dang, ATR Human Information Processing Research Labs (Japan)
Kiyoshi Honda, ATR Human Information Processing Research Labs (Japan)
Page (NA) Paper number 639
Abstract:
This report describes the development of a physiologically-based articulatory model, which consists of the tongue, mandible, hyoid bone and vocal tract wall. These organs are represented as a midsagittal quasi-3D layer with a thickness of 2 cm for tongue tissue and 3 cm for tract wall. The geometry of these organs and muscles are extracted from volumetric MR images of a male speaker. Both the soft and rigid structures are represented by mass-points and viscoelastic springs for connective tissue, where the springs for bony organs are set to extremely large stiffness. This design is suitable to compute soft tissue deformations and rigid organ displacements simultaneously using a single algorithm, and thus reduces computational complexities of the simulation. A novel control method is developed to produce dynamic actions of the vocal tract, as well as to handle the collision of the tongue to surrounding walls. Area functions are obtained for vowel sequences based on model's vocal tract widths in the midsagittal and parasagittal planes. The proposed model demonstrated plausible dynamic behaviors for human speech articulation.