Abstract

Due to the COVID-19 pandemic, the use of face mask was obligatory at all education levels in Greece during the school years 2020-21 and 2021-22. International research supports that concealment of the lower part of the face hinders the transmission of important oral and visual information from the speaker to the listener, impeding the recognition of speech as well as speaker identity and emotions, and thus reducing the efficacy of both verbal and non-verbal communication. The difficulty stemming from the use of face mask intensifies in environments with a high level of noise and reverberation, like the one found in most Greek classrooms, and can constitute a significant obstacle in communication and learning of children with normal hearing, but especially of children with hearing or/and learning difficulties. The aim of the present paper is to inform the education community of recent scientific findings involving the difficulties imposed by the use of face mask, discuss the effects of mask use in communication and learning, and propose possible remediation strategies.

Abstract

Previous studies have shown that noise has a detrimental effect on speech intelligibility and language comprehension, more so for children than adults. Furthermore, the mandatory use of face masks due to the COVID-19 pandemic presents an additional communication and learning barrier. The current study investigates the effect of face mask on low frequency word recognition in quiet and noise by adults and, more importantly, by young children who just completed the first grade of primary school. Preliminary results indicate that word identification is significantly compromised when produced with a surgical face mask, with more pronounced negative effects for children listening in noise. Implications on new word recognition and learning by first graders in noisy classrooms with compulsory mask usage are considered.

Abstract

The current study investigates voice quality characteristics of Greek adults with normal hearing and hearing loss, automatically obtained from glottal inverse filtering analysis using the Aalto Aparat toolkit. Aalto Aparat has been employed in glottal flow analysis of disordered speech, but to the best of the authors’ knowledge, not as yet in hearing impaired voice analysis and assessment. Five speakers, three women and two men, with normal hearing (NH) and five speakers with prelingual profound hearing impairment (HI), matched for age and sex, produced symmetrical /ˈpVpV/ disyllables, where V=/i, a, u/. A state-of-the-art method named quasi-closed phase analysis (QCP) is offered in Aparat and it is used to estimate the glottal source signal. Glottal source features were obtained using time- and frequency-domain parametrization methods and analysed statistically. The interpretation of the results attempts to shed light on potential differences between HI and NH phonation strategies, while advantages and limitations of inverse filtering methods in HI voice assessment are discussed.

Abstract

In this paper, we present distance-learning and assessment methods for engineering courses. The primary content of engineering courses involves mathematical equations, but there is an increasing need for drawing figures and tables on the spot, as well as demonstrating theory-in-practice using multimedia. In the present design, synchronous handwriting with a digital pen on pre-prepared slides presented on tablet-PCs was utilized in lectures via a videoconference platform in real-time, attempting to simulate the use of a blackboard while teaching in a physical classroom. The lectures are video-recorded and delivered to students, along with the annotated slides, for further study. Furthermore, a substantial delay in textbook delivery was introduced due to the pandemic that lead the authors to compile personal notes in PDF form to assist students in their study. Regarding course assessment, open book examination is adopted, with internet access allowed to all participants, while the assessment is invigilated via a videoconference platform. In addition, an automatic procedure is developed so that each student electronically receives a unique exam sheet, thus addressing the issue of academic dishonesty in online assessment pertaining wide student audiences. This is achieved by (a) randomizing the order of the exam questions using specific software, (b) including student serial number into the numerical data of each exam sheet, and (c) introducing a small variation in some of the exam questions. Advantages and disadvantages of the present distance-learning approach are discussed and statistics collected via online questionnaires are presented, suggesting that students show strong preference to the current course design over others, even over teaching in the physical classroom.

Abstract

This work involves the construction of a freely available sentence corpus in Modern Greek for research in audiology, speech technology and linguistics. Although such linguistic material has been developed in other languages, published work in Greek includes only lists of words and pseudowords. The present material has been based on the “Harvard/ΙΕΕΕ sentences” which are being used extensively in intelligibility tests in the English language. It consists of 720 sentences, each containing 5 to 9 words of one, two or -maximally- three syllables. The sentences are provided in Greek orthography and phonetic transcription, and statistical information regarding the material is also available (https://www.csd.uoc.gr/~asfakianaki/GrH.html).

Abstract

The paper investigates acoustic properties of the Greek voiceless plosives /p, t, k/, including the palatal allophone [c], by examining absolute and relative VOT and closure duration, relative burst intensity and spectral moments. Variability due to place of articulation, vowel context, gender and age is examined. The speech material comprised C1VC2V real words (C1=/p, t, k/, V=/i, a/, C2=dental/alveolar). Data from 12 adult speakers and 12 children (6 male and 6 female in each group) were analysed. Results showed that relative closure duration decreased and relative VOT duration increased in the order /p/, /t/, /k/ showing the anticipated inverse relationship reported in the literature. VOT was longer in the high vowel context for /t, k/. All spectral moments were significantly affected by place of articulation. Relative burst intensity was greater for the velar. Effects of gender and age were variable. Results are discussed in relation to theory and crosslinguistic evidence.

Abstract

Η καταγραφή της προφορικής ιστορίας συμβάλλει σε μέγιστο βαθμό τόσο στη διαφύλαξη της ιστορίας ενός λαού (Thompson, 1998) όσο και στη διατήρηση των διαλέκτων και ιδιωμάτων του. Ενώ η αξιοποίηση των διαλέκτων στη γλωσσική διδασκαλία προάγει τη μεταγλωσσική ενημερότητα και τον κριτικό γραμματισμό (Wolfram, 2003), το ελληνικό σχολείο δε φαίνεται να εκμεταλλεύεται τον διαλεκτικό πλούτο της ελληνικής γλώσσας (Καψάσκη & Τζακώστα, 2016, Μαγαλιά, 2000). Η δημιουργία του Ψηφιακού Μουσείου Ελληνικής Προφορικής Ιστορίας (ΨηΜοΠΙ) γεννήθηκε από την ανάγκη ενσωμάτωσης και αξιοποίησης των ελληνικών διαλέκτων στη γλωσσική διδασκαλία στο σχολείο (Τζακώστα, Σφακιανάκη, Πατσιάς, Κλειδής & Τσιγδινός, 2015). Το παρόν εργαστήριο έχει σκοπό να ενημερώσει και να ευαισθητοποιήσει τους συμμετέχοντες αναφορικά με την αξία των νεοελληνικών διαλέκτων και της προφορικής ιστορίας, να τους παρέχει πρακτικές οδηγίες και γνώσεις σχετικά με το σχεδιασμό ενός project προφορικής ιστορίας στο σχολείο, καθώς και βασική κατάρτιση σε δεξιότητες που αφορούν στην καταγραφή, επεξεργασία και επισημείωση αρχείων ομιλίας με εξειδικευμένα λογισμικά.

Abstract

Το εκπαιδευτικό εργαλείο «Μιλώ Ελληνικά-SpeakGreek» είναι ένα σύγχρονο σύστημα διδασκαλίας της προφορικής ελληνικής γλώσσας με ελεύθερη πρόσβαση στο διαδίκτυο. Το “SpeakGreek” μπορεί να χρησιμοποιηθεί στην εκπαίδευση μαθητών της ελληνικής ως δεύτερης/ξένης γλώσσας καθώς και ατόμων με προβλήματα ακοής και ομιλίας. Η ομιλία του μαθητή οπτικοποιείται στην οθόνη του Η/Υ σε πραγματικό χρόνο (βιοανατροφοδότηση) και συγκρίνεται με ηχογραφημένα δεδομένα από άλλους ομιλητές των ελληνικών (άνδρες, γυναίκες και παιδιά), με αποτέλεσμα να επιτυγχάνεται ευκολότερα η βελτίωση διαφόρων παραμέτρων της ομιλίας του. Είναι φιλικό στη χρήση και μπορεί να χρησιμοποιηθεί είτε αυτόνομα από τους μαθητές είτε με παράλληλη βοήθεια από τον δάσκαλο, το λογοθεραπευτή ή τον γονέα. Το «Μιλώ Ελληνικά-SpeakGreek» είναι το πρώτο εργαλείο στο είδος του για την ελληνική γλώσσα.

Abstract

Hearing-impaired speech presents a number of characteristics deviating from normal, which concern both the segmental level, i.e. vowel and consonant errors, and the suprasegmental level, i.e. stress, intonation, resonance, voice quality etc. Errors on both levels can compromise speech intelligibility, that is, the ability of being understood. The current paper presents a qualitative analysis of the main segmental and suprasegmental errors observed in the speech of 10 Greek adult speakers with profound prelingual hearing impairment and of various intelligibility levels. Audio files and/or spectrographic displays, useful to clinicians, teachers and researchers, are provided to illustrate the above features and to help clarify acoustic measurement issues pertaining to the acoustic analysis of hearing-impaired speech.

Abstract

In this paper we describe, first, the goals of the Digital Museum of Greek Oral History (diMOHI) which intends to cover the need for the involvement of pupils in the educational procedure, and, second, the ways the goals of diMOHI are materialized. Aim of the research program presented here is the design, the construction and the organization of a museum of oral history, which will be used for scientific and educational purposes. As a result, the diMOHI will be directed to pupils, educators and anyone interested in dialectology, education, more specifically language teaching. The diMOHI will take the shape of a digital platform where audiovisual data will be available. These data will be collected by Primary and High school pupils who will be trained on methods of data collection, indexation and processing. The audiovisual material will consist primarily of dialectal language data from areas of Western Crete, however, the museum will be designed in such a way so that it can host dialectal material from various areas of Greece. In addition, the audiovisual data together with language activities whose design will be based on the data will be used for the (combinatorial) teaching of standard Greek and its dialectal variants.

Abstract

Dialectal variants are complete linguistic systems just like standard languages (cf. Kontosopoulos 1997, Ntinas & Zarkogianni 2009). The teaching of different linguistic varieties of a standard language gives pupils the possibility a) to be acquainted with the treasures of the expressive means of their mother language, b) to embody the mother language in a broader cultural and historical context and c) to discover the grammatical adjacency of the dialects and the standard language. We believe that the above can be maximally achieved if there is active interaction between pupils and language teachers/ educators during the process of language teaching.

Aim of the project presented here is the design, construction and organization of the Digital Museum of Greek Oral History (DiMuGOHi) as a research and educational tool available to pupils and language teachers/ educators. Goals of the project are a) the training of primary and high school pupils in methods of language data collection and processing and, b) the collection, processing, filing and preservation of dialectal linguistic data, and, c) the design of authentic didactic/ pedagogical material to be used for the teaching of dialectal varieties in combination with the standard language. The DiMuGOHi will have the form of a digital platform in which all the materials described above will be available. We illustrate, on the one hand, the major axes on the bases of which the platform is designed, and, on the other hand, the CLIL principles underlying the suggested pedagogical/ didactic activities.

The DiMuGOHi is expected to contribute to the educational procedure in direct and indirect ways. A first direct outcome is the pupils’ sensitization regarding dialects, their linguistic properties and structural adjacency with standard Greek as well as the role of the dialects in the preservation and diffusion of local and national cultural heritage. In addition, the DiMuGOHi will accentuate issues of environmental education, social sciences, geography, local history (cf. Thompson 1998, 2000). Some indirect outcomes are that, first, it will facilitate the improvement of metalinguistic awareness regarding the Cretan dialect, second, it will enhance knowledge that dialects have complete linguistic systems, just like the standard language, third, it will improve pupils’ stylistic and sociolinguistic awareness, namely the conscious knowledge of the linguistic contexts in which dialectal material is used, and, fourth, it will preserve the linguistic treasures of different linguistic varieties of the standard language.

Abstract

This paper presents the structure and first applications of SpeakGreek, an online biofeedback speech training tool that can be used in second/foreign language education and in clinical intervention for individuals with speech disorders. The tool provides training in the perception and production of key segmental and suprasegmental aspects of Greek. The development of the tool is based on the analysis of acoustic and articulatory data of Greek and on the use of speech technology applications including automatic speech recognition.

Abstract

In many intelligibility studies, it was demonstrated that the speaking style referred to as clear speech is significantly more intelligible than conversational (or casual) speech. This intelligibility gain exists for both normal-hearing and hearing-impaired listeners (e.g. elderly persons and linguistically inexperienced listeners like non-native (L2) speakers and children). Also, in a two-way conversation in which one person is affected by an adverse listening condition and one is not (e.g. between one person speaking to another via telephone where the other is in a noisy club, or in a cafeteria, in the street etc.), the person who is not affected still manages to make adaptations (on acoustic-phonetic and linguistic levels) that are quite specifically tailored to counteract the specific communication barrier that the other person is experiencing. These adaptations show that clear speech is not defined in a uniform way, but that there are different styles of clear speech depending on the adverse condition that the speech is heard in. In this context, Active Speech Modifications refer to the speaking-style adaptations or strategies a speaker applies in order to maximize communication effectiveness.

Identification and effective manipulations of the most prominent acoustic-phonetic characteristics of different styles of clear speech can allow for the development of new, signal based, active speech modification algorithms to increase intelligibility. The algorithms can consequently improve speech intelligibility in many situations, such as in the design of hearing aids, telephony, and other speech signal processing technologies and applications (i.e., speech synthesis, recognition, enhancement, etc).

The purpose of this project was to use modern speech analysis and reconstruction algorithms to:

• identify which acoustic-phonetic characteristics are prominent in different styles of clear speech (e.g. babble-countering clear speech, vocoder-countering clear speech, L2-“countering” clear speech) and when they are realized in time.
• model a selection of these aspects so that they can be applied automatically on speech, to enact prosodic changes, changes in amplitude spectrum, modulation frequencies, etc.
• run a series of “proof of concept” perception experiments to see if the “specifically-enhanced” speech is better perceived in the “matched” adverse condition than other types of clear speech (there is evidence that this is the case with the naturally-enhanced speech).

The outcome of the project can be summarized as follows:

• a new speech corpus (P8-Harvard corpus) was linguistically and meta-linguistically annotated and acoustically analyzed with the goal of identifying which acoustic-phonetic characteristics differ between clear and casual speech and also between different styles of clear speech. Moreover, acoustic analyses on specific features were also performed on a different corpus, namely the LUCID database (specifically on read clear and read casual speech signals).
• among the different styles of clear speech, prosodic changes were most apparent. Therefore, signal modification algorithms were developed to mimic human adaptations on prosody in adverse conditions with the aim of increasing intelligibility.
• a user-friendly interface, XPlic8, for a large range of acoustic analyses was developed.
• a set of evaluation experiments was prepared to evaluate the different modifications.

Abstract

Research on other languages, mainly English, has focused on coarticulatory aspects in the productions of hearing impaired/deaf individuals (Waldstein & Baum, 1991; Okalidou & Harris, 1999; Ryalls et al., 2003, etc). For Greek, this is an area yet to explore. Electropalatography has been used to investigate consonant production by Greek hearing impaired (HI) speakers (Nicolaidis, 2004; 2007) and some first data on the acoustic characteristics of vowels produced by Greek speakers with HI have been presented (Nicolaidis & Sfakianaki, 2007). This study examines first formant vowel-to-vowel coarticulatory effects in /pVCV/ disyllables (V= [i], [a], [u], C= [p], [t], [s]) produced by five men and four women with profound hearing loss (>95 dB) and five normal hearing adults (control group). Bidirectional coarticulatory effects are examined as a function of hearing level, gender, vocalic and consonantal context and stress. Acoustic studies on coarticulation have relied on F2 rather than F1 data, although the latter can shed light on tongue dorsum and jaw height information (Recasens & Pallarès, 2000). Preliminary results reveal different coarticulation patterns due to contextual variability for speakers with HI which may indicate differences in their gestural organisation (Okalidou & Harris, 1999), and are discussed in the light of articulatory constraint (Recasens & Espinosa, 2009) and speech production theories.

Abstract

The study examines F1, F2 formant frequencies and duration of all five Greek vowels produced by six Greek speakers with profound hearing impairment and six speakers with normal hearing (three male and three female in each group). The speech material analysed was of the form /'pVCV/ where V=/ i, ɛ, ɐ, ɔ, u /, C=/p, t, k, s/. The study discusses differences in the above acoustic parameters as a function of hearing level, gender, stress and context. The results show longer vowel durations and a reduction of the vowel space for the speakers with hearing impairment. Significant variability due to stress and context was evident between the two groups. The paper discusses findings with reference to perceptual constraints affecting the speech of individuals with hearing impairment.

Abstract

Our goal is to examine the VOS order (e.g. efaje tin turta o Janis, ate- 3sg the cake-acc the John-nom) from an experimental point of view and thus to investigate to what extent the claims made so far on a theoretical basis about this particular order (see Alexiadou 1997, 1999; Philippaki-Warburton 2001; Haidou, 2000; Georgiafentis 2001) can be supported or not by our experimental analysis. At the same time, we intend to direct our attention to the properties of the VOS order with respect to information structure. In particular, we are interested in finding out which constituent/s of the VOS order tend/s to be more frequently part of focus and thus new information.

Abstract

To my knowledge, no research on the acoustic analysis of Greek children’s speech has been published. At the same time, the need for an expanded knowledge base, that would include information on children, is underlined in the literature (Deterding, 1990, Kent & Read, 1992, Clark & Yallop, 1995, Henton, 1995, Lass, 1996, Lee et al, 1997). This study, in an attempt to acknowledge this need, involves the acoustic analysis of Greek men’s, women’s and children’s vowels, with an emphasis on the last, and examines the relationship between adults’ and children’s acoustic data. Children’s vowels showed higher formant frequencies than adults’, as expected from English children’s data, and the scattergrams of adults’ versus children’s data revealed a relation that can be represented by a linear model quite satisfactorily.

Abstract

Το SPECO είναι ένα ευρωπαϊκό πρόγραμμα υπο εξέλιξη (INCO-COPERNICUS, Αριθ. Προγρ. 977126, 1998-2001) το οποίο αναπτύσσει ένα οπτικο-ακουστικό σύστημα για την εκμάθηση της προφοράς της γλώσσας και την εξάσκηση στην παραγωγή λόγου από παιδιά με ακουστική ή/και γλωσσική αναπηρία. Τα δύο κύρια μέρη του προγράμματος υποστηρίζουν την παραγωγή φωνηέντων και συμφώνων, ξεκινώντας από ασκήσεις έντασης ήχου και διακυμάνσεις του τόνου, και καταλήγουν σε εξάσκηση συλλαβών, λέξεων και φράσεων. Το SPECO αναπτύσσεται στα ουγγρικά, αγγλικά, σλοβενικά και σουηδικά, αλλά μπορεί να προσαρμοστεί σε οποιαδήποτε ευρωπαϊκή γλώσσα μέσω ενός γλωσσικά ανεξάρτητου συστήματος μετρήσεων (database editor). Είναι μέσα στους στόχους μας να αναπτύξουμε το SPECO και στα ελληνικά. Η βασική ιδέα του προγράμματος μπορεί επίσης να αναπροσαρμοστεί για τη δημιουργία ενός μέσου διαδασκαλίας της προφοράς της ελληνικής ως δεύτερης ή ξένης γλώσσας.

Abstract

The idea of computer systems which exploit developments in speech technology for pedagogical or remedial purposes is not new, and a number of such systems have been developed. The SPECO Project is attempting to produce a new system which uses advanced speech technology and is adaptable to different languages. Although its primary purpose is for clinical remediation of children’s speech problems, we believe that it has additional potential in pronunciation teaching.

Abstract

The SPECO Project was funded by the EU through the INCOCOPERNICUS program (Contract no. 977126) in 1999. In the frame of the project a system has been developed which is an audio-visual pronunciation teaching and training system for 5-10 year old children. Correction of disordered aspects of speech is done by real time visual presentation of the speech parameters, in a way that is understandable and interesting for young children, while remaining correct from the acoustic-phonetic point of view. The development of the speech by our teaching method is made mainly on the basis of visual information using the intact visual channel of the hearing impaired child. However during practice we use their limited auditory channel too, by giving auditory information synchronised with the visual one. This multi modal training and teaching system has been developed for languages of all the SPECO partners; these are English, Swedish, Slovenian and Hungarian.