Japanese Speech Research Articles

Acoustic evaluation of voice signal distortion by videoconferencing platforms and devices used in telepractice for cleft palate.

The usefulness and effectiveness of telepractice have been reported in recent years. Treatment of cleft palate patients with compensatory articulation is based on perceptual identification. Telepractice using videoconferencing platforms causes voice signal distortion and impacts auditory-perceptual perception. This study aimed acoustically examine voice signal distortion and determine the optimal videoconferencing platforms, in addition to the phonemes that can be discriminated with the same quality as in face-to-face interactions. ATR503 with 50 phoneme-balanced Japanese speech sentences was used as a reference corpus. Four videoconferencing platforms, -Zoom, Cisco Webex, Skype, and Google Meet, -and five devices, -iPhone, Android, iPad Air, Windows, and MacBook Pro were used as transmission conditions to examine voice signal distortions with the objective measure log-spectral distortion (LSD). Tukey's test was conducted to evaluate the degree of consonant distortion related to voicings (voiceless and voiced), places of articulation (bilabial, alveolar, alveolo-palatal, palatal, velar, labial-velar, and glottal), and manners of articulation (plosive, fricative, affricate, tap or flap, nasal, and approximant). With statistically significant differences, voiced, bilabial, labial-velar, nasal, and plosive consonants exhibited smaller distortions. In contrast, voiceless, alveolo-palatal, fricative, and affricate consonants exhibited larger distortions. Google Meet exhibited the lowest distortion among videoconferencing platforms and MacBook exhibited the lowest distortion among devices. This study provides significant insights into the telepractice strategies with the appropriate videoconferencing platform and device, and useful settings for cleft palate patients with compensatory articulations with respect to acoustics.

Articulatory-based Phonemic Paraphasia in Conduction Aphasia: A Dysfunction in Phoneme-to-Articulation Conversion Uncovered Through Crossed Aphasia.

Phonemic paraphasia, a common characteristic of conduction aphasia, has traditionally been attributed to phonological representation dysfunction. An alternative hypothesis posits that phonemic paraphasia arises from difficulty converting phonemes into their corresponding articulatory maneuvers. However, detailed case studies supporting this theory have been lacking. In this report, we present the case of a 61-year-old right-handed man with right temporo-parietal infarction who exhibited crossed aphasia characterized by typical conduction aphasia symptoms (eg, relatively fluent speech with intact comprehension, frequent phonemic paraphasia, and pronounced difficulties in oral repetition) in the absence of distorted articulation, syllable segmentation, and prosody impairment. Despite the frequent occurrence of phonemic paraphasia and articulatory challenges, our patient's phonological representations remained relatively intact. His phonemic paraphasia was often self-corrected to produce correct responses, a feature known as conduit d'approche. During the oral repetition of individual mora (ie, the smallest unit of speech in Japanese), we observed that the patient consistently traced the corresponding Hiragana phonetic symbol accurately, despite his difficulties in articulation. We substantiated this phenomenon through objective assessment and posit that it resulted from an unusual separation of language functions in crossed aphasia-specifically, a disconnection between phonological representations in the right temporo-parietal cortex and speech articulation engrams in the left hemisphere. In this case of conduction aphasia, articulatory-based phonemic paraphasia may be viewed as an inability to convert phonemes into the appropriate articulatory maneuvers rather than as phonological representation dysfunction or apraxia of speech.

Recognition of target domain Japanese speech using language model replacement

End-to-end (E2E) automatic speech recognition (ASR) models, which consist of deep learning models, are able to perform ASR tasks using a single neural network. These models should be trained using a large amount of data; however, collecting speech data which matches the targeted speech domain can be difficult, so speech data is often used that is not an exact match to the target domain, resulting in lower performance. In comparison to speech data, in-domain text data is much easier to obtain. Thus, traditional ASR systems use separately trained language models and HMM-based acoustic models. However, it is difficult to separate language information from an E2E ASR model because the model learns both acoustic and language information in an integrated manner, making it very difficult to create E2E ASR models for specialized target domain which are able to achieve sufficient recognition performance at a reasonable cost. In this paper, we propose a method of replacing the language information within pre-trained E2E ASR models in order to achieve adaptation to a target domain. This is achieved by deleting the “implicit” language information contained within the ASR model by subtracting the source-domain language model trained with a transcription of the ASR’s training data in a logarithmic domain. We then integrate a target domain language model through addition in the logarithmic domain. This subtraction and addition to replace of the language model is based on Bayes’ theorem. In our experiment, we first used two datasets of the Corpus of Spontaneous Japanese (CSJ) to evaluate the effectiveness of our method. We then we evaluated our method using the Japanese Newspaper Article Speech (JNAS) and CSJ corpora, which contain audio data from the read speech and spontaneous speech domain, respectively, to test the effectiveness of our proposed method at bridging the gap between these two language domains. Our results show that our proposed language model replacement method achieved better ASR performance than both non-adapted (baseline) ASR models and ASR models adapted using the conventional Shallow Fusion method.

Интонационные ошибки в японской речи русских обучающихся на примере вопросительных высказываний

The article presents a description of different types of interrogative intonation errors in the Japanese speech of L1 Russian learners of Japanese. The material for the research includes recordings of Japanese interrogative phrases uttered in spontaneous dialogues by speakers whose native language is Russian. Acoustic analysis of these realizations has revealed a few intonation errors, which may be considered when developing a national-oriented teaching approach for learners of Japanese intonation.

Testing an acoustic model of the P-center in English and Japanese.

The notion of the "perceptual center" or the "P-center" has been put forward to account for the repeated finding that acoustic and perceived syllable onsets do not necessarily coincide, at least in the perception of simple monosyllables or disyllables. The magnitude of the discrepancy between acoustics and perception-the location of the P-center in the speech signal- has proven difficult to estimate, though acoustic models of the effect do exist. The present study asks if the P-center effect can be documented in natural connected speech of English and Japanese and examines if an acoustic model that defines the P-center as the moment of the fastest energy change in a syllabic amplitude envelope adequately reflects the P-center in the two languages. A sensorimotor synchronization paradigm was deployed to address the research questions. The results provide evidence for the existence of the P-center effect in speech of both languages while the acoustic P-center model is found to be less applicable to Japanese. Sensorimotor synchronization patterns further suggest that the P-center may reflect perceptual anticipation of a vowel onset.

A Study of the Assistive Nature of Artificial Intelligence Technology for Japanese Translation and Interpretation

Abstract Traditional Japanese translation methods have certain disadvantages, and the introduction of artificial intelligence technology into them can enhance the effect of Japanese interpretation and translation. In this paper, the Japanese language data of Twitter and Facebook are used as the basis to construct a Japanese language interpretation and translation corpus, and the GPT-2 model is constructed on the basis of Transformer for Japanese text translation. To optimize the Seq2Seq model for Japanese speech interpretation, the Attention mechanism is introduced to establish a Japanese speech translation model. A Japanese oral and written corpus was used to analyze the validity of the methods mentioned above. The results show that the class/form ratio in the Japanese oral/translated corpus fluctuates between [0.1231, 0.1448], but the survival rate of borrowed words under the scientific category reaches the highest of 54.14%, and the average number of occurrences of each word is between [4.35, 4.95]. Japanese verbal and translated texts had an average sentence length of 40 hours, and their translation accuracy was approximately 74.16%. The quality of translation can be effectively improved, and cultural exchange between China and Japan can be enhanced by integrating AI technology with Japanese interpretation and translation.

Analysis of Potential Mistakes of Russian-Speaking Students When Mastering the Use of Honorific Speech in Japanese, Taking Aspectual Relations into Account

Analysis of Potential Mistakes of Russian-Speaking Students When Mastering the Use of Honorific Speech in Japanese, Taking Aspectual Relations into Account

How many languages can be broadcast simultaneously as an emergency announcement?

In spaces where people from various countries come and go, voice guidance is often broadcast sequentially in multiple languages. However, this method is disadvantageous to users of the language broadcast later, and this problem can be particularly severe in evacuation broadcasts. In addition, the broadcast time may be too long, which may increase the annoyance/noisiness and spoil the soundscape. We propose the use of simultaneous multilingual broadcasts as a challenging method of achieving language barrier-free and saving the time required for information transmission. In the present study, we examined how many languages can be broadcast simultaneously to clarify the feasibility of the method. The intelligibility of simultaneous broadcasts of up to four languages (Japanese, English, Chinese, and Korean) was examined through listening experiments. Participants were instructed to write down Japanese speech only. The experimental parameters were the arrival direction of the target/distracting speech and the background noise level. The results of the experiments suggested that broadcasting two languages simultaneously is sufficiently practical, and even three languages can be practical under certain conditions.

Strategi Pemagaran dalam Ujaran Bahasa Jepang: Analisis Wawancara Jalanan Kanal Youtube Ask Japanese

This study reviews the hedging phenomena in Japanese speech found in street interview Ask Japanese. The hedging can be seen in interviewees’ answers to the interviewer’s questions. The data source used in this study is videos uploaded to Youtube by channel Ask Japanese. The reason for choosing such data source is that the videos are recorded impromptu, and thus, they show the realization of hedging in utterance naturally without any make-up. There are, in total, two videos of interview that is analyzed. The theories used are hedging as semantic phenomenon and hedging as pragmatic phenomenon. This research employs qualitative method that is involving analyzing data that have been transcribed and codified beforehand. In the videos used, there are 234 cases of hedging found. After keen observation, it is revealed that those cases of hedging are not realized in the same way. They can be classified into 6 types of strategies according to their realization in speech, namely (i) hedging by emphasizing subjectivity, (ii) hedging by expressing uncertainty, (iii) hedging by creating common ground, (iv) hedging by approximation, (v) hedging by exemplification, and (vi) hedging by downtoning.

Two efforts towards natural speech synthesis: Incorporating disfluency and speaking style change based on the interlocutor

During the period 2000–2005, a Japanese female speaker recorded her everyday conversations with many different interlocutors using a head-set microphone. As a result, 600 hours of natural Japanese speech data were obtained. This study describes a DNN-based speech synthesis system which was trained on 300 hours of the data, focusing on two unique efforts to make it more expressive in a human-like way: (1) allowing for disfluencies, and (2) accounting for the category of interlocutor. Incorporating some frequently observed disfluent patterns in general Japanese speech such as fillers, phrase-final rising intonation, and word-internal prolongation or suspension, is believed to be effective in practical application as certain disfluencies are connected to a speaker’s attitude in Japanese communication. For example, having word-internal prolongations can show hesitation or politeness, and word-internal suspending can show the speaker’s surprised attitude. Interlocutors in the original data were categorized into four groups: family, friend, child and others. This information was used in the training process, and as a result, the synthesizer can generate different speaking styles according to the interlocutor setting. Being able to generate disfluent speech and change the speaking style depending on who you are talking to can make the synthesizer ever more expressive.

Directive acts as part of a quotation in Modern Japanese speech

In this paper, we consider directive acts as part of a quotation in Modern Japanese speech. Using the texts of electronic diaries as a kind of written expression of conversational discourse, we studied the ways of transmitting directive acts as part of a quotation, as well as to analyzed how the choice of a certain form of expression of directive acts as part of a quotation may indicate the direct or indirect nature of the citation. After giving a preliminary theoretical picture, we turn mainly to the direct (morphologized) imperative forms of the Japanese language. These forms are enough to show the complexity of the directive category in the "polite" language of the inhabitants of the Land of the Rising Sun. In any case, the analysis of morphologized forms inevitably moved to the syntactic-discursive level. The peculiarity of the approach is in the analysis of several texts by one author to determine the language techniques characteristic of this author, which is used, among other things, as an argument in favor of direct or indirect citation, since Japanese (as well as for a number of other languages) is characterized by a rather complex, subtle distinction between direct and indirect ways of transmitting someone else’s speech. The material convinces us that only a complete analysis of both the quotation itself and the context (for example, texts by the same author) allows us to get a more accurate picture of the phenomenon of the distinction between direct and indirect citation. Among other things, the conducted research has the theoretical significance in terms of thinking about the phenomenon of quoting in language in general. It is clear that this phenomenon (especially in the light of modern information-polymodal, postmodern and other trends, in particular trends in the psychologization of discourse) should be understood more difficult, as new research in this field proves, and as our paper shows

Listening evaluation of cochlear implant users: comparison of subjective and objective evaluation by visual analogue scale.

This study aimed to use short-form visual analogue scale cochlear implantation questionnaires to evaluate subjective aspects at each out-patient visit. The correlation between subjective hearing tests using the short-form visual analogue scale and objective hearing outcomes was evaluated. This study was conducted in a single centre. Cochlear implant users (n = 199) evaluated their hearing on a scale of 0 to 100 for the right, left and both ears. The Japanese speech perception test (CI-2004) Japanese monosyllable speech perception test (67-S) and cochlear implantation threshold were used for the objective cochlear implantation evaluation. A significant correlation was found between the short-form visual analogue scale questionnaire and objective hearing outcome, for words (r = 0.64) and sentences (r = 0.62) in CI-2004 and 67-S (r = 0.56) tests. No significant correlation was found between the short-form visual analogue scale score and cochlear implantation threshold (r = -0.18). Short-form visual analogue scale cochlear implantation questionnaires mean cochlear implant users spend less time answering subjective visual analogue scale questionnaires, and clinicians estimate a patient's cochlear implantation hearing and abnormality by chronological evaluation.

Sound signal analysis in Japanese speech recognition based on deep learning algorithm

Sound signal analysis in Japanese speech recognition based on deep learning algorithm

Phenomena of Old-Written Japanese in a Modern Newspape

Modern Japanese uses elements of Bungo, an old written Japanese language created a thousand years ago. These elements have already entered its grammatical system and are used in written speech in modern Japanese. At the same time, these phenomena are present in different forms, i.e. some are used freely and are taken into account in the description of the modern Japanese language, others have been lexicalised, and some have been preserved as part of phraseological units. The article aims at highlighting and describing bungo phenomena in the texts of modern Japanese newspapers for further use in grammar teaching and revision. This is necessary to achieve full and quality education in the process of teaching Japanese language in the framework of the implementation of the Federal Target Programme of Education Development

Экспрессивные средства языка в речи глав государств Японии и России (на материале обращений к нации в период пандемии COVID-19)

The paper considers the problem of the political discourse of Japan and Russia. The significance of the coverage of this issue is due to new social threats, which include the worldwide COVID-19 pandemic. The aim of the study is to identify similarities and differences in expressive language means in the speeches of the Japanese and Russian heads of state during the COVID-19 pandemic (Shinzo Abe (2020), Yoshihide Suga (2021), Fumio Kishida (2022), V. V. Putin (2020-2022)). The study is novel in that it is the first to identify the stylistic markedness of the speech of Japanese and Russian top officials using the material of the official addresses made during the COVID-19 pandemic. As a result of the study, it can be argued that in the official addresses of the Japanese and Russian heads of state, metaphors, epithets, set expressions, rhetorical questions are more often used, which makes the speech expressive, emphasizes its imagery, emotionality and evaluativeness. The speeches of the Prime Ministers of Japan are characterized by a more neutral connotation compared to the speeches of the President of the Russian Federation V. V. Putin, whose speech is dominated by the use of expressive means.

A segmental analysis of the Japanese negation verb variation: A phonological perspective

Within the field of Japanese linguistics, the topics on segmental analysis of verb variation have rarely been studied from the phonological perspective. On the other hand, verb modifications have been frequently observed to occur in oral communication. This study addresses a segmental analysis within the variation of Japanese negation verb from the phonological standpoint. It employed a qualitative descriptive approach with the data gathered from the Japanese animated series Orange (Hamasaki Nakayama, 2016). Excerpts of the data were in the form of conversation fragments, which were analyzed using the articulatory phonetic identity technique. The analysis resulted in five found variations: -nai, -nnai, -nee, -nnee, and -ai deletion. The presence of sounds in the phonological environment caused phonological variation, which affected the internal structure of the verb. As a result, the phonological rules are applied to showcase the phonological process in the verb. It was also found that segmental processing does not necessarily reduce the amount of mora in verbs. Segmental processes, on the other hand, can change the mora structure without reducing the mora. Additionally, the current study shows vowel coalescence associated with the male speech in Japanese.

ГЕНДЕРНАЯ СТРАТИФИКАЦИЯ ОБРАЩЕНИЙ-СИНТОНОВ В ЯПОНСКОЙ ОНЛАЙН КОММУНИКАЦИИ

In this article, from the point of view of gender and pragmalinguistics, a comprehensive lingua-statistic analysis of Japanese online communication is carried out. The research problem is updated by the continuous development of both global online communication and Japanese online communication, as well as existing manifestations of gender identity in it. The main goal of the study is to identify addresses-syntones in Japanese online communication and their gender stratification. Through the use of the continuous sampling method, the statistical data of the study were collected from various social networks and video hosting sites of the Japanese Internet space. Using the lingua-statistic method, the data were divided into several groups, reflecting the main patterns of using addresses in online communication: the absence of an address in the message; the presence of an address in the message; hyperlink as an address. Subsequently, from the group “The presence of an address in the message”, according to the gender of the addressers, the statistical data were distributed into several subgroups: men, women and anonymous. In order to identify their deepest components, descriptive, communicative-pragmatic and structural-noematic methods were also used. As a result of the study, it was revealed that addresses-syntones in Japanese online communication are a fairly low-frequency component of communication that creates a favorable environment for interaction. They reflect the gender component of Japanese speech, which is manifested in the preference for certain forms of address, as well as in the variety of these forms depending on the gender of addressers and addressees.

Japanese Accent Pronunciation Error by Japanese Learners in Elementary and Intermediate Level

This research discusses errors in the pronunciation of Japanese accents by Japanese learners. This study aims to determine the forms and causes of accents pronunciation errors in Japanese learners as well as Japanese learners’ comprehension of Japanese accent. The data used are speech data (word and sentence) of 18 people who are students of Japanese literature at a private university in Jakarta, in their 3rd and 4th years (hereafter written as respondents). As comparison data, this research uses Japanese native speaker data from Tokyo and electronic speech data from OJAD (Online Japanese Accent Dictionary). This research used a qualitative research method. Praat application also used to get a detailed visual description of Japanese speech from respondents. The results of this research found a form of error in the form of accent changes that do not match the speech data per word and sentence. The cause of accent errors was the lack of attention to the accent and the lack of learning about accents in the respondent’s learning environment. It was also found that 80% of respondents had minimal understanding of Japanese accents.

A Study on Particles used with Adverbs in Japanese

Japanese is a language among Transeurasian languages embedding a substantial number of similarities to Turkish; however, there appear many differences as deeper analyses are carried out. It bears outstanding features both in terms of the differences in the writing system and in terms of the structural differences of the elements used in sentences. As a matter of fact, from a grammatical point of view, there are some cases where some terms in Turkish do not exactly overlap their Japanese counterparts.
 The number of Japanese linguists who argue that the Turkish equivalent of the words called joshi (particle), which is one of the parts of speech in Japanese, does not exactly correspond to the part of speech, which is the ilgeç (edat), is increasing rapidly. There are also researchers categorizing this part of speech as a noun case suffix instead of ilgeç (edat) because it is affixed to nouns in Turkish. From this perspective, in order to remain in the realm of Japanese studies, to refer to the grammatical structure which is the subject of this study, the term ilgeç (edat) (particle) is employed in the present research.
 The presence of particles which are frequently used in sentences patterned with adverbs「adverb + particle(s)」shows that they are part of a formally different formation in the form of adverbs in Japanese. In this research, the forms of adverbs in this structure existing in sentences are discussed by analyzing the reading texts and sample sentences in various intermediate level grammar books. The aim of the study is to determine, in general terms, which of the particles of this type are used with adverbs and which cannot be used with adverbs in Japanese. Thanks to the light shed by the findings obtained, it was understood that the particles「と(to)」and「に(ni)」were the particles most frequently used as a pattern with the adverbs in Japanese. Moreover, the other particles used with adverbs are shown in tables.

Contrastive Analysis of Errors Analysis in the Use of Japanese and Sundanese Speech Level Categories for Themselves (Kenjougo and Lemes Keur Sorangan)

This study aims to describe some of the errors in the use of Japanese speech levels in the Kenjougo form by Japanese language learners and the smooth form of Sundanese language for themselves by some Sundanese people. What kinds of errors are found in the use of the two languages, the answers will be presented in this study.The data in this study for the use of the Japanese language were obtained from questionnaires addressed to students of the Japanese Literature Study Program, Faculty of Humanities, Gadjah Mada University semester III, V, and VII. While the data for the use of the Sundanese language consisted of two groups, the first group was obtained from the distribution of questionnaires addressed to Japanese language lecturers who spoke Sundanese and the second group was obtained from housewives, civil servants, retired civil servants, and others. It was concluded that the errors in using Kenjougo in the third semester were 36.36%, the fifth semester errors were 21.27%, and the seventh semester errors were 33.33%. Meanwhile, errors in the use of the lemes keur sorangan form among Japanese language lecturers who speak Sundanese as much as 22% and the public as much as 15%.