Pairs Of Phonemes Research Articles

Contrast preservation and other segmental effects in the formation of Xochistlahuaca Amuzgo plurals

In Amuzgo (Eastern Otomanguean), the formation of nominal plurals exhibits many realizations, ranging from the simple addition of a nasal prefix, to additional initial consonant fortition, initial consonant deletion, and sometimes also the replacement of the prefixal nasal by a lateral. In this paper, we argue that all of these changes follow from two main principles: (1) The underlying contrast between the two pairs of phonemes characterized by a delayed release – the [+anterior] /s, ʦ/ and the [‑anterior] /ʃ, ʧ/ – must be maintained; and (2) /s, ʃ/ cannot be faithfully realized after [n]. These principles, in interaction with other considerations, lead to an establishment of a push chain (/s/→/ʦ/→/t/) among [+anterior] consonants and to a case of saltation (/ʧ/→/ʧ/; /ʃ/→/k/) among [-anterior] consonants.

Perception of temporal synchrony not a prerequisite for multisensory integration

Temporal alignment is often viewed as the most essential cue the brain can use to integrate information from across sensory modalities. However, the importance of conscious perception of synchrony to multisensory integration is a controversial topic. Conversely, the influence of cross-modal incongruence of higher level stimulus features such as phonetics on temporal processing is poorly understood. To explore the nuances of this relationship between temporal processing and multisensory integration, we presented 101 participants (ranging from 19 to 73 years of age) with stimuli designed to elicit the McGurk/MacDonald illusion (either matched or mismatched pairs of phonemes and visemes) with varying degrees of stimulus onset asynchrony between the visual and auditory streams. We asked them to indicate which syllable they perceived and whether the video and audio were synchronized on each trial. We found that participants often experienced the illusion despite not perceiving the stimuli as synchronous, and the same phonetic incongruence that produced the illusion also led to significant interference in simultaneity judgments. These findings challenge the longstanding assumption that perception of synchrony is a prerequisite to multisensory integration, support a more flexible view of multisensory integration, and suggest a complex, reciprocal relationship between temporal and multisensory processing.

A Study on Fricatives and Affricates Produced by Chinese English Learners

The resemblance between equivalent phonemes in Mandarin and English may be crucial for Chinese English learners. Three pairs of comparable fricative and affricates exit in the two languages. Prior research on fricatives and affricatives has tended to ignore the outputs of second language learners and instead concentrated on those of monolingual speakers. As a result, the research question of the current paper is to study fricatives and affricates produced by Chinese English learners. After extracting the main acoustic features of fricatives and affricates that occurred in both English and Mandarin produced by Chinese English learners, one-way ANOVA tests are used to examine the significance. Multidimensional Euclidean distance is used to compare the similarity between each pair of comparable phonemes. The results of these tests led to the conclusion that Mandarin has a detrimental impact on English when learning a second language. What is also worth noticing is that gender may work in the acquisition process, which needs further research to verify.

Decoding speech sounds from neurophysiological data: Practical considerations and theoretical implications.

Machine learning techniques have proven to be a useful tool in cognitive neuroscience. However, their implementation in scalp-recorded electroencephalography (EEG) is relatively limited. To address this, we present three analyses using data from a previous study that examined event-related potential (ERP) responses to a wide range of naturally-produced speech sounds. First, we explore which features of the EEG signal best maximize machine learning accuracy for a voicing distinction, using a support vector machine (SVM). We manipulate three dimensions of the EEG signal as input to the SVM: number of trials averaged, number of time points averaged, and polynomial fit. We discuss the trade-offs in using different feature sets and offer some recommendations for researchers using machine learning. Next, we use SVMs to classify specific pairs of phonemes, finding that we can detect differences in the EEG signal that are not otherwise detectable using conventional ERP analyses. Finally, we characterize the timecourse of phonetic feature decoding across three phonological dimensions (voicing, manner of articulation, and place of articulation), and find that voicing and manner are decodable from neural activity, whereas place of articulation is not. This set of analyses addresses both practical considerations in the application of machine learning to EEG, particularly for speech studies, and also sheds light on current issues regarding the nature of perceptual representations of speech.

Application of Discriminative Training Algorithm Based on Intelligent Computing in English Translation Evaluation

Abstract Aiming at the higher correlation between the objective evaluation of computer English speech and the subjective evaluation of experts, an acoustic model based on discriminative training is proposed to improve the confidence score of objective evaluation. First, the process of obtaining the pronunciation quality evaluation score of the speech vector by the forced matching algorithm is introduced, and then the mathematical theory of hypothesis testing is used to prove that the acoustic model trained based on the discriminative algorithm ‘minimum phoneme error’ is more effective than the acoustic model based on the traditional maximum likelihood algorithm. Confidence scores close to subjective assessments are obtained. By calculating the correlation coefficient of the subjective and objective evaluation results, the experiment verifies that the speech evaluation system using the discriminative acoustic model can give a higher confidence score and proposes a data selection method based on dynamic weighting, which is applied to continuous speech recognition in the discriminative training of the acoustic model. This method combines the posterior probability and the phoneme accuracy rate to select the data. First, the Beam algorithm of the posterior probability is used to trim the word graph. On this basis, according to the error rate of the candidate path where the candidate word is located, the probability dynamically assigns different weights to the candidate words; second, by calculating the confusion degree between the phoneme pairs, different penalty weights is dynamically added to the easily confused phoneme pairs to calculate the phoneme accuracy; finally, the expected accuracy of the obtained arc is calculated on the basis of the probability distribution. The Gaussian function is used to softly weight the expected phoneme accuracy of all competing arcs. The experimental results show that compared with the minimum phoneme error criterion, the dynamic weighting method has higher recognition accuracy and can effectively reduce the training time.

Auditory Word Recognition Ability in Babble Noise and Phonological Development in Children at 3;6 Years of Age

ABSTRACT Background noise challenges auditory recognition of speech and may reveal the underlying deficits in auditory word recognition skills. Previous studies have reported an association between children’s auditory skills and various linguistic skills, including phonology, although in some languages only. However, language-specific features influence these connections. This study describes the associations between auditory word recognition ability and phonological skills in a representative group of 3;6-year-old children in less studied language, Finnish (N = 65). Auditory recognition was assessed using a closed-set word recognition task presented in multi-talker babble noise (+13 dB) and the phonological skills using Finnish Phonology Test. A significant, moderate correlation was found between Finnish-speaking children’s auditory word recognition and phonological skills. The late-developing phoneme /r/ challenged word recognition. The sibilant /s/ was noted to be the most noise resistant consonant, and the phoneme pair /r/ and /l/ created the most mutual confusion. The accuracy of phonological representations may be a moderating factor for both auditory recognition and phonological skills. Children with strong phonological skills may recognize spoken words more accurately in noisy everyday situations than children with weaker phonological skills. This should be taken into consideration in children’s daily environments, such as daycare centers and preschools.

한국어 음소의 기능부담량 연구: 《한국어 자연발화 음성 코퍼스》를 대상으로

The purpose of this study was to investigate the functional load of Korean. To accomplish this research purpose, the study computes the functional loads of phoneme pairs using the change in system entropy from the Korean Corpus of Spontaneous Speech and compares it with the functional load quantified based on the dictionary corpus. The data analysis show that the order of functional loads of the onsets is ‘/ㄱ/-/ㅈ/, /ㄱ/-/ㄴ/, /ㄱ/-/ㅁ/, … ’, that of the nucleus is ‘/ㅓ/-/ㅡ/, /ㅏ/-/ㅓ/, /ㅗ/-/ㅜ/, …’, and that of the coda is ‘/ㄴ/-/ㅇ/, /ㄴ/-/ㅁ/, /ㄴ/-/ㄹ/, …’. The similarities and differences in the functional loads of the two types of corpora are also discussed. The results of this study reveal the basic and important knowledge of the characteristics of the phonology of spoken Korean.

AN OVERVIEW OF THE EXISTING TEXT-TO-SPEECH(TTS) ALGORITHMS BASED ON DEEP LEARNING

Text-to-speech technology has been around for a long time, but it is still considered challenging. The creation of human-like synthesized speech requires the solution of many tasks. As an example, the definition of intonation, the polyphony of systems, the combination of pairs of phonemes without audible flaws, and the definition of prosody. Various methods exist for synthesizing speech from the text, including parametric synthesis, compilation synthesis, domain-specific synthesis, and full speech synthesis by rules. With the advent of the necessary computing resources and the accumulation of sufficient data for training, machine learning methods have become widely used in speech synthesis. With significant improvements in text-to-speech algorithms, text-to-speech is now widely used in mobile devices, smart speakers, and voice assistants. In spite of this, the algorithms of this class, developed at the present time, are not always able to complete the assigned task correctly. In some cases, they are unable to correctly put stress on the right parts of the text or speak with the right intonation. In this study, the existing text-to-speech deep learning algorithms and models are reviewed and compared. The comparison is made by taking into account the positive and negative aspects of the algorithms, the complexity of implementing the algorithms, and the availability of the existing libraries for implementation. We consider the following algorithms as the main ones: WaveNet, DeepVoice 3, Tacatron 2, and Transformer.

Improving Tactile Codes for Increased Speech Communication Rates in a Phonemic-Based Tactile Display.

Previous research has shown evidence of tactile speech acquisition of up to 500 English words presented as tactile phonemic patterns using a 4-by-6 tactor array worn on the forearm. This article describes modifications to some of the tactile codes encoding the 39 English phonemes, and ten additional codes as abbreviated patterns for the ten most frequent phoneme pairs in spoken English. The re-design aimed to reduce the duration of phonemes and phoneme pairs that occur most frequently, with the goal to increase tactile speech transmission rates. Code identification experiments were conducted with ten participants over three weeks using a video game. The average identification rate of the 49 modified codes (39 phonemes plus 10 phoneme pairs) was 83.3% with an average learning time of 6.2 hours. The average identification rate of the 49 codes in a retention test with 7 of the 10 participants after more than 90 days of no exposure to the tactile codes was 75.7%. An analysis using ideal transmission rates showed a 58% increase in transmission rate with the modified tactile codes as compared to the original codes, demonstrating that the improved codes can speed up tactile speech communication.

On the revealing the emotional valence in communication by text

The interaction between humans and between humans and machines is often based on texts. The statistical analysis, presented here, provides one more reason to suppose that meaningful language communication integrates an emotional component involved at the phonemic level. The performed analysis is of written English language, presented by means words’ phonetic transcripts. The phonemic content of the texts is thought in layers, supposing that the phonological forms involved in the communication have the role of phonological markers assisting emotion transfer. The words are considered as composed of biphones – phonemic pairs consisting of one vowel and one consonant, in either order. Phonemic metadata for this sublexical representation was obtained from a vast corpus of emotionally evaluated texts and submitted to statistical analysis. The result of the applied principle component regression showed that the phonemic content is very strongly related (r = 0.96) with the ratings of emotional valence (positive-negative emotion), provided by readers. The analysis showed the big importance of the first biphone in the words, in alliance with results for single phonemes reported in recent studies. The result shows that the phonemic content presented as a stream of biphones is strongly related to the emotional impact that a text has on the reader. This leads to suppose that the mechanisms that have shaped the phonological forms to communicate conceptual meaning involve emotional valence too.

Single-cell activity in human STG during perception of phonemes is organized according to manner of articulation

One of the central tasks of the human auditory system is to extract sound features from incoming acoustic signals that are most critical for speech perception. Specifically, phonological features and phonemes are the building blocks for more complex linguistic entities, such as syllables, words and sentences. Previous ECoG and EEG studies showed that various regions in the superior temporal gyrus (STG) exhibit selective responses to specific phonological features. However, electrical activity recorded by ECoG or EEG grids reflects average responses of large neuronal populations and is therefore limited in providing insights into activity patterns of single neurons. Here, we recorded spiking activity from 45 units in the STG from six neurosurgical patients who performed a listening task with phoneme stimuli. Fourteen units showed significant responsiveness to the stimuli. Using a Naïve-Bayes model, we find that single-cell responses to phonemes are governed by manner-of-articulation features and are organized according to sonority with two main clusters for sonorants and obstruents. We further find that ‘neural similarity’ (i.e. the similarity of evoked spiking activity between pairs of phonemes) is comparable to the ‘perceptual similarity’ (i.e. to what extent two phonemes are judged as sounding similar) based on perceptual confusion, assessed behaviorally in healthy subjects. Thus, phonemes that were perceptually similar also had similar neural responses. Taken together, our findings indicate that manner-of-articulation is the dominant organization dimension of phoneme representations at the single-cell level, suggesting a remarkable consistency across levels of analyses, from the single neuron level to that of large neuronal populations and behavior.

Predictability of higher-order temporal structure of musical stimuli is associated with auditory evoked response

Sound predictability resulting from repetitive patterns can be implicitly learned and often neither requires nor captures our conscious attention. Recently, predictive coding theory has been used as a framework to explain how predictable or expected stimuli evoke and gradually attenuate obligatory neural responses over time compared to those elicited by unpredictable events. However, these results were obtained using the repetition of simple auditory objects such as pairs of tones or phonemes. Here we examined whether the same principle would hold for more abstract temporal structures of sounds. If this is the case, we hypothesized that a regular repetition schedule of a set of musical patterns would reduce neural processing over the course of listening compared to stimuli with an irregular repetition schedule (and the same set of musical patterns). Electroencephalography (EEG) was recorded while participants passively listened to 6–8 min stimulus sequences in which five different four-tone patterns with temporally regular or irregular repetition were presented successively in a randomized order. N1 amplitudes in response to the first tone of each musical pattern were significantly less negative at the end of the regular sequence compared to the beginning, while such reduction was absent in the irregular sequence. These results extend previous findings by showing that N1 reflects automatic learning of the predictable higher-order structure of sound sequences, while continuous engagement of preattentive auditory processing is necessary for the unpredictable structure.

The Relationship between Speech Recognition Ability and Short-term Memory in Children with Down Syndrome

<p>In order to compare the difference in speech recognition ability and short-term memory between children with Down syndrome (DS) and normal children (TD), and to explore the relationship between DS children’s speech recognition ability and short-term memory, this study tested two groups of children with Chinese minimal phoneme pairs. Results: (1) The speech recognition ability of DS children was significantly lower than that of TD children in the control group; The short-term verbal memory of DS children is significantly lower than that of TD children in the control group. (2) The difference between DS children’s speech recognition ability and speech short-term memory performance of four stimuli is significantly lower than the difference between speech short-term memory performance of four stimuli and six stimuli; There was no significant difference between the speech recognition ability of TD children and the speech short-term memory scores of four stimuli and between the speech short-term memory scores of four stimuli and six stimuli. Conclusion: DS children have defects in speech recognition and short-term memory. DS children’s speech recognition ability is not the main reason for their poor short-term memory.</p>

Smart media and application

Smart media and application

한국어 음소 최소대립쌍의 계량언어학적 연구: 초성 자음을 중심으로

The paper investigates the minimal pair of Korean phonemes quantitatively. To achieve this goal, I calculated the number of consonant minimal pairs in the syllable-initial position as both raw counts and relative counts, and analyzed the part of speech relations of the two words in the minimal pair. 『Urimalsaem』 was chosen as the object of this study because it was judged that the minimal pair analysis should be done through a dictionary and it is the largest among Korean dictionaries. The results of the study are summarized as follows. First, there were 153 types of minimal pairs out of 337,135 examples. The ranking of phoneme pairs from highest to lowest was ‘ㅅ-ㅈ, ㄱ-ㅅ, ㄱ-ㅈ, ㄱ-ㅂ, ㄱ-ㅎ, ..., ㅆ-ㅋ, ㄸ-ㅋ, ㅉ-ㅋ, ㄹ-ㅃ, ㅃ-ㅋ’. The phonemes that played a major role in the formation of the minimal pair were /ㄱ, ㅅ, ㅈ, ㅂ, ㅊ/, in that order, which showed a high proportion of palatals. The correlation between the raw count of minimal pairs and the relative count of minimal pairs was found to be quite high r=0.937. Second, 87.91% of the minimal pairs shared the part of speech (same syntactic category). The most frequently observed type has been ‘noun-noun’ pair (70.25%), and ‘vowel-vowel’ pair (14.77%) was the next ranking. It can be indicated that the minimal pair could be grouped into similar categories in terms of semantics. The results of this study can be useful for various research in Korean linguistics, speech-language pathology, language education, language acquisition, speech synthesis, and artificial intelligence-machine learning as basic data related to Korean phonemes.

Effective computer‐assisted pronunciation training based on phone‐sensitive word recommendation

SummaryOwing to their autonomous nature, computer‐aided pronunciation training systems are regarded as a useful tool for pronunciation training for beginner‐level students, particularly those who feel uncomfortable having their pronunciation corrected in front of other students or who have difficulty with face‐to‐face training for other reasons. In this study, we propose a computer‐assisted pronunciation training system targeting unacceptable pronunciation due to confusion among contextual allophones, a problem that often emerges from phoneme‐based feedback provided during pronunciation training. To consider the different pronunciation of phonemes according to position, the proposed system is implemented to recommend a set of words focusing on phoneme pairs. Experimental results show that the proposed system results in the improvement of pronunciation skills through training sessions that use recommended words containing phoneme pairs that were initially pronounced incorrectly.

ЧАСТОТНЫЕ ХАРАКТЕРИСТИКИ ФОНЕМ В ЭВЕНКИЙСКОМ ЯЗЫКЕ: ДАННЫЕ ПИЛОТНОГО ЭКСПЕРИМЕНТА

The current paper aimed at studying the 2 types of frequency of occurrence of phoneme groups and in some phoneme pairs - in the Evenki phonological system and speech. The results of the analysis of speech samples obtained from 4 Evenki subjects showed that there are objective grounds to consider this language as highly vocalic. These are notably higher frequencies of occurrence of vocalic sounds (vowels + sonorants) compared to obstruents both in the system and speech. Evenki is more vocalic than English and Russian and less vocalic than Yakut as far as system frequency of occurrence. This tendency is even better expressed in speech for Evenki, English and Russian. As far as there are no data on Yakut phoneme objective frequency, this comparison is a matter of further study. For both sonorants and obstruents, system and objective frequency showed similar tendencies: prevailing of stops over fricatives, nasals over approximants, fore-lingual consonants over ones of other loci...

Variation in a tense/lax vowel pair in Dutch youngsters with different ethnic backgrounds

Abstract How do young bilingual speakers of current Turkish and Moroccan ethnolects of Dutch deal with phoneme contrasts that do not exist in their heritage languages and that are at the same time subject to regional and social variation in the Dutch speech community at large, such as that between Dutch phonemes /a:/ and /α/? Data from speakers from the Amsterdam and Nijmegen urban areas were analyzed and compared. Two variable properties of /a:/ and /α/ defining the phoneme pair were examined: (1) duration, and (2) place of articulation. We found clear differences between the two urban areas (regional effect) and between the Amsterdam ethnic groups (social effect). In addition, we found variation dependent on the interlocutor (socio-stylistic effect) and the linguistic context. The main dimension of social and linguistic variation was place of articulation, length remaining the primary distinctive feature of the vowel pair. No heritage language effect was found. Young ethnolect speakers take part in existing patterns of regional and social variation.

Statistical and acoustic effects on the perception of stop consonants in Kaqchikel (Mayan)

This paper investigates the relationship between speech perception and linguistic experience in Kaqchikel, a Guatemalan Mayan language. Our empirical focus is the perception of plain, ejective, and implosive stops. Drawing on an AX discrimination task, a corpus of spoken Kaqchikel, and a text corpus, we make two claims. First, we argue that speech perception is mediated by phonemic representations which include acoustic detail drawn from prior phonetic experience, as in Exemplar Theory. Second, segmental distributions also condition speech perception: The perceptual distinctiveness of a pair of phonemes is affected by their functional load and relative contextual predictability. These top-down factors influence phoneme discrimination even at relatively fast response times. We take this result as evidence that distributional factors like functional load may affect speech perception by shaping perceptual tuning during linguistic development. This study replicates and extends some key findings in speech perception in the context of a language (Kaqchikel) which is structurally and sociolinguistically different from the majority languages (like English) which have served as the basis of most work in the speech perception literature. At the practical level, our research illustrates methods for conducting corpus-based laboratory phonology with lesser-studied and under-resourced languages.

A resource model of phonological working memory in language production and perception

A resource model of phonological working memory in language production and perception