Intonation Pattern Analysis of Phonemes in MISING Language

Similar documents
The Acquisition of English Intonation by Native Greek Speakers

Mandarin Lexical Tone Recognition: The Gating Paradigm

L1 Influence on L2 Intonation in Russian Speakers of English

Word Stress and Intonation: Introduction

Revisiting the role of prosody in early language acquisition. Megha Sundara UCLA Phonetics Lab

Rhythm-typology revisited.

Proceedings of Meetings on Acoustics

Modern TTS systems. CS 294-5: Statistical Natural Language Processing. Types of Modern Synthesis. TTS Architecture. Text Normalization

Phonological Processing for Urdu Text to Speech System

A Cross-language Corpus for Studying the Phonetics and Phonology of Prominence

Acoustic correlates of stress and their use in diagnosing syllable fusion in Tongan. James White & Marc Garellek UCLA

A survey of intonation systems

Demonstration of problems of lexical stress on the pronunciation Turkish English teachers and teacher trainees by computer

Journal of Phonetics

Quarterly Progress and Status Report. Voiced-voiceless distinction in alaryngeal speech - acoustic and articula

Rachel E. Baker, Ann R. Bradlow. Northwestern University, Evanston, IL, USA

Quarterly Progress and Status Report. VCV-sequencies in a preliminary text-to-speech system for female speech

Automatic intonation assessment for computer aided language learning

Pobrane z czasopisma New Horizons in English Studies Data: 18/11/ :52:20. New Horizons in English Studies 1/2016

ADDIS ABABA UNIVERSITY SCHOOL OF GRADUATE STUDIES MODELING IMPROVED AMHARIC SYLLBIFICATION ALGORITHM

**Note: this is slightly different from the original (mainly in format). I would be happy to send you a hard copy.**

The influence of metrical constraints on direct imitation across French varieties

The Perception of Nasalized Vowels in American English: An Investigation of On-line Use of Vowel Nasalization in Lexical Access

SEGMENTAL FEATURES IN SPONTANEOUS AND READ-ALOUD FINNISH

Speech Recognition using Acoustic Landmarks and Binary Phonetic Feature Classifiers

Program Matrix - Reading English 6-12 (DOE Code 398) University of Florida. Reading

Universal contrastive analysis as a learning principle in CAPT

Copyright by Niamh Eileen Kelly 2015

Eli Yamamoto, Satoshi Nakamura, Kiyohiro Shikano. Graduate School of Information Science, Nara Institute of Science & Technology

THE PERCEPTION AND PRODUCTION OF STRESS AND INTONATION BY CHILDREN WITH COCHLEAR IMPLANTS

Consonants: articulation and transcription

Florida Reading Endorsement Alignment Matrix Competency 1

Speech Segmentation Using Probabilistic Phonetic Feature Hierarchy and Support Vector Machines

Consonant-Vowel Unity in Element Theory*

Beginning primarily with the investigations of Zimmermann (1980a),

GOLD Objectives for Development & Learning: Birth Through Third Grade

age, Speech and Hearii

THE SURFACE-COMPOSITIONAL SEMANTICS OF ENGLISH INTONATION MARK STEEDMAN. University of Edinburgh

1. REFLEXES: Ask questions about coughing, swallowing, of water as fast as possible (note! Not suitable for all

SOUND STRUCTURE REPRESENTATION, REPAIR AND WELL-FORMEDNESS: GRAMMAR IN SPOKEN LANGUAGE PRODUCTION. Adam B. Buchwald

Discourse Structure in Spoken Language: Studies on Speech Corpora

Unvoiced Landmark Detection for Segment-based Mandarin Continuous Speech Recognition

Speech Recognition at ICSI: Broadcast News and beyond

Body-Conducted Speech Recognition and its Application to Speech Support System

2,1 .,,, , %, ,,,,,,. . %., Butterworth,)?.(1989; Levelt, 1989; Levelt et al., 1991; Levelt, Roelofs & Meyer, 1999

Phonological and Phonetic Representations: The Case of Neutralization

Design Of An Automatic Speaker Recognition System Using MFCC, Vector Quantization And LBG Algorithm

Infants learn phonotactic regularities from brief auditory experience

INTRODUCTION. 512 J. Acoust. Soc. Am. 105 (1), January /99/105(1)/512/10/$ Acoustical Society of America 512

Local and Global Acoustic Correlates of Information Structure in Bulgarian

Speech Emotion Recognition Using Support Vector Machine

Cambridgeshire Community Services NHS Trust: delivering excellence in children and young people s health services

Contrastiveness and diachronic variation in Chinese nasal codas. Tsz-Him Tsui The Ohio State University

Phonological encoding in speech production

18 The syntax phonology interface

Designing a Speech Corpus for Instance-based Spoken Language Generation

Analysis of Emotion Recognition System through Speech Signal Using KNN & GMM Classifier

The analysis starts with the phonetic vowel and consonant charts based on the dataset:

DEVELOPMENT OF LINGUAL MOTOR CONTROL IN CHILDREN AND ADOLESCENTS

Prosody in Speech Interaction Expression of the Speaker and Appeal to the Listener

Bitonal lexical pitch accents in the Limburgian dialect of Borgloon

An Acoustic Phonetic Account of the Production of Word-Final /z/s in Central Minnesota English

Surface Structure, Intonation, and Meaning in Spoken Language

Class-Discriminative Weighted Distortion Measure for VQ-Based Speaker Identification

Christine Mooshammer, IPDS Kiel, Philip Hoole, IPSK München, Anja Geumann, Dublin

Large Kindergarten Centers Icons

Fix Your Vowels: Computer-assisted training by Dutch learners of Spanish

Clinical Application of the Mean Babbling Level and Syllable Structure Level

Intra-talker Variation: Audience Design Factors Affecting Lexical Selections

REVIEW OF CONNECTED SPEECH

Speaker Recognition. Speaker Diarization and Identification

Voice conversion through vector quantization

Get Your Hands On These Multisensory Reading Strategies

Phonological Encoding in Sentence Production

A comparison of spectral smoothing methods for segment concatenation based speech synthesis

Collecting dialect data and making use of them an interim report from Swedia 2000

Structure and Intonation in Spoken Language Understanding

Online Publication Date: 01 May 1981 PLEASE SCROLL DOWN FOR ARTICLE

THE MULTIVOC TEXT-TO-SPEECH SYSTEM

To appear in the Proceedings of the 35th Meetings of the Chicago Linguistics Society. Post-vocalic spirantization: Typology and phonetic motivations

Bare Root Nodes in Basaa

Correspondence between the DRDP (2015) and the California Preschool Learning Foundations. Foundations (PLF) in Language and Literacy

Segregation of Unvoiced Speech from Nonspeech Interference

Infants Perception of Intonation: Is It a Statement or a Question?

The IRISA Text-To-Speech System for the Blizzard Challenge 2017

DOWNSTEP IN SUPYIRE* Robert Carlson Societe Internationale de Linguistique, Mali

LISTENING STRATEGIES AWARENESS: A DIARY STUDY IN A LISTENING COMPREHENSION CLASSROOM

Regional variation in the realization of intonation contours in the Netherlands

Perceived speech rate: the effects of. articulation rate and speaking style in spontaneous speech. Jacques Koreman. Saarland University

Phonetics. The Sound of Language

A Fact in Historical Phonology from the Viewpoint of Generative Phonology: The Underlying Schwa in Old English

CLASSIFICATION OF PROGRAM Critical Elements Analysis 1. High Priority Items Phonemic Awareness Instruction

An argument from speech pathology

English Language and Applied Linguistics. Module Descriptions 2017/18

ABSTRACT. Some children with speech sound disorders (SSD) have difficulty with literacyrelated

Perceptual scaling of voice identity: common dimensions for different vowels and speakers

Linking object names and object categories: Words (but not tones) facilitate object categorization in 6- and 12-month-olds

IEEE Proof Print Version

Expressive speech synthesis: a review

Eyebrows in French talk-in-interaction

Transcription:

Intonation Pattern Analysis of Phonemes in MISING Language Rizwan Rehman Centre for Computer Studies Dibrugarh University Gopal Chandra Hazarika Department of Mathematics Dibrugarh University Devid Kardong Department of Life Sciences Dibrugarh University ABSTRACT Intonation is one the basic building blocks of the prosody. The fundamental frequency (F0) represents the intonation acoustically. This is the first study into the analysis of intonation pattern in MISING language. In MISING language it is found that the syllables occur at a regular interval and hence it is syllable timed. Since the syllables are the phonological building blocks of a word they can influence the rhythm of the language and also its pattern of stress apart from other factors. The present study aims at finding the intonation pattern of words in MISING language. Words of the structure VCV, VCVC, VCCV, VCCVC, CVCV, CVCVC, CVCCV and VCCCVC are observed and analyzed both in terms of pitch variations and fundamental frequency (F0) the pitch contour are also analyzed in this paper. The tone pattern are marked with the help of Tone and Break Index model (ToBI) Keywords Natural Language Processing, Intonation, Fundamental frequency, Pitch, ToBI, Syllable, MISING 1. INTRODUCTION Intonation refers to the pitch or tune of the utterance [Ladd 1996]. It is one of the major components of the prosody. Intonation as such is not used to distinguish words but it has some major function such as indicating the attitude and emotion of the speaker as well as differentiating factors between a statement and a question [Ladd 2000]. The speaker can convey the meaning of the word by employing pitch in different ways. Pitch can be analyzed using fundamental frequency (F0) of the speaker s vocal cord while speaking a sentence. There are four classes into which intonation pattern are categorized: i. Pitch increase over time ii. Pitch decrease over time iii. Pitch fall and rise iv. Pitch rise and fall In linguistics it is often found that a same set of words can be associated with different tunes which can be done by rising or falling pitch [Liberman and Prince 1977]. While speaking a sentence the intonation which is associated can be heard along with some relatively prominent word or accented words. This phenomenon is found because the pitch accent which is the unit of intonation is synchronized with the vowel of the primary stress syllable of each word that is accented. These intonation patterns are language dependent. 2. STRUCTURE OF SYLLABLE AND WORD Syllables are the phonological building block of the words. Syllables have an impact on the rhythm of the language its prosody and stress pattern. Syllabic nucleus is the main component of the syllable which is occupied by vowel. In almost all the major language across the globe, vowels form the nucleus. The vowels that precede the nucleus are the ONSET and vowels after the nucleus are called the CODA [Selkirk 1984]. The nucleus of the syllable is sonorant, which makes the vowel sound in the form of monopthong, dipthong and tripthong. O (Onset) Fig 1: Syllable Structure Every syllable will consist of nucleus. Many languages in their syllable part have Onset such as CV, CVC. If the coda is absent such as V, CCV then such syllable are open syllable or free syllable while the syllable which have Coda such as VC, CVC are called closed syllable. Open syllable are found in almost all the languages. 3. MISING LANGUAGE Mising language is a subgroup of Tibeto-Burman group of languages and most of the syllables in MISING language are disyllabic in nature. The basic syllabic structure in MISING language is represented as (C) V(C) where V is a vowel and C is consonant. MISING language is syllable timed or pitched time in nature. Table 1. Syllabic Structure in MISING Language STRUCTURE S (Syllable) N (Nucleus) WORD (DISYLLABIC) R (Rhyme) MEANING V C V alé leg V C V C amír body V C C V émpu large mat C (Coda) 36

V C C V C argom rust CV CV ke:di: mango CV CVC sípund rainwater CVC CV kampo white VCC CVC sormon crocodile Word stress in MISING language is a weak phenomenon. MISING speech is therefore syllable timed. The prominent fact that makes word stress a weak phenomenon is that for every short vowel in MISING language there is a corresponding long vowel. Other fact being that the short vowels are all equal in terms of prominence and the long vowels are relatively more prominent but may or may not be stressed in the normal speech. As against the absence of tone there exists multiple pattern of intonation in MISING language. These patterns vary from dialect to dialect. However, the PAGRO and DELU group of MISING community have homogeneity in pattern of intonations. 4. PITCH CONTOUR Speech being a physical phenomenon is a product of sound produced by the vocal chords and filtering which is done by the tongue, lips and teeth. When the sound is produced by the vocal tract the frequency which is produced is called fundamental frequency and its perception is called PITCH frequency generally denoted by F0 (Fundamental Frequency). Pitch is used in almost many languages in the world to express emotions and other paralingual information. Pierrehumbert s [Pierrehumbert 1980] argued that intonation could be described in terms of two basic tones which is called H(high) and L (low). Pitch accents can be represented as either single or double tone [Pike 1945]. The possible pitch accents are H*, L*, H* +L, H+L*, L+H*. The phrase boundary and boundary tones can be found which are marked with %. 4.1.Tobi Tone and Break Index represent the core prosodic analysis. The break index represents the number that exists between the consecutive words. Prominence is used to indicate the strength of the word, syllable or phrase when used in a sentence. Accent is the intonational phenomenon associated with pitch [Ladd, Mennen and Schepman 2000]. Pitch accented tones are marked using TOBI for every accented syllable. If there is no corresponding pitch accent assignment it would mean that the syllable is not accented [Liberman 1975]. The following four categories of assignment can be done using TOBI model: i. H* - Peak Accent (Upper part of speakers pitch range in phrase) ii. L* - Low Accent (Lowest part of speakers pitch) iii. L* + H Scooped accent (low tone target which is followed by relatively sharp rise to a peak) iv. L + H* - Rising peak accent (High peak target which is immediately preceded by relatively sharp rise) 5. EXPERIMENTAL SETUP In the present analysis some phonetically balanced phonemes of MISING language were selected with the help of language experts in the MISING language from the Centre for Language Studies, Dibrugarh University. The objective is to study the intonation pattern. The data set consist of a block of 29 words of MISING language and 14 vowels. A group of 5 male and 5 female native speaker of age between 20 to 45 years has been considered for recording the data. The recording was done using PRAAT software and PHILIPS head mounted microphone in a small room with minimum background noise. The speech signal was then digitized at 22.5 KHz. The theoretical model used is TOBI a system developed by Beckham and Pierrehumbart for making high and low tone targets in a word. Pitch setting for the recording was as follows: i. For male: floor-75 Hz ceiling-300 Hz ii. For female: floor- 100 Hz ceiling-600 Hz 4.2.Experimental Results Table 2. Tone pattern of MISING words for both male and female utterances Word Structure V-CV VC-VC VC-CV VC-CVC Word Max F0(Hz) Min F0 (Hz) Tone Pattern Alé 155.16 117.59 L Akí 161.85 108.24 L Anyi 151.28 117.67 L Épo 149.12 119.45 L Okum 185.19 125.49 L Apong 176.32 116.77 L Oyan 173.61 122.46 L Amir 157.27 119.52 L Anné 183.88 128.88 L Agro 158.72 108.39 L Amro 153.17 119.66 L Émpo 174.01 126.48 L Argom 171.68 120.41 L Éllung 176.12 132.65 L Otsur 163.33 109.18 L 37

CV-CV CV-CVC CVC-CV CVC-CVC Ogyum 176.52 123.67 L Péki 178.95 130.67 L Péji 178.75 124.26 L Koné 167.67 129.85 L Bíkí 174.97 116.69 L Tapad 159.23 125.42 L Siram 176.83 127.11 L Sorog 160.78 123.18 L Ponta 175 117.88 L Gempa 174.64 116.77 L Kunno 181.30 125.57 L Sarkod 168.37 128.22 L Sormon 181.47 126.65 L Dongkal 129.16 93.58 L 38

Fig 2. Pitch Contour of Few MISING Disyllabic Words 39

Fig. 3. Example of Intonation Pattern of a Typical MISING Sentence 6. RESULTS 1. F0 Contour for vowels in Mising Language shows a rising trend as evident from Fig.2. 2. If the third syllable does not have coda, the stress falls on the first syllable as evident from Fig.2 for words argom, anne, kone, siram, kumno, sormon. 3. It is found that most of the pitch contour in Fig. 2 starts at a little higher position and then falls and then shows a rising pattern. 4. F0 Contour for vowels in Mising Language shows a rising trend as evident from Fig.2. 5. Fig. 3 shows a typical example of intonation pattern of Mising language. The overall intonation pattern is found to consist of local (minimal) rise and falls of pitch across words or syllable, thus the perception of a sentence is dependent on the pitch pattern of the syllable or word and intonation in general. 7. CONCLUSION On the average the tone pattern shown is L. MISING words do not change their meaning with the change in the pitch of the voice. There are multiple patterns of intonation in the MISING language largely due to the absence of tone. Since the vowel sounds in the syllable are not weakened by the giving stress on adjacent syllable, MISING utterances are therefore syllable timed. The data gathered and the analysis performed in the present study will further help in speech synthesis process. 8. REFERENCES [1] Ladd,D.R., Intonation Phonology, Cambridge: Cambidge University Press,1996. [2] Ladd, D.R., Declination reset and the hierarchical 40

organization of utterances, Journal of the Acoustical Society of America,1988. [3] Liberman, M.Y., and Prince, A., On stress and linguistic rhythm, Linguistic Inquiry 8, 1977. [4] Selkirk, E.O., Phonology and Synyax, Cambridge, MA: MIT Press,1984. [5] Pierrehumbert,J.B., The phonology and phonetics of English Intonation, MIT,1980. [6] Pike,K.L., The Intonation of American English, Michigan: University of Michigan(1945) [7] Ladd,D., Mennen,I. and Schepman, A., Phonological conditioning of peak alignment in rising pitch accent in Dutch, The journal of Acoustical Society of America,2000. IJCA TM : www.ijcaonline.org 41