Phonetic Segments Research Articles

Audio-Only Phonetic Segment Classification Using Embeddings Learned From Audio and Ultrasound Tongue Imaging Data

Audio-Only Phonetic Segment Classification Using Embeddings Learned From Audio and Ultrasound Tongue Imaging Data

The relation of velopharyngeal coupling area and vocal tract scaling to identification of stop-nasal cognates.

The purpose of this study was to determine whether the threshold of velopharyngeal (VP) coupling area at which listeners switch from identifying a consonant as a stop to a nasal in North American English was different for speech produced by a model based on an adult male, an adult female, and a 4-year-old child. V1CV2 stimuli were generated with a speech production model that encodes phonetic segments as relative acoustic targets imposed on an underlying vocal tract and laryngeal structure that can be scaled according to sex and age. Each V1CV2 was synthesized with a set of VP coupling functions whose maximum area ranged from 0 to 0.1 cm2. Results showed that scaling the vocal tract and vocal folds had essentially no effect on the VP coupling area at which listener identification shifted from stop to nasal. The range of coupling areas at which the crossover occurred was 0.037-0.049 cm2 for the male model, 0.040-0.055 cm2 for the female model, and 0.039-0.052 cm2 for the 4-year-old child model, and overall mean was 0.044 cm2. Calculations of band limited peak nasalance indicated that 85% peak nasalance during the consonant was well aligned with listener responses.

Slip of the Tongue in English-Language Debates: A Psycholinguistic Study

Debaters need to plan, speak fluently, and think critically during a debate. However, a common error occurs during debates known as slips of the tongue. They are caused by nervousness, tension and unpreparedness. This research aimed to identify types of slips of the tongue and their units of error through a descriptive qualitative study. The data were collected from the video recording of an English Debate Competition at a university. The data were analyzed using two theories by Gleason and Ratner Click or tap here to enter text.and by DardjowidjojoClick or tap here to enter text. to investigate units of error in slips of the tongue. The second theory was presented by Altiparmak and Koruoglu which described nine types of slips of the tongue: shifts, exchanges, anticipations, persistence, additions, deletions, substitutions, blends, and deviations. There were 78 data points found in slips of the tongue - 39 units of error and 39 types of error. The units of error found in slips of the tongue were 11 distinctive feature errors, 9 syllable errors, 6 phonetic segment errors and 14 word errors. Types of the slips of the tongue were 15 deletions, 9 substitutions, 11 blends, and 4 deviations.
 Keywords: Debate, Slips of Tongue, Unit of Error

Analysis of Phonetic Segments of Oesophageal Speech in People Following Total Laryngectomy

This paper presents an approach to extraction techniques for speaker recognition following total laryngectomy surgery. The aim of the research was to develop a pattern of physical features describing the oesophageal speech in people after experiencing laryngeal cancer. Research results may support the speech rehabilitation of laryngectomised patients by improving the quality of oesophageal speech. The main goal of the research was to isolate the physical features of oesophageal speech and to compare their values with the descriptors of physiological speech. Words (in Polish) used during speech rehabilitation were analyzed. Each of these words was divided into phonetic segments from which the physical features of speech were extracted. The values of the acquired speech descriptors were then used to create a vector of the physical features of oesophageal speech. A set of these features will determine a model that should allow us to recognize whether the speech-rehabilitation process is proceeding correctly and also provide a selection of bespoke procedures that we could introduce to each patient. This research is a continuation of the analysis of oesophageal speech published previously. This time, the effectiveness of parameterization was tested using methodologies for analyzing the phonetic segments of each word.

Self-Supervised Learning With Segmental Masking for Speech Representation

Self-supervised learning has achieved remarkable success for learning speech representations from unlabeled data. The masking strategy plays an important role in the self-supervised learning algorithm. Most of the masking techniques operate at a frame level. In linguistics, phone is the smallest unit of sound. Hence, we believe that a masking technique that operates at a phoneme level will effectively encode the phonotactic and prosodic constraints of a spoken language, thus eventually benefits the downstream speech recognition tasks. In this work, we explore a novel segmental masking strategy. Specifically, we mask phonetically motivated speech segments according to the phonetic segmentation in an utterance. By doing so, we implicitly incorporate the properties of a spoken language, such as phonotactic constraints and duration of phonetic segments, into the pre-training. Through extensive experiments, we confirm that the segmental masking strategy consistently outperforms the frame-based masking counterpart. We also further investigate the effect of segmental masking unit size, i.e. phoneme, phoneme span, and lexical word. This work presents an important finding about masking strategy in speech representation learning.

Intelligibility of synthetic words generated by transformation of a sequence of discrete acoustic events into modulation of the vocal tract shape

Within the paradigm of a recent model of speech production [Story and Bunton, JASA 146(4), 2522–2528], an utterance is specified as a sequence of relative acoustic events along a time axis. These events consist of directional changes of the vocal tract resonance frequencies called resonance deflection patterns (RDPs) that, when associated with a temporal event function, are transformed via acoustic sensitivity functions, into time-varying modulations of the vocal tract shape. RDPs specifying the targeted directional shift of the first three resonances for bilabial, alveolar, and velar consonants would be coded as [−1 −1 −1], [−1 1 1], and [−1 1 −1], respectively. In this study, these RDPs were combined with four vowels (“ih, ae, eh, uh”) to construct a set of 40 American English words (CVCs). A word intelligibility test was conducted in which listeners heard a synthesized target word and were asked to indicate what they heard by choosing a word from a matrix that included the target and seven near-neighbor words. Results indicate listener word recognition was aligned with the RDP settings, suggesting that they are an effective discrete representation of phonetic segments that can be transformed into speech by modulation of the vocal tract shape.

Does Automatic Speech Recognition (ASR) Have a Role in the Transcription of Indistinct Covert Recordings for Forensic Purposes?

The transcription of covert recordings used as evidence in court is a huge issue for forensic linguistics. Covert recordings are typically made under conditions in which the device needs to be hidden, and so the resulting speech is generally indistinct, with overlapping voices and background noise, and in many cases the acoustic record cannot be analyzed via conventional phonetic techniques (i.e. phonetic segments are unclear, or there are no cues at all present acoustically). In the case of indistinct audio, the resulting transcripts that are produced, often by police working on the case, are often questionable and despite their unreliable nature can be provided as evidence in court. Injustices can, and have, occurred. Given the growing performance of automatic speech recognition (ASR) technologies, and growing reliance on such technologies in everyday life, a common question asked, especially by lawyers and other legal professionals, is whether ASR can solve the problem of what was said in indistinct forensic audio, and this is the main focus of the current paper. The paper also looks at forced alignment, a way of automatically aligning an existing transcriptions to audio. This is an area that needs to be explored in the context of forensic linguistics because transcripts can technically be “aligned” with any audio, making it seem as if it is “correct” even if it is not. The aim of this research is to demonstrate how automatic transcription systems fare using forensic-like audio, and with more than one system. Forensic-like audio is most appropriate for research, because there is greater certainty with what the speech material consists of (unlike in forensic situations where it cannot be verified). Examples of how various ASR systems cope with indistinct audio are shown, highlighting that when a good-quality recording is used ASR systems cope well, with the resulting transcript being usable and, for the most part, accurate. When a poor-quality, forensic-like recording is used, on the other hand, the resulting transcript is effectively unusable, with numerous errors and very few words recognized (and in some cases, no words recognized). The paper also demonstrates some of the problems that arise when forced-alignment is used with indistinct forensic-like audio—the transcript is simply “forced” onto an audio signal giving completely wrong alignment. This research shows that the way things currently stand, computational methods are not suitable for solving the issue of transcription of indistinct forensic audio for a range of reasons. Such systems cannot transcribe what was said in indistinct covert recordings, nor can they determine who uttered the words and phrases in such recordings, nor prove that a transcript is “right” (or wrong). These systems can indeed be used advantageously in research, and for various other purposes, and the reasons they do not work for forensic transcription stems from the nature of the recording conditions, as well as the nature of the forensic context.

Foreign accent strength and intelligibility at the segmental level

The relationship between strength of foreign accent and intelligibility is not straightforward. This relationship resists a simple characterisation due in part to the multiplicity of cues that carry accent in the word- and sentence-level materials typically used in the study of accent. One of the principal conveyors of accent is the phonetic segment. The current study attempts to isolate this segmental contribution to foreign accent and consequently measure the relationship between segmental accent and intelligibility for listeners with differing linguistic correspondence to the target and accented language. English, Spanish and Czech listeners identified English words in which the initial consonant was either intact, or had been replaced by a Spanish-accented counterpart; in a second task, they rated the accent strength of the same tokens. All speech material was produced by an English-Spanish bilingual talker. Overall, Spanish listeners displayed a smaller loss of intelligibility due to the accented segment than native English listeners, while the Czech cohort experienced the largest intelligibility loss. However, the relationship between accent strength and intelligibility loss was not linear, varying with phoneme identity and its role in a listener's first language. These findings suggest that how accented and intelligible a sound is depends strongly on the interactions between the phonological systems of speakers and listeners.

ANALISIS SENYAPAN DAN KILIR LIDAH PADA PRODUKSI UJARAN “BELIBETNYA ISYANA SARASVATI”

This study explains about the Pause and slip of tongue on the production of Isyana Sarasvati utterances in several video uploads on Youtube, one of which is on uploading a 20detik youtube video channel titled Isyana Sarasvati Bongkar Aib Sendiri, as well as several other videos as additional data. This analysis aims to describe the forms of the tongue's tongue and tongue and the cause of the tongue's tongue and tongue's speech that is contained in Isyana Sarasvati's speech. The analysis uses psycholingusitic studies that language is related to the mental processes that humans go through in forming an utterance. The method of data collection is done by using note- taking techniques, namely by listening to Isyana Sarasvati's utterances to be recorded, recorded and transcribed into writing. While the data analysis method uses descriptive qualitative, with tanskrispsi utterance as primary data, while books and journals as secondary data, with analysis using psycholinguistic theory from Noam Chomsky namely performance in conversation. From the analysis, it was found that silent due to breathing, and silent due to doubt. Tongue can also occur as a result of talking in a hurry, unintentionally and not concentrating. The units in the gloss of the tongue are errors in phonetic segments, syllables, and word errors.

The relation of velopharyngeal coupling area to the identification of stop versus nasal consonants in North American English based on speech generated by acoustically driven vocal tract modulations

The purpose of this study was to determine the threshold of velopharyngeal coupling area at which listeners switch from identifying a consonant as a stop to a nasal in North American English, based on V1CV2 stimuli generated with a speech production model that encodes phonetic segments as relative acoustic targets. Each V1CV2 was synthesized with a set of velopharyngeal coupling functions whose area ranged from 0 to 0.1 cm2. Results show that consonants were identified by listeners as a stop when the coupling area was less than 0.035-0.057 cm2, depending on place of articulation and final vowel. The smallest coupling area (0.035 cm2) at which the stop-to-nasal switch occurred was found for an alveolar consonant in the /ɑCi/ context, whereas the largest (0.057 cm2) was for a bilabial in /ɑCɑ/. For each stimulus, the balance of oral versus nasal acoustic energy was characterized by the peak nasalance during the consonant. Stimuli with peak nasalance below 40% were mostly identified by listeners as stops, whereas those above 40% were identified as nasals. This study was intended to be a precursor to further investigations using the same model but scaled to represent the developing speech production system of male and female talkers.

INTERFERENCE OF STUDENTS' NATIVE LANGUAGE IN THEIR ORAL PRODUCTION OF ENGLISH

Abstract: This research aims to describe the phonological interference in the students' English pronunciation. This research is descriptively designed. The subjects are the students with the Banjarese background. The instruments used in this research are questionnaires and oral test. Data were contrastively analyzed. The research showed the students’ pronunciation problems due to the use of Banjarese phonology. There are 23 types of phonological interference found in the English pronunciation. The factors of the linguistic phenomena are: (1) the existence of a given sound in English, but absence from the Banjarese. The Banjarese (L-1) has vowels: /a/, /u/, /i/, /e/, and /o/, whereas English (L-2) has vowels: /I/, /e/, /æ/, /ʌ/, /ɒ/, /ʊ/, /ə/ /i:/, /ɜ:/, /ɑ:/, /ɔ:/, /u:/. The L-1 has three diphthongs, namely: /ai/, /au/, and /ui/, while the L-2 has eight diphthongs, namely: /Iə/, /oʊ/, /aɪ/, /aʊ/, /ɔɪ/, /ɪə/, /ʊə/, and /ɔə/. The L-1 has the consonants: /b/, /d/, /g/, /p/, /t/, /k/, /c/, /j/, /s/, /h/, /m/, /n/, /ny/, /ŋ/, /l/, /r/, /w/, and /y/, while the L-2 has the following consonants: /b/, /d/, /g/, /p/, /t/, /k/, /f/, /v/, /ɵ/, /ð/, /s/, /z/, /ʃ/, /ʒ/, /tʃ/, /dʒ/, /m/, /n/, /ŋ/, /l/, /r/, /w/, /j/ and /h/. Second, both languages have the same phonetic segments but differ in their distributions. Third, both languages have voiced stop consonants /p, k, and t/, but they are differently realized in a particular position. One of the phonological rules is what is called ‘aspiration. Thus, the students need to more practice in oral production.

Dynamic Formant Trajectories in German Read Speech: Impact of Predictability and Prominence

Phonetic structures expand temporally and spectrally when they are difficult to predict from their context. To some extent, effects of predictability are modulated by prosodic structure. So far, studies on the impact of contextual predictability and prosody on phonetic structures have neglected the dynamic nature of the speech signal. This study investigates the impact of predictability and prominence on the dynamic structure of the first and second formants of German vowels. We expect to find differences in the formant movements between vowels standing in different predictability contexts and a modulation of this effect by prominence. First and second formant values are extracted from a large German corpus. Formant trajectories of peripheral vowels are modeled using generalized additive mixed models, which estimate nonlinear regressions between a dependent variable and predictors. Contextual predictability is measured as biphone and triphone surprisal based on a statistical German language model. We test for the effects of the information-theoretic measures surprisal and word frequency, as well as prominence, on formant movement, while controlling for vowel phonemes and duration. Primary lexical stress and vowel phonemes are significant predictors of first and second formant trajectory shape. We replicate previous findings that vowels are more dispersed in stressed syllables than in unstressed syllables. The interaction of stress and surprisal explains formant movement: unstressed vowels show more variability in their formant trajectory shape at different surprisal levels than stressed vowels. This work shows that effects of contextual predictability on fine phonetic detail can be observed not only in pointwise measures but also in dynamic features of phonetic segments.

STYLISTIC FEA YLISTIC FEATURES OF IN TURES OF INTONATION

Abstract. It is known that descriptive phonetics refers to phonetic phenomena such as speech sounds, their formation, types, changes, syllables, intonation, pause, stress, and their peculiarities. These phenomena recorded in the process of speech combine into a single system and form a speech act. This article discusses and comparatively analyzes stylistic features of intonation in the English and Uzbek languages. Background. Therefore, in the analysis and study of such phenomena, it is necessary to study them separately, and to keep in mind that these phenomena require one another, in other words, units that cannot exist without the other. Particularly, vowel sounds, which are stressed, form a syllable, and words are formed from the syllables, a sentence is formed from words, and each word and sentence is pronounced with a specific intonation and pause. In the formation of such a speech chain, the phonetic segment (main part) and supersegment (additional parts) elements are almost identically involved. But they also have some peculiarities while using them

"The Diagnostic Evaluation of Switchboard-corpus Automatic Speech Recognition Systems"

To see whether the related mistake patterns can be linked to a particular set of variables, a Eight Control equipment recognizing (and six forcedalignment) algorithms were evaluated for clinical diagnosis. Each recognizing service's result was converted to a standardized way and evaluated to a comparative record made from pronunciations labelled data (which included 54 minutes of information from several hundred speakers). A job evaluation was used to relate a combination of acoustic, morphological, etc. speaker attributes to acknowledgment occurrences throughout this reference data. The decision trees show that correct categorization of phonetic segments and characteristics is one of the most constant variables linked with better recognition performance. These findings indicate that enhancing the pronouncing modelling used in verbal pairing, including the acoustic modeling techniques utilized for morphological classification, might improve future-generation recognition systems.

Development of Minimal Pair Test in Tamil (MPT-T)

Introduction: Speech perception testing provides an accurate measurement of the child’s ability to perceive and distinguish the various phonetic segments and patterns of the sounds. From among the many types of speech stimuli used, minimal pairs can also be used to assess the phoneme recognition skills. Thus, the study focused on developing Minimal Pair Test in Tamil (MPT-T). Aim: The aim of the present study was to develop and validate the MPT in Tamil on Normal Hearing (NH) children and paediatric cochlear implantees (CI). Materials and Methods: It was an experimental study which included school going children in the age range of six to eight years and the duration of the study was 12 months. The test was developed in two phases. The first phase focussed on the construction of the word list, recording of the word pairs and the preparation of the test. The second phase was administration of the test on NH children and paediatric cochlear implantees. The test scores were analysed using Mann Whitney U test, Kruskal Wallis and Wilcoxon signed-rank test. The results showed a statistical significance between the NH group and the paediatric cochlear implantees. Results: The present study included 40 NH children and 15 paediatric cochlear implantees through purposive sampling method. The specific speech feature analysis of the paediatric cochlear implantees revealed that there was difficulty identifying the word pairs differing in Vowel Length (VL) and the best performed feature was Place of Articulation (POA). The results showed statistical significance between the NH group and the paediatric cochlear implantees. Conclusion: The developed test can be effectively used in clinic for assessing speech perception abilities of pediatric Cochlear Implantees and also in planning the rehabilitative goals.

Translating Fieldwork into Datasets: The Development of a Corpus for the Quantitative Investigation of Grammatical Phenomena in Eibela

This extended abstract details the process of constructing an annotated XML corpus suitable for quantitative analysis of morphosyntactic and phonetic phenomena in the Eibela language of Papua New Guinea. Preliminary results will also be included, which investigate the semantic, phonetic, and discourse correlates of argument realization. The goal of this paper is to illustrate how legacy materials can be enriched and investigated using computational methodologies including forced alignment of phonetic segments using bulk processing of data in Python and R, the Montreal Forced Aligner (MFA), and morphosyntactic annotation developed as part of the Multilingual Corpus of Annotated Spoken Texts (Multi-CAST).

Influence of the Covering Area of Major Connectors of Mandibular Dentures on the Accuracy of Speech Production: A Pilot Study

Background/Objectives: Speech impairment during the initial phase of removable partial denture (RPD) treatment can prevent patient adaptation to RPDs. This quasi-experimental study aimed to investigate the influence of the covering area of major connectors located in the mandibular anterior region on the accuracy of speech production. Methods: Fifteen adults (12 men and 3 women aged 26–40 years) with healthy dentition were recruited. Four target syllables were used for the assessments: [ʃi], [ʧi], [çi], and [ki]. There were four speaking conditions: without a connector (control), wearing lingual bar (5 mm width), wearing lingual plates (12 mm width), and wearing connectors covering the occlusal surface of lower dentition (16 mm width; CO). The accuracy of speech production was evaluated by mean appearance ratio of correct labels (MARC) and incorrect labels (MARIC) of the phonetic segments which were calculated with a speech evaluation system (VoiceAnalyzer). Results: The MARC in the 3 target syllables [ʃi], [ʧi], and [ki] significantly decreased with CO. For the target syllable [ʃi], the MARICs of [ʧi], [ʤi], and [çi] with CO were significantly lower than that in the control. For the target syllable [ʧi], the MARIC of [ki] with CO was significantly higher than that in the control. For the target syllable [çi], the MARIC of [gi] with CO was significantly higher than that in the control. Conclusion: RPDs with major connectors covering the lower dentition could disturb speech production during the initial phase of RPD treatment. Prosthodontic treatment with such RPD designs requires further consideration regarding the information given to patients before starting the treatment.

Perceptual biases in consonant perception—Where do we find them? What is behind them?

There is no doubt that language experience has an early and profound impact on the perception of phonetic segments. There is now clear evidence that phonetic perception is also shaped by universal perceptual biases that can be exposed as directional asymmetries in phonetic discrimination. Much of this work has focused on vowel perception in which we observe robust perceptual asymmetries in infant and adult L2 perception. To explain these patterns, Polka and Bohn (2003) and (2011) outlined the Natural Referent Vowel (NRV) framework which proposes that vowel perception is shaped by both generic (universal) and language-specific processing. In this talk, I will present data showing perceptual biases in consonant perception which suggest that we can extend NRV principles to consonant perception.There is no doubt that language experience has an early and profound impact on the perception of phonetic segments. There is now clear evidence that phonetic perception is also shaped by universal perceptual biases that can be exposed as directional asymmetries in phonetic discrimination. Much of this work has focused on vowel perception in which we observe robust perceptual asymmetries in infant and adult L2 perception. To explain these patterns, Polka and Bohn (2003) and (2011) outlined the Natural Referent Vowel (NRV) framework which proposes that vowel perception is shaped by both generic (universal) and language-specific processing. In this talk, I will present data showing perceptual biases in consonant perception which suggest that we can extend NRV principles to consonant perception.

The role of input in native Spanish Late learners’ production and perception of English phonetic segments

Abstract This study evaluated the effect of input variation on the production and perception of English phonetic segments by native Spanish adults who had immigrated to the United States after the age of 16 years. The native Spanish (NS) participants were assigned to three groups of 20 each according to years of English input (years of U.S. residence multiplied by percent English use outside the home). Experiment 1 assessed the perceived relation between English and Spanish vowels. It yielded similar results for the NS groups designated “Low input” (M = 0.2 years of input), “Mid” (M = 1.2 years) and “High” (M = 3.0 years). Experiments 2–4 examined English vowel discrimination, vowel production and consonant discrimination. Apart from a modest improvement in vowel discrimination, these experiments showed little improvement as years of English input increased. One possible explanation for the essentially null finding of this study is that input matters little or not at all when an L2 is learned naturalistically following the closure of a critical period. Another possibility is that adequate native speaker input is crucial for L2 speech learning but the input differences evaluated here were insufficient to yield measurable improvements in performance. We conclude the article by illustrating a new technique that might be used to choose between these competing explanations.

Phonetic Segments and the Organization of Speech

According to mainstream linguistic phonetics, speech can be modeled as a string of discrete sound segments or “phones” drawn from a universal phonetic inventory. Recent work has argued that a mature phonetics should refrain from theorizing about speech and speech processing using sound segments and that the phone concept should be eliminated from linguistic theory. The article lays out the tenets of the phone methodology and evaluates its prospects in light of the eliminativist arguments. I claim that the eliminativist arguments fail to show that the phone concept should be eliminated from linguistic theory.