Portions Of Speech Research Articles

Modulation and Energy Components for Speaker Recognition System

This paper presents a model use the modulation and energy components for speaker recognition application, that is mainly follows theshort-term scenario in speech signal processing, and also introduce a parameter combination that includes the instantaneous components and the energy parameters. This will describe the importance of short-term speech analysis in estimating the modulation parameters and the role of the instantaneous energy in estimating the speaker-dependent parameters. Simply, the short-term scenario is used to, first; avoid the silent and background noise speech portions that present in speech signals, and also to benefitfrom the stationary concept of the short-term processing in the speech signal. The energy components, on the other hand, are adopted purely in many speech parameterisation models, such as, linear predictive coding (LPC) and Mel-frequency cepstral coefficients (MFCCs). The main ideal of our mixture parameter (or MFCC/AM-FM model) is to determined the extent of these components to contribute together in extracting the parameters that are more related to the speaker more than anything else presented in the speech signal. We evaluated both models using the text-dependent and text-independent speech corpora. The accuracy results show that the frame-based AM-FM model achieve better performance comparing with the traditional structure of the AM-FM modulation model(the model presented in [1]). The MFCC/AM-FM parameters, on the other hand, perform much better, in terms of text-dependent, comparing with the AM-FM parameters and the MFCC parameters. In the case of the text-independent, however, the MFCC/AM-FM model provide better results than the MFCC features but less performance comparing to the AM-FM modulation parameters.

Differential sensitivity to speech rhythms in young and older adults.

Sensitivity to the temporal properties of auditory patterns tends to be poorer in older listeners, and this has been hypothesized to be one factor contributing to their poorer speech understanding. This study examined sensitivity to speech rhythms in young and older normal-hearing subjects, using a task designed to measure the effect of speech rhythmic context on the detection of changes in the timing of word onsets in spoken sentences. A temporal-shift detection paradigm was used in which listeners were presented with an intact sentence followed by two versions of the sentence in which a portion of speech was replaced with a silent gap: one with correct gap timing (the same duration as the missing speech) and one with altered gap timing (shorter or longer than the duration of the missing speech), resulting in an early or late resumption of the sentence after the gap. The sentences were presented with either an intact rhythm or an altered rhythm preceding the silent gap. Listeners judged which sentence had the altered gap timing, and thresholds for the detection of deviations from the correct timing were calculated separately for shortened and lengthened gaps. Both young and older listeners demonstrated lower thresholds in the intact rhythm condition than in the altered rhythm conditions. However, shortened gaps led to lower thresholds than lengthened gaps for the young listeners, while older listeners were not sensitive to the direction of the change in timing. These results show that both young and older listeners rely on speech rhythms to generate temporal expectancies for upcoming speech events. However, the absence of lower thresholds for shortened gaps among the older listeners indicates a change in speech-timing expectancies with age. A further examination of individual differences within the older group revealed that those with better rhythm-discrimination abilities (from a separate study) tended to show the same heightened sensitivity to early events observed with the young listeners.

The Textual Theme Potrayed on the Talks of Education in the Indonesian’s TEDx

This study sought to describe education-related conversations or talks of education (TE) from a systemic functional linguistics (SFL) perspective, focusing on how textual themes were deployed by speakers. TEDx talk, a reliable online source, was chosen as the source for the data used to support this study. The analysis’s validated data was the transcription of an education-related speech. Documentation, transcription, and in-depth textual fragmentation were the three primary techniques utilized during data processing and analysis. This study demonstrates, based on its findings, that the textual theme was not utilized in all portions of speech; it was the type of theme for which the speaker required to provide further information. Despite the limitations of performance, a particular construction and functionalization have been applied to the textual theme. There were phrasal and clausal formulations of textual theme, as well as their role within multiple themes.

Supervised Contrastive Learning for Voice Activity Detection

The noise robustness of voice activity detection (VAD) tasks, which are used to identify the human speech portions of a continuous audio signal, is important for subsequent downstream applications such as keyword spotting and automatic speech recognition. Although various aspects of VAD have been recently studied by researchers, a proper training strategy for VAD has not received sufficient attention. Thus, a training strategy for VAD using supervised contrastive learning is proposed for the first time in this paper. The proposed method is used in conjunction with audio-specific data augmentation methods. The proposed supervised contrastive learning-based VAD (SCLVAD) method is trained using two common speech datasets and then evaluated using a third dataset. The experimental results show that the SCLVAD method is particularly effective in improving VAD performance in noisy environments. For clean environments, data augmentation improves VAD accuracy by 8.0 to 8.6%, but there is no improvement due to the use of supervised contrastive learning. On the other hand, for noisy environments, the SCLVAD method results in VAD accuracy improvements of 2.9% and 4.6% for “speech with noise” and “speech with music”, respectively, with only a negligible increase in processing overhead during training.

Using automated syllable counting to detect missing information in speech transcripts from clinical settings

Speech rate and quantity reflect clinical state; thus automated transcription holds potential clinical applications. We describe two datasets where recording quality and speaker characteristics affected transcription accuracy. Transcripts of low-quality recordings omitted significant portions of speech. An automated syllable counter estimated actual speech output and quantified the amount of missing information. The efficacy of this method differed by audio quality: the correlation between missing syllables and word error rate was only significant when quality was low. Automatically counting syllables could be useful to measure and flag transcription omissions in clinical contexts where speaker characteristics and recording quality are problematic.

Muscle Sympathetic Nerve Reactivity to Trier Social Stress Test

Muscle sympathetic nerve activity (MSNA) responses to laboratory mental stress are highly variable, making characterization of the sympathoneural response difficult to infer. However, most commonly used mental stress tasks (i.e., mental arithmetic, Stroop Color Word Task, etc.) are primarily cognitive tasks. Further, interpretation of MSNA reactivity to these stressors can be challenging due to irregular breathing patterns, limb movement, etc. The current study assesses MSNA reactivity to the Trier Social Stress Test (TSST), an ecologically valid evaluative social stress task that utilizes an anticipatory stress period free of confounds such as irregular breathing and movement found in other mental stress tasks. The primary purpose of the present study was to assess MSNA during the stress‐anticipation period, as well as during the socially evaluated portion of the task. We hypothesized that MSNA would be elevated during both phases of the task. Sixteen healthy adults (12 female, 26 ± 1 years, 24 ± 1 kg/m2) participated in a morning autonomic function test consisting of simultaneous recordings of heart rate (HR, electrocardiogram), beat‐to‐beat blood pressure (finger plethysmography), and MSNA (microneurography) during a 10‐minute resting baseline followed by the TSST. The TSST consists of a speech preparation period (5 min), a public speech (5 min), and mental arithmetic task (5 min) in front of two laboratory judges and video recording. After completion of the TSST, participants were assessed throughout a 5‐minute recovery. MSNA was obtained throughout all tasks in 12 individuals. Compared to baseline, mean arterial pressure (MAP) significantly increased during the anticipatory phase (Δ8 ± 2 mmHg, p<.001), increased further during the speech (Δ25 ± 3 mmHg, p<.001) and arithmetic (Δ23 ± 3 mmHg, p<.001), and returned to anticipatory levels during recovery (Δ10 ± 1 mmHg, p<.001). Similarly, HR was elevated during the anticipatory (Δ8 ± 2 bpm, p=.002), and increased further during the speech (Δ24 ± 5 bpm, p<.001) and arithmetic (Δ19 ± 4 bpm, p<.001) phases, but fully returned to baseline levels during recovery (Δ2 ± 1 bpm p=.232). MSNA burst frequency was significantly reduced during the anticipatory phase (Δ‐5 ± 2 bursts/min, p=.014) and tended to be reduced during mental arithmetic (Δ‐5 ± 2 bursts/min, p=.058), while MSNA burst incidence was reduced during the anticipatory (Δ‐9 ± 3 bursts/100hb, p=.008) and arithmetic (Δ‐11 ± 4 bursts/100hb, p=.016) phases. In 7 individuals, MSNA signal quality allowed assessment of MSNA total area. While total MSNA appeared to increase during the speech portion of the TSST, this effect did not reach significance (Δ8906 ± 4547 arbitrary units, p=.098). Contrary to our hypothesis, while blood pressure and HR were significantly elevated, MSNA was reduced in response to social stress, likely due to baroreflex inhibition. Further, our findings exhibit previously observed inter‐individual variability in MSNA responsiveness to mental stress, even during anticipatory stress, suggesting that this variability is conserved between various laboratory stressors.

Parental Self-Efficacy and Physiological Responses to Stress among Mothers of Early Adolescents

Little research addresses how parental self-efficacy is related to stress responses, and no research does so among parents of early adolescents. To fill this research gap, the current study examined the association between maternal self-efficacy and physiological stress responses during early adolescence. Participants were 68 mother-early adolescent dyads with youth in the 6th grade (M = 11 years; 56% female). Physiological responses (i.e., skin conductance, respiratory sinus arrythmia, cortisol) were measured before and after mothers observed their children engage in a modified Trier Social Stress Test for Children. Mothers reported on parental self-efficacy. Mothers with higher parental self-efficacy exhibited a more moderate skin conductance response to the speech portion of the task, and a smaller increase in cortisol, compared to mothers with lower parental self-efficacy. Respiratory sinus arrhythmia change was not related to parental self-efficacy. The findings are consistent with a "caring but confident" physiological profile among mothers with high parental self-efficacy, suggesting that greater confidence about parental influence might reduce parents' experience of stress/anxiety as they observe children face certain challenges.

Modifying the Trier Social Stress Test to Induce Positive Affect.

The online version contains supplementary material available at 10.1007/s42761-021-00074-6.

Cortisol Reactivity and Socially Anxious Behavior in Previously Institutionalized Youth.

The current study investigated the association between cortisol stress reactivity to a social stressor and observed socially anxious behaviors both concurrently and over time among previously institutionalized (PI) (N = 132; ages 7-17) youth and a comparison non-adopted (NA) sample (N = 176). Cortisol reactivity was captured during the Trier Social Stress Test for Children (TSST-C; Yim et al., 2015) and youths' social anxiety behaviors were coded during the speech portion of the TSST-C. Autoregressive cross-lagged panel models with structured residuals showed that for PI youth, greater cortisol reactivity predicted increases in socially anxious behavior during the TSST-C across three sessions. However, greater cortisol reactivity was negatively associated with concurrent social anxiety behavior. Thus, increases in cortisol reactivity across adolescence may aid in behavioral control in social situations in the short-term but may exacerbate PI youths' socially anxious behavior over time. No significant associations emerged for NA youth.

Standardized stress reduction interventions and blood pressure habituation: Secondary results from a randomized controlled trial.

The present article reports new analyses of a published randomized controlled trial comparing the effects of mindfulness-based stress reduction (MBSR), cognitive behavioral therapy (CBT), and a passive control condition on blood pressure habituation-a secondary outcome. Psychological mediators of intervention effects were examined. Participants (138 healthy adults reporting moderate/high stress) were randomly assigned to 6-week MBSR, CBT, or a waitlist control. Analyses were conducted on 86 participants who subsequently completed stressful speech and mental arithmetic tasks during two posttreatment visits scheduled 48 hr apart. Blood pressure was measured -15, +0, +5, +10, +25, +35, and +60 min poststressor onset. There were no between-condition differences in blood pressure habituation (all ps > .05). However, both MBSR and CBT led to increased perceived control over thoughts, F(2, 72) = 5.20, p = .008, and individuals who displayed a greater change in perceived control over thoughts also displayed greater habituation to the speech portion of the stressor, F(6, 799) = 2.32, p = .020. Results implied an indirect effect of stress reduction interventions on blood pressure habituation via change in perceived control over thoughts (b = -3.93, SE = 1.98, 95% CI: [-8.392, -0.701]). Stress reduction interventions that increase perceived control over thoughts may benefit cardiovascular health by promoting blood pressure habituation. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

Comparative Analysis of Majority Language Influence on North Sámi Prosody Using WaveNet-Based modeling.

The Finnmark North Sámi is a variety of North Sámi language, an indigenous, endangered minority language spoken in the northernmost parts of Norway and Finland. The speakers of this language are bilingual, and regularly speak the majority language (Finnish or Norwegian) as well as their own North Sámi variety. In this paper we investigate possible influences of these majority languages on prosodic characteristics of Finnmark North Sámi, and associate them with prosodic patterns prevalent in the majority languages. We present a novel methodology that: (a) automatically finds the portions of speech (words) where the prosodic differences based on majority languages are most robustly manifested; and (b) analyzes the nature of these differences in terms of intonational patterns. For the first step, we trained convolutional WaveNet speech synthesis models on North Sámi speech material, modified to contain purely prosodic information, and used conditioning embeddings to find words with the greatest differences between the varieties. The subsequent exploratory analysis suggests that the differences in intonational patterns between the two Finnmark North Sámi varieties are not manifested uniformly across word types (based on part-of-speech category). Instead, we argue that the differences reflect phrase-level prosodic characteristics of the majority languages.

ConflictNET: End-to-End Learning for Speech-Based Conflict Intensity Estimation

Computational paralinguistics aims to infer human emotions, personality traits and behavioural patterns from speech signals. In particular, verbal conflict is an important example of human-interaction behaviour, whose detection would enable monitoring and feedback in a variety of applications. The majority of methods for detection and intensity estimation of verbal conflict apply off-the-shelf classifiers/regressors to generic hand-crafted acoustic features. Generating conflict-specific features requires refinement steps and the availability of metadata, such as the number of speakers and their speech overlap duration. Moreover, most techniques treat feature extraction and regression as independent modules, which require separate training and parameter tuning. To address these limitations, we propose the first end-to-end convolutional-recurrent neural network architecture that learns conflict-specific features directly from raw speech waveforms, without using explicit domain knowledge or metadata. Additionally, to selectively focus the model on portions of speech containing verbal conflict instances, we include a global attention interface that learns the alignment between layers of the recurrent network. Experimental results on the SSPNet Conflict Corpus show that our end-to-end architecture achieves state-of-the-art performance in terms of Pearson Correlation Coefficient.

Dealing with the unknown – addressing challenges in evaluating unintelligible speech

ABSTRACT When investigating the interaction between speech production and intelligibility, unintelligible speech portions are often of particular interest. Therefore, the fact that the standard quantification of speech production – the Percentage of Consonants Correct (PCC) – is only computed on intelligible speech is unsatisfying. Our purpose was to evaluate a new quantification of speech production: the Percentage of Intelligible and Correct Syllables (PICS) designed to address this limitation. A secondary purpose was to explore a task designed to elicit connected speech – concurrent commenting – offering more control of intended speech production compared to free conversation. Nine children with SSD participated in four speech elicitation tasks, varying with respect to ecological validity, and to degree of control: (1) word imitation, (2) picture naming, (3) concurrent commenting of a silent short video, and (4) free conversation. Speech accuracy was analysed in terms of PCC-Revised (PCC-R) and PICS, and intelligibility with regards to the Proportion of Intelligible Syllables (PINTS). A strong correlation was observed between PICS and PCC-R, supporting the construct validity of PICS. Further, a moderate correlation was seen between PICS and PINTS, presumably reflecting that these measures capture different – although related – constructs. No difference was seen between concurrent commenting and free conversation in terms of articulation proficiency or intelligibility; however, this needs further investigation based on more data. Nevertheless, we suggest concurrent commenting as a useful method for eliciting connected speech; in retaining more control over intended target words compared to free conversation, this task may be particularly useful in the context of unintelligible speech.

Remembering what was said and done: The activation and facilitation of memory for gesture as a consequence of retrieval.

The gestures that occur alongside speech provide listeners with cues that both improve and alter memory for speech. The present research investigated the interplay of gesture and speech by examining the influence of retrieval on memory for gesture. In three experiments, participants watched video clips of an actor speaking a series of statements with or without gesture before being asked to retrieve the speech portions of half of those statements. Participants were then tested on their ability to recall whether the actor had gestured during each statement and, if so, to recall the nature of the gesture that was produced. Results indicated that attempting to retrieve the speech portion of the statements enhanced participants' ability to remember the gesture portion of the statements. This result was only observed, however, for representational gestures when the speech and gesture components were meaningfully related (Experiments 1 & 2). It was not observed for beat gestures or nonsense gestures (Experiments 2 & 3). These results are consistent with the idea that gestures can be coactivated during the retrieval of speech and that such coactivation is due to the integrated representation of speech and gesture in memory. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

Emotion regulation and cortisol reactivity during a social evaluative stressor: A study of post-institutionalized youth.

In the current study, we compared emotion regulation abilities between post-institutionalized (PI; N=124) and never-institutionalized non-adopted (NA; N=172) children and adolescents (7-15years). We assessed cortisol reactivity and coded emotion regulation during the speech portion of Trier Social Stress Test (TSST-M). Parents reported on their children's social, academic, and behavioral adjustment. Results suggest that emotion regulation abilities increased with age, but this increase was greater for NA than PI youth. With regard to cortisol, piecewise growth modeling revealed that at higher levels of emotion regulation PI youth had greater baseline values (after a period of time allowing for acclimation to the laboratory) and had steeper recovery slopes than NA youth. There was also a main effect of emotion regulation on the reactivity slope suggesting that for both groups, as emotion regulation increased, the cortisol reactivity slope decreased. Finally, greater emotion regulation predicted fewer internalizing behavior problems for PI youth but not for NA youth.

Effect of speech segment samples selection in stutter block detection and remediation

Speech remediation by identifying those segments which take away from the substance of the speech content can be performed by identifying portions of speech which may be deleted without diminishing from the speech quality, but rather improving the speech. Speech remediation is important when the speech is disfluent as in the case of stuttered speech. We describe two stuttered speech remediation approaches based on the identification of those segments of speech which, when removed, would enhance speech understandability in terms of both, speech content and speech flow. We adopted two approaches, in the first approach we identify and extract speech segments that have weak semantic significance due to their low relative intensity; we subsequently trained several classifiers using a large set of inherent and derived features which provided a second layer filtering stage. The first approach was effective but required a two step process. In order to streamline the detection and remediation process, we introduced an enhancement which expands the realm of disfluency detection to include a broader range of speech anomalies by eliminating the need for a domain-dependent pre-qualification stage. The results of the new approach offer improved accuracy with enhanced simplicity, flexibility and extensibility.

Linguistic effect on speech perception observed at the brainstem

Linguistic experience affects speech perception from early infancy, as previously evidenced by behavioral and brain measures. Current research focuses on whether linguistic effects on speech perception can be observed at an earlier stage in the neural processing of speech (i.e., auditory brainstem). Brainstem responses reflect rapid, automatic, and preattentive encoding of sounds. Positive experiential effects have been reported by examining the frequency-following response (FFR) component of the complex auditory brainstem response (cABR) in response to sustained high-energy periodic portions of speech sounds (vowels and lexical tones). The current study expands the existing literature by examining the cABR onset component in response to transient and low-energy portions of speech (consonants), employing simultaneous magnetoencephalography (MEG) in addition to electroencephalography (EEG), which provide complementary source information on cABR. Utilizing a cross-cultural design, we behaviorally measured perceptual responses to consonants in native Spanish- and English-speaking adults, in addition to cABR. Brain and behavioral relations were examined. Results replicated previous behavioral differences between language groups and further showed that individual consonant perception is strongly associated with EEG-cABR onset peak latency. MEG-cABR source analysis of the onset peaks complimented the EEG-cABR results by demonstrating subcortical sources for both peaks, with no group differences in peak locations. Current results demonstrate a brainstem-perception relation and show that the effects of linguistic experience on speech perception can be observed at the brainstem level.

Enhancing speech intelligibility in reverberant spaces by a speech features distributions dependent pre-processing

In this paper, we deal with a pre-processing based on speech envelope modulation for intelligibility enhancement in reverberant large dimension public enclosed spaces. In fact, the blurring effect due to reverberation alters the speech perception in such conditions. This phenomenon results from the masking of consonants by the reverberated tails of the previous vowels. This is particularly accentuated for elderly persons suffering from presbycusis. The proposed pre-processing is inspired from the steady-state suppression technique which consists in the detection of the steady-state portions of speech and the multiplication of their waveforms with an attenuation coefficient in order to decrease their masking effect. While the steady-state suppression technique is performed in the frequency domain, the pre-processing described in this paper is rather performed in the temporal domain. Its key novelty consists in the detection of the speech voiced segments using a priori knowledge about the distributions of the powers and the durations of voiced and unvoiced phonemes. The performances of this pre-processing are evaluated with an objective criterion and with subjective listening tests involving normal hearing persons and using a set of nonsense Vowel–Consonant–Vowel syllables and railway station vocal announcements.

How repetition influences speech understanding by younger, middle-aged and older adults

Objective: To examine benefit from immediate repetition of a masked speech message in younger, middle-aged and older adults.Design: Participants listened to sentences in conditions where only the target message was repeated, and when both the target message and its accompanying masker (noise or speech) were repeated. In a follow-up experiment, the effect of repetition was evaluated using a square-wave modulated noise masker to compare benefit when listeners were exposed to the same glimpses of the target message during first and second presentation versus when the glimpses differed. Study Sample: Younger, middle-aged and older adults (n = 16/group) for the main experiment; 15 younger adults for the follow-up experiment.Results: Repetition benefit was larger when the target but not the masker was repeated for all groups. This was especially true for older adults, suggesting that these individuals may be more negatively affected when a background message is repeated. Data obtained using noise maskers suggest that it is slightly more beneficial when listeners hear different (versus identical) portions of speech between initial presentation and repetition.Conclusions: Although subtle age-related differences were found in some conditions, results confirm that repetition is an effective repair strategy for listeners spanning the adult age range.

Social evaluative threat with verbal performance feedback alters neuroendocrine response to stress

Laboratory stress tasks such as the Trier Social Stress Test (TSST) have provided a key piece to the puzzle for how psychosocial stress impacts the hypothalamic-pituitary-adrenal axis, other stress-responsive biomarkers, and ultimately wellbeing. These tasks are thought to work through biopsychosocial processes, specifically social evaluative threat and the uncontrollability heighten situational demands. The present study integrated an experimental modification to the design of the TSST to probe whether additional social evaluative threat, via negative verbal feedback about speech performance, can further alter stress reactivity in 63 men and women. This TSST study confirmed previous findings related to stress reactivity and stress recovery but extended this literature in several ways. First, we showed that additional social evaluative threat components, mid-task following the speech portion of the TSST, were still capable of enhancing the psychosocial stressor. Second, we considered stress-reactive hormones beyond cortisol to include dehydroepiandrosterone (DHEA) and testosterone, and found these hormones were also stress-responsive, and their release was coupled with one another. Third, we explored whether gain- and loss-framing incentive instructions, meant to influence performance motivation by enhancing the personal relevance of task performance, impacted hormonal reactivity. Results showed that each hormone was stress reactive and further had different responses to the modified TSST compared to the original TSST. Beyond the utility of showing how the TSST can be modified with heightened social evaluative threat and incentive-framing instructions, this study informs about how these three stress-responsive hormones have differential responses to the demands of a challenge and a stressor.