Conversational Speech Research Articles

Relacje online. Ekstrawersja i inteligencja emocjonalna w kontekście wybranych aspektów komunikacji internetowej

Interpersonal communication is a dynamic, complex process, and is conditioned by many factors, among which one of the most important roles is played by the personality of interaction partners. The article presents results of research on the influence of extraversion and emotional intelligence on selected aspects of communication when the process is mediated by the Internet. The study involved young adults (N = 51) aged 21–27 (M = 24.27; SD = 1.65; Md = 24), mostly women (Nk = 26; K/M = 1.08), who responded to the invitation to the study published on Facebook. The survey was conducted in the period from December 2019 to February 2020, through Messenger and GoogleForms to collect the results in personality questionnaires. On the basis of the correlation coefficients obtained, it can be assumed that the level of extraversion (r = .397, p = .002; r = 414, p < .001 and the level of emotional intelligence (r = .331, p = .009; r = .411, p < .001) may be responsible for the expression of emotions and productivity of the speech in an intermediary conversation. Due to the results of the conducted research, we put forward a thesis that the mediation of contact with another person thanks to the currently developed systems of smileys, does not lead to impoverishment of the emotional layer of the message.

DISCOURSE AS A UNIT OF LINGUISTICS

The present article is devoted to the culture, as a specific form of human relations, is represented by objects, actions, words and that the cross-cultural communication this is the process of direct interaction of cultures. Conversational analysis and discursive analysis are general, and the basis for the study is the selection of natural real dialogues. The analysis of conjugation offers to consider dialogical relations in their specific manifestation. Discursive analysis includes the study of the strategies and intentions of the speakers, the processes of intelligence: ethnographic, psychological and socio-cultural rules and strategies for the formation and understanding of speech. The study of communication can be seen as an integral part of conversation analysis and speech analysis.

Hitz-bukaerako herskariak: kontsonante talkak

Euskaraz hitz-bukaerako herskariek oso maiztasun txikia dute hiztegian, baina oso maiztasun handia solasean; /-t/ eta /-k/ kontsonanteak maiztasun handiko atzizki batzuetan aurkitzen baitira. Lan honetan hitz-bukaerako herskarien gauzatze fonetikoa aztertzen dugu, eta bereziki bi hitzen arteko kontsonante taldeetan jartzen dugu arreta. Datuak elkarrizketa libretik atera ditugu. Intentsitate neurketak erakusten du talde horietan kontsonanteak oso ahulak direla gehienetan. Espektrogramak miatuz baieztatzen dugu /-k/ eta /-t/ kontsonanteak beste kontsonante baten aurrean gehiengoetan galtzen edo hurbilkari ahostun gisa ahoskatzen direla. Bestalde, ahoskabetze gertakariak (adib. /k#b/ --> [p]) ez dira batere normalak gure corpusean.

Основные принципы и подходы использования проектной методики при обучении разговорной речи студентов

The article describes and analyzes the basic principles and approaches of using the project methodology in teaching spoken English to students. The analysis has identified the basic provisions of the fundamental principles and approaches on which the process of developing a project methodology for teaching students' conversational speech is based.

Fundamental frequency range and other acoustic factors that might contribute to the clear-speech benefit

The main goal of the present study was to assess the role of the fundamental frequency (F0) range on the clear-speech benefit. Conversational- and clear-speech sentences were recorded for four male speakers: the speakers' clear-speech productions had slower speaking rates, wider F0 range, more high-frequency energy, expanded vowel space, and higher vocal intensity level relative to their conversational-speech productions. To examine if F0 range contributes to the clear-speech benefit, the F0 range of clear-speech sentences was compressed to match that of the speakers' conversational-speech sentences. Fifteen listeners were presented with conversational, clear, and F0-compressed sentences in sustained speech-shaped noise. All talkers elicited substantial intelligibility benefits (keyword percent correct) from clear and F0-compressed speech when compared with conversational speech. There was no significant difference in performance between clear and F0-compressed speech. These results leave open the possibility that a clear-speech benefit could be a result of its F0 contours rather than its wide F0 range. Intelligibility predictions based on acoustic characteristics of clear speech, specifically high-frequency emphasis and pauses, accounted for either small or negligible amounts of the clear-speech benefit.

LSTM and CNN based ensemble approach for spoof detection task in automatic speaker verification systems

Nowadays, fingerprint and retina scans are the most reliable and widely used biometric authentication systems. Another emerging biometric approach for authentication, though vulnerable, is speech-based systems. However, speech-based systems are more prone to spoofing attacks. Through such malicious attacks, an unauthentic person tries to present himself as legitimate user, in order to acquire illegitimate advantage. Therefore, these attacks, created by synthesis or conversion of speech, pose an enormous threat to the reliable functioning of automatic speaker verification (ASV) authentication systems. The work presented in this paper tries to address this problem by using deep learning (DL) methods and ensemble of different neural networks. The first model is a combination of time-distributed dense layers and long short-term memory (LSTM) layers. The other two deep neural networks (DNNs) are based on temporal convolution (TC) and spatial convolution (SC). Finally, an ensemble model comprising of these three DNNs has also been analysed. All these models are analysed with Mel frequency cepstral coefficients (MFCC), inverse Mel frequency cepstral coefficients (IMFCC) and constant Q cepstral coefficients (CQCC) at the frontend, where the proposed ensemble performs best with CQCC features. The proposed work uses ASVspoof 2015 and ASVspoof 2019 datasets for training and testing, with the evaluation set having speech synthesis (SS) and voice conversion (VC) attacked utterances. Performance of proposed system trained with ASVspoof 2015 dataset degrades with evaluation set of ASVspoof 2019 dataset, whereas performance of the same system improves when training is also done with the ASVspoof 2019 dataset. Also, a joint ASVspoof 1519 dataset is created to add more variations into the single dataset, and it has been observed that the trained ensemble with this joint dataset performs even better, while performing evaluation using single datasets. This research promotes the development of systems that can cope with completely unknown data in testing. It has been further observed that the models can reach promising results, paving the way for future research in this domain using DL.

Clinical Importance of Sound-Field Speech Recognition Test for Measurement of Hearing Aid Benefit

Background and Objectives Although sound-field speech recognition test is an essential evaluation method to verify the benefit of hearing aids, the aided sound-field speech recognition test has not been used as a standard audiometric protocol. This study investigated whether a traditional measure of unaided speech recognition measured at an individual’s most comfortable level (MCL) can predict aided speech recognition performance and the benefit of hearing aids.Subjects and Method One hundred twenty-one ears (72 adult hearing aid users) with moderate, moderately severe, and severe degrees of hearing loss were tested. The word recognition score (WRS) and sentence recognition score (SRS) were measured at MCLs using earphones (earphone WRS and earphone SRS). From the sound-field speech audiometry, both unaided and aided WRS and SRS were evaluated at an average conversational speech level (at 65 dB SPL) using a loudspeaker (located at 0° azimuth, 1 m). This study determined the differences between earphone scores (at MCL) and aided recognition scores (at 65 dB SPL), which is known as earphone to aided difference (EAD). Also, the hearing aid benefit was simply derived by differences between the unaided and aided scores, both measured at 65 dB SPL.Results The differences between the unaided and aided scores were significantly lower with a higher degree of hearing loss, yet the earphone scores measured at MCLs did not differ across varied hearing loss. More than half of the tested ears had the EAD of greater than ±10%. The ears with a higher degree of hearing loss tended to have +EAD (earphone score>aided score), whereas the ears with less hearing loss showed -EAD (earphone score≤aided score). The earphone score measured at MCLs was not a predominant contributor to predict either aided performance or hearing aid benefit.Conclusion Clinicians should not assume that the unaided speech recognition score measured with earphones at MCL can accurately predict aided speech recognition or the benefit of hearing aids. Instead, it is crucial to directly assess aided sound-field speech recognition performance at a conversational speech level.

SPEECH TRANSFER PATTERNS IN FORMAL CLASSROOM

This study aims to explain the patterns of formal conversation speech transfer in the Law Studies program at Wisnuwardhana University of Malang. This study is a descriptive qualitative one-site study. The results of the study include transaction patterns in the initial, core, and final activities of learning. Conversation patterns are divided into (a) transaction patterns, (b) speech transfer patterns, and (c) movement patterns informal conversations. The conversation strategy is divided into initiation, negotiation, and an elicitation strategy. The function of conversation is divided into the function of declaring, asking, ordering, and expressing feelings. The use of negotiation strategies, speech expressing apologies, and expressions of gratitude are new phenomena informal conversations.
 
 <p> </p><p><strong> Article visualizations:</strong></p><p><img src="/-counters-/edu_01/0747/a.php" alt="Hit counter" /></p>

COMPARING HEARING AID FITTING OF THE MANUFACTURER'S FIRST FIT VERSUS NAL-NL2 USING REAL EAR MEASUREMENTS

Real ear measurments (REMs) or the probe microphone measurments is an objective method to verify hearing aids fitting taking into account the ear canal acoustics of the patient.(1) It is the best method to provide a patient with an audible signal.(2)Prescriptive fitting formulas, such as NAL-NL1, NAL-NL2, and DSL V5a are evidence-based methods for determining how much amplification is needed for a particular hearing loss. By using prescriptive formula verified with REMs at the initial hearing aid fitting allows the audiologist to ensure audibility for conversational speech, to set appropriate amounts of amplification for soft sounds, and to prevent loud sounds from becoming uncomfortable.(3) Some audiologists use the first-fit formulas of the manufacturer to have a proper fitting gain, however, many studies revealed that the first fit of the manufacturer is insufficient for the required gain.(4)Self-report outcome measure is also mandatory to determine the benefit of hearing aids.(5) The Abbreviated Profile of Hearing Aid Benefit questionnaire (APHAB) is used to measure disability caused by hearing loss and measure the reduction of that disability by using hearing aids. The APHAB uses 24 elements including 4 dimensions: ease of communication, reverberation, background noise, and aversiveness to sounds.(6)

Smart Glass Using IoT and Machine Learning Technologies to aid the blind, dumb and deaf

One of the earliest written records of a sign language is from the fifth century BC, in Plato’s Cratylus, where Socrates says: “If we hadn’t a voice or a tongue, and wanted to express things to one another, wouldn’t we try to make signs by moving our hands, head, and the rest of our body, just as dumb people do at present?”[Source: Wikipedia]. Technology is far more developed(advanced) than time then. Now, we can detect what the deaf and mute person is saying just by recording the same and comparing it with predefined datasets, revealing what the individual is trying to accomplish. In this paper, we have proposed a solution by which we can help disabled people, people who are Deaf, Dumb, or Blind by providing them a new technological eye, ear, and brain. Machine learning algorithms are used for object detection to provide the eye for the blind. Speech to text conversion to aid the deaf and text to speech conversion to aid the dumb while communicating. The amalgamation of all these technologies with IOT technologies will help resolve the issues related to these differently abled people.

Articulatory-to-Acoustic Conversion of Mandarin Emotional Speech Based on PSO-LSSVM

The production of emotional speech is determined by the movement of the speaker’s tongue, lips, and jaw. In order to combine articulatory data and acoustic data of speakers, articulatory-to-acoustic conversion of emotional speech has been studied. In this paper, parameters of LSSVM model have been optimized using the PSO method, and the optimized PSO-LSSVM model was applied to the articulatory-to-acoustic conversion. The root mean square error (RMSE) and mean Mel-cepstral distortion (MMCD) have been used to evaluate the results of conversion; the evaluated result illustrates that MMCD of MFCC is 1.508 dB, and RMSE of the second formant (F2) is 25.10 Hz. The results of this research can be further applied to the feature fusion of emotion speech recognition to improve the accuracy of emotion recognition.

Real-time Assistive Reader Pen for Arabic Language

Disability is an impairment affecting an individual's livelihood and independence. Assistive technology enables the disabled cohort of the community to break the barriers to learning, access information, contribute to the community, and live independently. This article proposes an assistive device to enable people with visual disabilities and learning disabilities to access printed Arabic material in real-time, and to help them participate in the education system and the professional workforce. This proposed assistive device employs Optical Character Recognition (OCR) and Text To Speech (TTS) conversion, using concatenation synthesis. OCR is achieved using image processing, character extraction, and classification, while Arabic speech synthesis is achieved through concatenation synthesis, followed by Multi Band Re-synthesis Overlap-Add (MBROLA). Waveform generation in the second phase produces vocal output for the disabled user to hear. OCR character and word accuracy tests were conducted for nine Arabic fonts. The results show that six fonts were recognized with over 60% character accuracy and two fonts were recognized with over 88% accuracy. A Mean Opinion Score (MOS) test for speech quality was conducted. The results showed an overall MOS score of 3.53/5 and indicated that users were able to understand the speech. A real-time usability testing was conducted with 10 subjects. The results showed an overall average of agreements scores of 3.9/5 and indicated that the proposed Arabic reader pen meets the real-time constraints and is pleasant and satisfying to use and can contribute to make printed Arabic material accessible to visually impaired persons and people with learning disabilities.

Prosody leaks into the memories of words

The average predictability (aka informativity) of a word in context has been shown to condition word duration (Seyfarth, 2014). All else being equal, words that tend to occur in more predictable environments are shorter than words that tend to occur in less predictable environments. One account of the informativity effect on duration is that the acoustic details of probabilistic reduction are stored as part of a word's mental representation. Other research has argued that predictability effects are tied to prosodic structure in integral ways. With the aim of assessing a potential prosodic basis for informativity effects in speech production, this study extends past work in two directions; it investigated informativity effects in another large language, Mandarin Chinese, and broadened the study beyond word duration to additional acoustic dimensions, pitch and intensity, known to index prosodic prominence. The acoustic information of content words was extracted from a large telephone conversation speech corpus with over 400,000 tokens and 6000 word types spoken by 1655 individuals and analyzed for the effect of informativity using frequency statistics estimated from a 431 million word subtitle corpus. Results indicated that words with low informativity have shorter durations, replicating the effect found in English. In addition, informativity had significant effects on maximum pitch and intensity, two phonetic dimensions related to prosodic prominence. Extending this interpretation, these results suggest that predictability is closely linked to prosodic prominence, and that the lexical representation of a word includes phonetic details associated with its average prosodic prominence in discourse. In other words, the lexicon absorbs prosodic influences on speech production.

Robust vowel region detection method for multimode speech

The aim of this paper is to explore a robust method for vowel region detection from multimode speech. In realistic scenario, speech can be classified into three modes namely; conversation, extempore, and read. The existing method detects the vowel form the speech recorded in clean environment which may not be appropriate for the multimode speech tasks. To address this issue, we proposed an approach based on continuous wavelet transform coefficients and phone boundaries for detecting the vowel regions from different modes of the speech signal. For evaluation of the proposed vowel region (VR) detection technique, TIMIT (read speech) and Bengali (read, extempore, and conversation speech) corpora are used. The proposed VR detection technique is compared to the state-of-the-art methods. The experiments has recorded significant gain in the performance of the proposed technique than the state-of-the-art methods. The efficiency of the proposed technique is shown by extracting vocal tract and excitation source features from automatically detected VRs for developing the multilingual speech mode classification (MSMC) model. The evaluation results report that the performance of the MSMC model is significantly improved when features are extracted from the vowel regions than the entire speech utterance.

VOP detection for read and conversation speech using CWT coefficients and phone boundaries

In this paper, we propose a novel approach for accurate detection of vowel onset points (VOPs). A VOP is the instant at which a vowel begins in a speech signal. Precise identification of VOPs is important for various speech applications such as speech segmentation and speech rate modification. Existing methods detect the majority of VOPs to an accuracy of 40 ms deviation, which may not be appropriate for the above speech applications. To address this issue, we proposed a two-stage approach for accurate detection of VOPs. At the first stage, VOPs are detected using continuous wavelet transform coefficients, and the position of the detected VOPs are corrected using phone boundaries in the second stage. The phone boundaries are detected by the spectral transition measure method. Experiments are done using TIMIT and Bengali speech corpora. Performance of the proposed approach is compared with two standard signal processing based methods as well as with a recent VOP detection technique. The evaluation results show that the proposed method performs better than the existing methods.

Many-to-Many Unsupervised Speech Conversion From Nonparallel Corpora

We address a nonparallel data-driven many-to-many speech modeling and multimodal style conversion method. In this work, we train a speech conversion model for multiple domains rather than a specific source and target domain pair, and we generate diverse output speech signals from a given source domain speech by transferring some speech style-related characteristics while preserving its linguistic content information. The proposed method comprises a variational autoencoder (VAE)-based many-to-many speech conversion network with a Wasserstein generative adversarial network (WGAN) and a skip-connected autoencoder-based self-supervised learning network. The proposed conversion network trains the models by decomposing the spectral features of the input speech signal into a content factor that represents domain-invariant information and a style factor that represents domain-related information to automatically estimate the various speech styles of each domain, and the network converts the input speech signal to another domain using the computed content factor with the target style factor we want to change. Diverse and multimodal outputs can be generated by sampling different style factors. We also train models in a stable manner and improve the quality of generated outputs by sharing the discriminator of the VAE-based speech conversion network and that of the self-supervised learning network. We apply the proposed method to speaker conversion and perform the perceptual evaluations. Experimental results revealed that the proposed method obtained high accuracy of converted spectra, significantly improved the sound quality and speaker similarity of the converted speech, and contributed to stable model training.

Intelligibility Enhancement Via Normal-to-Lombard Speech Conversion With Long Short-Term Memory Network and Bayesian Gaussian Mixture Model

Speech communications and interactions frequently occur in a variety of environments. Noise in the environment significantly degrades speech intelligibility when speaking and listening. Especially in the listening stage, even if the multimedia terminal outputs clean speech, it is still difficult for listeners to obtain information. Intelligibility enhancement (IENH) of speech is a technique for overcoming the environmental noise in the listening stage. It implements a perceptual enhancement of non-noisy speech. This study focuses on IENH via normal-to-Lombard speech conversion, inspired by a well known acoustic mechanism named the Lombard effect. Our method combines the long short-term memory (LSTM) network and Bayesian Gaussian mixture model (BGMM) to build a conversion architecture. Compared with baselines, it has three main advantages: 1) an LSTM network is used for spectral tilt mapping with fully considering short-term correlations and high-dimensional expression abilities; 2) the aperiodicity (AP) is mapped together with the fundamental frequency ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$F_0$</tex-math></inline-formula> ) by a BGMM, which considers their relevance constraints and the importance of APs; 3) the gender-dependent mapping is used for <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$F_0$</tex-math></inline-formula> and APs to consider distribution differences between genders. Experiments indicate that our method gets better performance in both objective and subjective tests.

Conversation Training Therapy: Let's Talk It Through.

This article introduces a novel approach to voice therapy called conversation training therapy (CTT). CTT is the first voice therapy approach to remove the therapeutic hierarchy common in most treatment programs. Rather, CTT uses patient-driven conversation as the sole stimuli in therapy to increase perceptual awareness of voice production in conversational speech. The genesis as to why CTT was developed, as well as the conceptual, theoretical, and component parts of CTT, will be explained. In addition, this article will offer examples of the language of therapy, as it applies to CTT and how to trouble-shoot if problems arise. Medical documentation relevant to CTT will also be outlined. Last, results from a recent efficacy study on CTT will be reported.

Text Reader

The historical and theoretical bases of contemporary high-performance text-to-speech (TTS) systems and their current designs are discussed. The major elements of TTS system are described, with particular reference to vocal tract models. The stages involved in the process of converting text into speech parameters are examined, covering text normalization, word pronunciation, prosodies, phonetic rules, voice tables, and hardware implementation. It is an image to text and speech conversion system developed for visually impaired as well as physically challenging people to be able to get information from the images easily. Core idea for image to text and speech conversion is to overcome the challenges faced by a visually impaired person in real life. The system goes through various phases such as image processing, text extraction, text-to-speech (TTS) conversion. This device opens the camera using the app into it out to the text and you get the audio. The primary motivation is to provide a visually impaired person with a friendly speech interface with computer and to allow such people who are physically and visually challenged to use the system to read printed text on the go.

Discursive Analysis of Hausa Language Use in a Kannywood Film: Basaja Gidan Yari

The paper studies Hausa films language through the analysis of communication strategies, namely figurative language (proverbs, metaphor), Code switching/Code mixing and Coded-language. The paper attempts to demonstrate how the characters’ use of these devices is deliberate, conscious. It shows that proverbs are commonly used as communication strategy not only to “embellish” but also to “convey deeper meaning” (Aduku, 2018) which helps reveal character competence and proficiency in the language. It also discovers that code switching/code mixing is used purposely as a communication strategy to structure speech in film conversation. As for coded message, it reflects a change in handling matters where communication is done in a restricted secret manner through coded message. Also, the examples extracted from the Hausa Film Basaja are to show that for proper understanding of the language, the surface meaning of words is to be examined.