Place Of Articulation Research Articles

A Study on the Impact of Lombard Effect on Recognition of Hindi Syllabic Units Using CNN Based Multimodal ASR Systems

Research work on the design of robust multimodal speech recognition systems making use of acoustic and visual cues, extracted using the relatively noise robust alternate speech sensors is gaining interest in recent times among the speech processing research fraternity. The primary objective of this work is to study the exclusive influence of Lombard effect on the automatic recognition of the confusable syllabic consonant-vowel units of Hindi language, as a step towards building robust multimodal ASR systems in adverse environments in the context of Indian languages which are syllabic in nature. The dataset for this work comprises the confusable 145 consonant-vowel (CV) syllabic units of Hindi language recorded simultaneously using three modalities that capture the acoustic and visual speech cues, namely normal acoustic microphone (NM), throat microphone (TM) and a camera that captures the associated lip movements. The Lombard effect is induced by feeding crowd noise into the speaker’s headphone while recording. Convolutional Neural Network (CNN) models are built to categorise the CV units based on their place of articulation (POA), manner of articulation (MOA), and vowels (under clean and Lombard conditions). For validation purpose, corresponding Hidden Markov Models (HMM) are also built and tested. Unimodal Automatic Speech Recognition (ASR) systems built using each of the three speech cues from Lombard speech show a loss in recognition of MOA and vowels while POA gets a boost in all the systems due to Lombard effect. Combining the three complimentary speech cues to build bimodal and trimodal ASR systems shows that the recognition loss due to Lombard effect for MOA and vowels reduces compared to the unimodal systems, while the POA recognition is still better due to Lombard effect. A bimodal system is proposed using only alternate acoustic and visual cues which gives a better discrimination of the place and manner of articulation than even standard ASR system. Among the multimodal ASR systems studied, the proposed trimodal system based on Lombard speech gives the best recognition accuracy of 98%, 95%, and 76% for the vowels, MOA and POA, respectively, with an average improvement of 36% over the unimodal ASR systems and 9% improvement over the bimodal ASR systems.

Phonetic variation in Italian L2: An acoustic analysis of sibilant fricatives in the speech of L1 Spanish learners


 
 The present study investigates the acoustic features of Italian sibilant fricatives (/s/, /z/ and /ʃ/) in the speech of L1 Spanish learners. Segmental duration, degree of voicing and place of articulation of learners’ productions are analysed alongside those of a control group of L1 Italian speakers to investigate the fine-grained phonetic differences between native and non-native pronunciation. Results from a quantitative analysis suggest that factors such as the degree of typological markedness of the phoneme and the influence of the L1 in perception and production affect L2 pronunciation to a different extent for each of the target sounds.
 

Theory Of English Constant In Phonology

The theory of consonants in English phonology is an important aspect in understanding the sound system in that language. Phonology is the study of how the sounds in language are used to form meaning. In English, consonants and vowels have different roles in forming word structure and sound articulation.Consonants in English are defined as sounds that experience resistance or airflow shift when spoken. There are various types of consonants in English, such as hissing consonants, popping consonants, sliding consonants, and so on. These consonants have certain characteristics, such as the place and manner of articulation, which distinguish one consonant from another.Places of articulation refer to the parts in the mouth or throat that are used to produce consonant sounds. For example, the lips, teeth, tongue, palate, and throat can play a role in the production of consonants. For example, the consonant /p/ is pronounced by bringing the lips together, while the consonant /k/ is pronounced by moving the back palate with the back of the tongue.The manner of articulation refers to the type of resistance or shift in airflow that occurs when pronouncing the consonant. Some examples of how consonants are articulated in English include complete closure of the airflow (eg /p/), partial obstruction of the airflow (eg /s/), or passage of air through the nose (eg /m/). The theory of consonants and vowels in English phonology helps to explain how these sounds form a distinctive language system. Knowledge of the place and manner of consonant articulation, as well as the position of the tongue and the tilt of the vowel lips, allows linguistics to analyze the differences in sounds in English. In the study of phonology, this theory also helps to understand sound patterns, sound changes, and the placement of sounds in words as well as in the broader structure of language.

Aging and sex effects on phoneme perception: An exploratory mismatch negativity and P300 investigation

Background & aimsThe mismatch negativity (MMN) and P300 event-related potentials (ERPs) have been studied in relation to phoneme discrimination and categorization, respectively. Although the effects of aging and sex on pure-tone perception have been widely investigated using these ERPs, evidence relating to phoneme perception is scarce. The current study aimed to provide insight into the effects of aging and sex on phoneme discrimination and categorization, as measured through the MMN and P300. MethodAn inattentive and attentive oddball paradigm containing a phonemic articulation place contrast were administered during EEG registration in sixty healthy individuals (thirty males and females), of which an equal number of young (20–39 years), middle-aged (40–59 years) and elderly (60+ years) subjects were included. The amplitude, onset latency and topographical distribution of the MMN and P300 effect, as well as the amplitude of the P1-N1-P2 complex, were analyzed for age group and sex differences. ResultsWith respect to aging, elderly subjects demonstrated a reduced MMN and P300 amplitude compared to the young group, whereas the scalp distribution of both components was unaffected. No aging effects on the P1-N1-P2 complex were found. In elderly individuals, the P300 was found to be delayed compared to the young group, while no such effect on MMN latency could be observed. No differences in MMN and P300 measures could be identified between males and females. ConclusionDifferential effects of aging were found on the MMN and P300, specifically in terms of latency, in relation to phoneme perception. In contrast, sex was found to scarcely affect both processes.

Phonetic variation in Italian L2: An acoustic analysis of sibilant fricatives in the speech of L1 Spanish learners

The present study investigates the acoustic features of Italian sibilant fricatives (/s/, /z/ and /ʃ/) in the speech of L1 Spanish learners. Segmental duration, degree of voicing and place of articulation of learners’ productions are analysed alongside those of a control group of L1 Italian speakers to investigate the fine-grained phonetic differences between native and non-native pronunciation. Results from a quantitative analysis suggest that factors such as the degree of typological markedness of the phoneme and the influence of the L1 in perception and production affect L2 pronunciation to a different extent for each of the target sounds.

The association between longitudinal declines in speech sound accuracy and speech intelligibility in speakers with amyotrophic lateral sclerosis

ABSTRACT The purpose of this study was to examine how neurodegeneration secondary to amyotrophic lateral sclerosis (ALS) impacts speech sound accuracy over time and how speech sound accuracy, in turn, is related to speech intelligibility. Twenty-one participants with ALS read the Bamboo Passage over multiple data collection sessions across several months. Phonemic and orthographic transcriptions were completed for all speech samples. The percentage of phonemes accurately produced was calculated across each phoneme, sound class (i.e. consonants versus vowels), and distinctive feature (i.e. features involved in Manner of Articulation, Place of Articulation, Laryngeal Voicing, Tongue Height, and Tongue Advancement). Intelligibility was determined by calculating the percentage of words correctly transcribed orthographically by naive listeners. Linear mixed effects models were conducted to assess the decline of each distinctive feature over time and its impact on intelligibility. The results demonstrated that overall phonemic production accuracy had a nonlinear relationship with speech intelligibility and that a subset of features (i.e. those dependent on precise lingual and labial constriction and/or extensive lingual and labial movement) were more important for intelligibility and were more impacted over time than other features. Furthermore, findings revealed that consonants were more strongly associated with intelligibility than vowels, but consonants did not significantly differ from vowels in their decline over time. These findings have the potential to (1) strengthen mechanistic understanding of the physiological constraints imposed by neuronal degeneration on speech production and (2) inform the timing and selection of treatment and assessment targets for individuals with ALS.

Acoustic and Perceptual Categorization of Sibilants for Mandarin Children With Ankyloglossia.

To deepen our understanding of the effects of ankyloglossia on articulation, the purpose of this study is to examine consonant production and the perceived accuracy of Mandarin-speaking children with ankyloglossia. Ten tongue-tie (TT) and 10 typically developing (TD) children produced nine Mandarin sibilants that contrast in three places of articulation. Their speech productions were analyzed from six acoustic measurements. To further examine the perceptual consequences, an auditory transcription task (N = 21) was conducted. The acoustic analyses discovered that the TT children failed to distinguish the three-way place contrast and produced significant acoustic deviations compared with the TD peers. The perceptual transcription results found that TT children's production was significantly misidentified, suggesting severely affected intelligibility. The preliminary findings provide strong support for a correlation between ankyloglossia and distorted speech signals and suggest important interactions between sound errors and linguistic experience. We also propose that ankyloglossia should not be a purely appearance-based diagnosis and that speech production is a crucial index of tongue function in clinical decision making and monitoring.

ASYMMETRY IN THE SIMPLIFICATION OF REVERSED SONORITY CLUSTERS IN (A)TYPICAL PHONOLOGICAL DEVELOPMENT: EVIDENCE FROM GREEK

This paper explores the asymmetry in the disordered (atypical) Greek L1 phonological productions of a child with Developmental Language Disorder (DLD) compared to the productions of two non-disordered (typical) children. The study focuses on the simplification patterns of reversed sonority consonantal sequences, namely [s/Fricative+Stop] and [s+Fricative]. The data show that, while the non-disordered children uniformly simplify reversed sonority sequences, reducing them to the less sonorous consonant, the child with DLD applies two different simplification patterns, resulting from constraints that disallow featural Markedness. We propose that the asymmetry attested in the reduction is due to the employment of two distinct grammars by the two groups of children. The typically developing children employ the, widely attested cross-linguistically, sonority-driven reduction. Meanwhile, the grammar of the child with DLD is not motivated by sonority, but rather, by a general avoidance for Markedness, retaining the unmarked [-continuant] Manner of Articulation in [s/Fricative+Stop] sequences, while favoring the consonant with unmarked Place of Articulation in [s+Fricative] sequences, where the Manner of Articulation of both segments is marked, i.e. [+continuant].

The evolution of similarity avoidance: a phylogenetic approach to phonotactic change

The cross-linguistic under-representation of adjacent consonants sharing a place of articulation within uninflected lexical items is well documented. At the same time, little is known regarding the specific diachronic mechanisms involved in the emergence and maintenance of this pattern. Phylogenetic analyses provide some support for the idea that adjacent identical consonants within words arise infrequently, but stronger support for the idea that words containing such a pattern die out more frequently than those without. I highlight the value of tools used in this paper for exploring the evolution of sound patterns, and also discuss some limitations of the implementation used in the paper to be improved upon.

Knowledge-Based Features for Speech Analysis and Classification: Pronunciation Diagnoses

Accurate pronunciation of speech sounds is essential in communication. As children learn their native language, they refine the movements necessary for intelligible speech. While there is variability in the order of acquisition of speech sounds, there are some sounds that are more complex and are later developing. The rhotic /r/ is a later-developing sound in English, and some children require intervention to achieve accurate production. Additionally, individuals learning English as a second language may have difficulty learning accurate /r/ production, especially if their native language does not have an /r/, or the /r/ they produce is at a different place of articulation. The goal of this research is to provide a novel approach on how a knowledge-based intelligence program can provide immediate feedback on the accuracy of productions. In the proposed approach, the audio signals will first be detected, after which features of audio signals will be extracted, and finally, knowledge-based intelligent classification will be performed. Based on the obtained knowledge and application scenarios, novel features are proposed and used to classify various speaker scenarios.

Voice Onset Time in a language without voicing contrast: An acoustic analysis of Blackfoot oral stops

This paper presents an acoustic analysis of Voice Onset Time (VOT) in oral stop consonants in Blackfoot, an Algonquian language without contrastive voicing. We focus on VOT as one of the key temporal acoustic correlates of voicing and investigate VOT variation in relation to (i) place of articulation (labial vs. alveolar vs. velar); (ii) length (long vs. short), quality (/a/ vs. /o/), and accent pattern (accented vs. unaccented) of the following vowel; (iii) word position (initial vs. medial); (iv) gender; and (v) age. We analyzed 2096 stop consonant tokens produced by 13 participants, who completed two different tasks: an English-to-Blackfoot translation task and a picture naming task. The key findings are as follows: (i) Blackfoot stop consonants fall into the short-lag range that overlaps with the English voiced category, with mean VOT values ranging from 11.6–32.7 ms; (ii) VOT values become progressively longer as the place of articulation moves to more posterior positions; (iii) VOT values are longer before high vowels than before low vowels; (iv) VOT values are longer before long vowels than before short vowels; (v) no statistically significant effect was found for the linguistic factors word position and accent pattern; (vi) no statistically significant effect was found for the socio-indexical factors age and gender; and (vii) no statistically significant effect was found for the experimental factor task type. The implications of our findings for the Blackfoot writing system and ongoing work on language documentation and revitalization are briefly discussed.

Quantitative Acoustic versus Deep Learning Metrics of Lenition

Spanish voiced stops /b, d, ɡ/ surfaced as fricatives [β, ð, ɣ] in intervocalic position due to a phonological process known as spirantization or, more broadly, lenition. However, conditioned by various factors such as stress, place of articulation, flanking vowel quality, and speaking rate, phonetic studies reveal a great deal of variation and gradience of these surface forms, ranging from fricative-like to approximant-like [β⊤, ð⊤, ɣ⊤]. Several acoustic measurements have been used to quantify the degree of lenition, but none is standard. In this study, the posterior probabilities of sonorant and continuant phonological features in a corpus of Argentinian Spanish estimated by a deep learning Phonet model as measures of lenition were compared to traditional acoustic measurements of intensity, duration, and periodicity. When evaluated against known lenition factors: stress, place of articulation, surrounding vowel quality, word status, and speaking rate, the results show that sonorant and continuant posterior probabilities predict lenition patterns that are similar to those predicted by relative acoustic intensity measures and are in the direction expected by the effort-based view of lenition and previous findings. These results suggest that Phonet is a reliable alternative or additional approach to investigate the degree of lenition.

The Euclidean metrics applied to the interphonemic distance measurements

The similarity or dissimilarity of the spoken words is generally rendered by intuition, depending on the personal orientation or the personal traits of a listener/speaker. The existing methods of phonetic encoding of words suffer from a number of shortcomings, the main one being the impossibility of weighing spoken words in quantitative terms. Moreover, the existing methods may be related to a certain language or language family. The algorithm advanced in the present paper compares the characteristics of different phonemes that make up a word. The paper treats phonemic frequency and sonority as elements common both for consonants and vowels, backness and openness, as features pertaining to vowels, and the place of articulation pertaining to consonants only. The algorithm in question permits to compare in quantitative terms the words of different length, whether formed by open or closed syllables. The inter-phonemic distances are calculated by employing Euclidean metrics. The paper suggests fields of application of the method treated in the paper: this scheme can be applied in the fields of comparative linguistics, in medicine, when the hearing disorders are scrutinized, as well as in the brain cortex mapping.

The McGurk Illusion: A Default Mechanism of the Auditory System

Recent studies have questioned past conclusions regarding the mechanisms of the McGurk illusion, especially how McGurk susceptibility might inform our understanding of audiovisual (AV) integration. We previously proposed that the McGurk illusion is likely attributable to a default mechanism, whereby either the visual system, auditory system, or both default to specific phonemes-those implicated in the McGurk illusion. We hypothesized that the default mechanism occurs because visual stimuli with an indiscernible place of articulation (like those traditionally used in the McGurk illusion) lead to an ambiguous perceptual environment and thus a failure in AV integration. In the current study, we tested the default hypothesis as it pertains to the auditory system. Participants performed two tasks. One task was a typical McGurk illusion task, in which individuals listened to auditory-/ba/ paired with visual-/ga/ and judged what they heard. The second task was an auditory-only task, in which individuals transcribed trisyllabic words with a phoneme replaced by silence. We found that individuals' transcription of missing phonemes often defaulted to '/d/t/th/', the same phonemes often experienced during the McGurk illusion. Importantly, individuals' default rate was positively correlated with their McGurk rate. We conclude that the McGurk illusion arises when people fail to integrate visual percepts with auditory percepts, due to visual ambiguity, thus leading the auditory system to default to phonemes often implicated in the McGurk illusion.

Sonority sequencing and its relationship to articulatory timing in Georgian

This work asks how the syllable as a unit is delimited in space and time. To do this we bring together two theoretical approaches to the syllable: a sonority-based approach which emphasizes spatial organization, and the coupled oscillator model (Nam, Goldstein & Saltzman 2009) which emphasizes temporal organization. Many languages present challenges to these theories, and here we focus on Georgian, which is problematic for both approaches, as Georgian permits onset clusters of up to seven consonants and of any sonority shape. Specifically, this study examines the relationship between gestural overlap and sonority shape in Georgian via electromagnetic articulography. Drawing on data from three speakers, we examine gestural overlap in two-consonant onsets that vary in sonority shape and order of place of articulation. Using two measures of gestural overlap, we find (i) long lag between consonant gestures, which we suggest is language-specific and (ii) that lag is largest in sonority rises and smallest in sonority falls. These results suggest that neither phonemic recoverability nor a hierarchical effect of increasingly open constrictions is the primary motivator behind inter-consonantal timing. Instead, the high degree of overlap in sonority falls ensures their tautosyllabic parse, which would otherwise be threatened by intrusive vocoids. Thus, we argue that recoverability of the syllable as a unit is the major motivating factor behind the timing patterns observed, and that syllables emerge not solely from either spatial or temporal properties, but from the interplay of the two.

The Historical Changes of /k/ and /q/ in Najdi Arabic: A Phonological Analysis

This paper seeks to explore the potential phonological reasons for the historical change of /k/ and /q/ in Najdi Arabic and the influence of these two consonants on the frequency and duration of vowels, such as /i/ sound through the prism of acoustic analysis. The /k/ phoneme transforms into [ts] and /q/ transforms into [ɡ] and [dz] in many Najdi Arabic words. This transformational process is called ‘palatalization’. Although it is a universal phenomenon, palatalization in Najdi Arabic has specific characteristic features. The main reason for this specificity is that the Najdi dialect had not been in contact with any other foreign languages in Najd region (the region of Najdi speakers) to create the affrication. This historical change is thus attributable to phonological reasons motivated by specific phonological features. Data were extracted from two recordings by two native speakers of the Najdi dialect who produced 217 words. They were phonologically transcribed and analyzed by the researcher to understand the motivation for the phonological change of these two consonants. Additionally, Pratt software was used to analyze the data to explore the influence of consonants on the /i/ sound frequency and duration. The results revealed that the palatalization of [ts] and [dz] does not occur in all the Najdi Arabic words, but only in some due to specific phonological features. The phonological patterns of /k/ and /q/ rely on the place of vocalic articulation preceding and following them.

모음 스펙트럼에 기반한 전후 비자음 조음위치 판별*

This study examined the utility of the acoustic features of vowels as cues for the place of articulation of Korean nasal consonants. In the acoustic analysis, spectral and temporal parameters were measured at the 25%, 50%, and 75% time points in the vowels neighboring nasal consonants in samples extracted from a spontaneous Korean speech corpus. Using these measurements, linear discriminant analyses were performed and classification accuracies for the nasal place of articulation were estimated. The analyses were applied separately for vowels following and preceding a nasal consonant to compare the effects of progressive and regressive coarticulation in terms of place of articulation. The classification accuracies ranged between approximately 50% and 60%, implying that acoustic measurements of vowel intervals alone are not sufficient to predict or classify the place of articulation of adjacent nasal consonants. However, given that these results were obtained for measurements at the temporal midpoint of vowels, where they are expected to be the least influenced by coarticulation, the present results also suggest the potential of utilizing acoustic measurements of vowels to improve the recognition accuracy of nasal place. Moreover, the classification accuracy for nasal place was higher for vowels preceding the nasal sounds, suggesting the possibility of higher anticipatory coarticulation reflecting the nasal place.

Neural networks’ posterior probability as measure of effects of alcohol on speech

Intoxication has a well-known effect on speech production. Lester and Skousen (1974) reported that the place of articulation for /s/ is retracted and /tʃ/ and /ʤ/ are deaffricated (i.e., substituted by a non-affricate segment) in drunken speech. Zihlmann (2017) further established the robustness of deaffrication as it cannot be consciously suppressed under intoxication. Using these prevalent speech errors as test cases, this study extends a phonologically-informed neural network approach to the study of intoxicated speech. The approach has success in measuring pathological speech and lenition patterns in healthy speakers. Degrees of place retraction for /s/ and deaffrication of /tʃ/ and /ʤ/ are estimated from posterior probabilities calculated by recurrent neural networks trained to recognize [anterior], [continuant] and [strident] features. When applied to a corpus of alcohol English speech, preliminary results suggested that sober versus drunken state could be reliably predicted by the three posterior probabilities. The directions of the effects are largely in line with previous studies. For example, /tʃ/ and /ʤ/ are more fricated (higher strident and continuant probabilities), and /s/ is more retracted (lower anterior probability) in drunken compared to sober speech. The results suggest that the intoxicated speech can be reliably quantified by this new approach.

A Corpus-based Study on Speech Errors in Pronouncing the Fricative // by Chinese Learners of English

There were quite a few studies investigating English pronunciation errors of Chinese learners of English. However, studies focusing on one typical sound and its detailed features are still inadequate. Besides, limited studies used speech corpora to investigate this topic due to a lack of available speech corpora. To fill such a research gap, this study uses the corpus L2-ARCTIC to investigate the speech errors of voiceless dental fricative // among four Chinese learners of English, and analyzes the possible reasons for the speech errors. Data are collected by analyzing the speech recordings from the corpus L2-ARCTIC, and the pronunciation errors of speakers are classified into different types. Results show that, in this corpus, the types of speech errors of the fricative // are the substitution and deletion. The sounds that have very close manners and places of articulation are easier to become the error sounds to replace the // sound.

Characteristics of Alveolo-palatal Affricates Produced by Mandarin-speaking Children with Repaired Cleft Palate.

Objectives: In this study, examined the acoustic properties of affricates /t/ and /th/ in Mandarin Chinese, and analyzed the differences of the acoustic characteristics of these affricates produced by children with repaired cleft palate and normally developing children. We also explored the relationship between the affricates and high-front vowel /i/. Methods: We analyzed 16 monosyllabic words with alveolo-palatal affricates as the initial consonants produced by children with repaired cleft palate (N=13, Mean=5.9 years) and normally developing children (N=6, Mean age=5.3 years). We used several acoustic parameters to investigate the characteristics of these affricates, such as the center of gravity, VOT and the formants of vowels. Results: Compared with normally developing children, children with cleft palate exhibited a lower center of gravity for the 2 affricates /t/ and /th/. Data from the control group showed that the affricate /th/ had a significantly greater center of gravity than that of /t/. The accuracy of /t , th/ produced by speakers of cleft palate was significantly correlated with that of /i/ (r=0.63). High-front vowel /i/ is a significant index in diagnosing speech intelligibility which is more valuable than /a/ and /u/. There was a significant difference in F2 of vowel /i/ between children with cleft palate without speech therapy (CS1) and after speech therapy (CS2). After speech intervention, the accuracy of affricates produced by children with cleft palate was improved, the acoustic properties "stop + noise segments" appeared. Conclusion: Children with cleft palate can be distinguished better from children with normal development by 2 significant acoustic characteristics: center of gravity and VOT. As alveolo-palatal affricates /t , th/ and high-front vowel /i/ have a similar place of articulation, front-tongue-blade, their production accuracy can be improved mutually. The analysis showed that the articulation of Chinese /i/ has a higher frontal lingual position and less variability, which is more conducive to articulation training and improves the effect of cleft palate training. These findings provide a potential relationship on affricates /t, th/ and vowel /i/. Children with cleft palate have difficulty pronouncing the /t, t h/ and /i/. It is better to start with a vowel /i/, resulting in improvement in overall speech intelligibility.