Articulatory Modelling 3
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
An Electropalatographic, Kinematic, and Acoustic Analysis of Supralaryngeal Correlates of Word-Level Prominence Contrasts in English
Authors:
Jonathan Harrington, Speech, Hearing and Language Research Centre, Macquarie University, Sydney (Australia)
Mary E. Beckman, Department of Linguistics, Ohio State University, Columbus, Ohio (USA)
Janet Fletcher, Department of Linguistics, University of Melbourne, Victoria (Australia)
Sallyanne Palethorpe, Speech, Hearing and Language research Centre, Macquarie University, Sydney (Australia)
Page (NA) Paper number 646
Abstract:
This study examines the phonetic characteristics of primary versus secondary stress on the first syllables of the surname 'Wheateron' and related adjective 'Wheateresque' in post-nuclear, deaccented position in a dialogue produced 40 times by 3 Australian English talkers. Synchronised acoustic, electromagnetometer, and electropalatographic recordings were analysed. One subject had a higher F0 in the primary stressed syllable. The other two had a longer acoustic duration for the syllable's voiced portion, corresponding to a longer lip closing movement. One of these two also had a larger and faster lip opening movement into the vowel. Taken together, the results show that primary versus secondary lexical stress may be differentiated even when accent contrasts are neutralised, although the differences are inconsistent across talkers and small by comparison to those that have been shown to characterise the accented-unaccented contrast.
SL980646.PDF (From Author)
SL980646.PDF (Rasterized)
TOP
Consistencies and Inconsistencies Between EPG and Locus Equation Data on Coarticulation
Authors:
Marija Tabain, Speech, Hearing and Languages Research Centre, Department of Linguistics, Macquarie University, Sydney. (Australia)
Page (NA) Paper number 668
Abstract:
Following a previous study using locus equation (LE) and electropalatographic (EPG) data to examine coarticulation of voiced consonants and vowels in CV syllables [1], the present study examines voiceless stops and fricatives using the same analysis techniques. It is found that when LE data for stops is sampled at the stop burst, rather than at vowel onset, the correlation between LE data on coarticulation and EPG data on coarticulation is quite high. By contrast, results for the fricatives are quite poor. It is suggested that the LE is capable of capturing rather gross differences in coarticulatory resistance, such as that involving a tongue tip rather than a tongue body, but that it is not capable of capturing more subtle differences in coarticulation, such as those involving different coronal articulations. This explanation is supported by work in progress on Australian Aboriginal languages which have up to four coronal places of articulation [2].
SL980668.PDF (From Author)
SL980668.PDF (Rasterized)
TOP
Synergy Between Jaw And Lips/Tongue Movements : Consequences In Articulatory Modelling
Authors:
Gérard Bailly, Institut de la Communication Parlée (ICP) (France)
Pierre Badin, Institut de la Communication Parlée (ICP) (France)
Anne Vilain, Institut de la Communication Parlée (ICP) (France)
Page (NA) Paper number 66
Abstract:
Linear component articulatory models are built using an iterative subtraction of linear predictors of the vocal tract geometry. In this paper we consider the contribution of jaw displacement to tongue and lips movements using sets of cineradiographic data from three different speakers. We show that linear prediction overestimates this contribution by capturing not only the intrinsic mechanical jaw-tongue coupling but also the synergetic control observed in the corpus. We then propose a subtraction of the jaw contribution which do not affect the performance of the model in terms of data prediction.
SL980066.PDF (From Author)
SL980066.PDF (Rasterized)
TOP
Modelling Tongue Configuration in German Vowel Production
Authors:
Philip Hoole, Phonetics Institute, Munich University (Germany)
Page (NA) Paper number 1096
Abstract:
The PARAFAC method of factor analysis was used to investigate patterns of tongue shaping in a corpus of 15 German vowels spoken in 3 consonant contexts by 7 speakers at 2 speech rates, using data from electromagnetic articulography. A two-factor model was extracted, giving a succinct, speaker-independent characterization of the German vowel space and of some important coarticulatory effects on vowel articulation. Moreover, the factors appeared to have a plausible physiological substrate. The PARAFAC model places strong constraints on the form that speaker-specific effects can take, since speaker differences must be captured in a single multiplicative weight per speaker and factor. While these constraints appeared acceptable for modelling vocalic aspects of articulation, more consonantally-related aspects, such as coarticulatory behaviour of the tongue-tip, appeared much more difficult to capture in the PARAFAC framework.
SL981096.PDF (From Author)
SL981096.PDF (Rasterized)
TOP
Optopalatograph: Real-time Feedback of Tongue Movement in 3D
Authors:
Alan A. Wrench, Queen Margaret College (U.K.)
Alan D. McIntosh, Queen Margaret College (U.K.)
Colin Watson, Queen Margaret College (U.K.)
William J. Hardcastle, Queen Margaret (U.K.)
Page (NA) Paper number 1117
Abstract:
In this paper the latest prototype Optopalatograph (OPG) is described and its operation is demonstrated graphically and in comparison to theoretical predictions. The system is divided into three parts - the optopalate itself; a separate self contained unit composed of 16 switched infra-red light sources, associated control logic and 16 receivers; and a computer with A/D card running software to analyse and interpret graphically the sensor outputs. The current prototype measures distances of up to 20mm between all of the 16 pre-selected points on the hard palate and the surface of the tongue at a frame rate of 100Hz. We conclude that the new prototype provides a practical measurement system with a subjectively informative real-time display but further development is required in order to obtain objective accuracy.
SL981117.PDF (From Author)
SL981117.PDF (Rasterized)
TOP
Effects of Contrastive Focal Accent on Linguopalatal Articulation and Coarticulation in the French [kskl] Cluster
Authors:
Yohann Meynadier, Laboratoire Parole et Langage, CNRS, ESA 6057 and Institut de Phonetique d'Aix-en-Provence (France)
Michel Pitermann, Department of Psychology, Queen's University of Kingston (Canada)
Alain Marchal, CNRS, DR 19, Caen (France)
Page (NA) Paper number 542
Abstract:
This paper investigates the effects of contrastive focal accent placement on lingual articulation and coarticulation of French [kskl] clusters in word-medial position. The EPG results show that (i) this type of accent does not systematically increase the amplitude, but the duration of linguopalatal constrictions (particularly their release); (ii) it directly lengthens the temporal interval between the articulatory hold phases of two contiguous consonants; (iii) no matter what the accent position is, it can affect the whole cluster; (iv) the gestural co-ordination of biconsonant sequences seemed to vary with the focal accent more according to articulatory than syllable boundary rhythmic constraints.