Text-to-speech Research Articles

THE COHESION AND COHERENCE IN BARACK OBAMA’S SPEECH TEXT “FINAL PRESIDENTIAL PRESS CONFERENCE OF YEAR”

This research aimed at analyzing the cohesion and coherence in Barack Obama’s speech text “Final presidential press conference of year”. This research designed by the qualitative method with descriptive approach. Then, researcher used the Miles, Huberman, & Saldana theory to analyze the cohesion and coherence in the speech text. The result of this research showed all of cohesion and coherence devices were found. The data of this research were words and phrases. Based on summary of the result, that Barack Obama’s speech text “Final Presidential Press Conference of Year” contained grammatical cohesion devices, they are reference (34), subtitution (1), ellipsis (12), and conjuntion (60). And then, it contained the lexical cohesion devices. They are; reiteration (47) and collocation (11). Moreover, it contained the coherence devices. They are; repeat key nouns (9), consistent pronoun (5), transition signals (53), and the ideas in the speech text are arranged structurally. Based on the research finding, in the spech text, the cohesion devices dominanted in grammatical cohesion, especially in using conjuction. And the coherence devices dominanted in using transition signals. It is suggested that English learners’ should pay attention about the usage of cohesion and coherence in order to create a good text as well as a speech. For the teachers, hopefully this research could be used to enrich their teaching material. And for the next researcher, can get information and knowledge to create a text grammatically and semantically.

English Pronunciation Quality Evaluation System Based on Continuous Speech Recognition Technology for Multi-Terminal

This paper presents a novel approach for evaluating the pronunciation quality of English speech using continuous speech recognition technology. The research focuses on the application of artificial intelligence in speech recognition, utilizing web browsers on various terminal devices such as computers, mobile phones, and tablets to allow users to read the provided text aloud. The web program captures audio input from the microphone, records it in MP3 format, and uploads it to the server. The server employs the Whisper model to transcribe the audio into semantic text, which is then compared with the displayed text. By calculating the semantic distance and assessing the accuracy of pronunciation, the system provides an evaluation of pronunciation quality, marking correct and incorrect words. To achieve real-time processing, the compact tiny model is employed, and further optimization is performed using Ctranslate 2, resulting in significant performance improvements.

PVGAN: A Pathological Voice Generation Model Incorporating a Progressive Nesting Strategy

The voice generation task is to solve the problem of limited samples in the voice dataset using computer technology. By increasing the number of samples, the accuracy of voice disorder diagnosis can be improved, which has a wide range of application value in medical diagnosis and other fields. At present, there are insufficient models for detailed features such as pitch, timbre, and different frequency components in pathological voice data. Therefore, this paper proposes a PVGAN network for learning different frequency information of audio to generate pathological voice data. The proposed network captures the multi-scale features and different periodic patterns of audio signals by designing multiscale perceptual residual blocks and periodic discriminators. At the same time, a progressive nesting strategy was proposed to combine the generator and the discriminator to improve the learning ability of different resolution information. In addition, a latent mapping network is designed to fuse the latent vector with the condition information to generate sound features related to specific diseases or pathological states. The loss function is optimized to further improve the model performance. On the Saarbruecken Voice Database(SVD), the average values of each index of the data generated after training with different pathological types as conditional information are similar to the original data. Finally, the generated data were used to expand the SVD dataset, and the accuracy of the two classification experiments was improved to a certain extent.

Data Augmentation based Cross-Lingual Multi-Speaker TTS using DL with Sentiment Analysis

Text to S peech (TTS) algorithms have made tremendous strides in recent years in terms of their ability to generate speech in a single language that sounds as natural as possible. However, because to a lack of available training data, the synthesis of speech from the same person in various languages continues to be difficult. It might be challenging to locate people who are proficient in numerous languages at a level equivalent to native speakers. Voice conversion is one method that may be used to create a polyglot corpus, which can then be used to solve this problem. This entails making use of a voice representation model that has been trained on 53 different languages through the application of hybrid deep learning in order to capture speaker-invariant qualities. In this study, we present a novel approach for the conversion of voices across different languages by employing Generated Adversarial Networks (GANs) to train a multilingual TTS system. The concept of individual likeness loss in order to address the special difficulty of maintaining one's individual speaking identity during the training process can be offered. This work is focused to provide the impression that voice data coming from a variety of languages and speakers was produced by the same individual, and one way to do so is by using this word. In order to determine the extent to which our model is useful, two experiments that compared it against benchmarks that made use of varying degrees of parameter sharing between languages are carried out. The purpose of these experiments was to evaluate the accuracy of pronunciation as well as the caliber of the synthetic voice during transitions between different languages.

Enhancing virtual instruction: leveraging AI applications for success

Purpose This paper aims to describe several artificial intelligence (AI)-based applications that librarians can use to serve and design virtual library instruction, so it will be more effective and efficient. Design/methodology/approach The approach involves a comprehensive review of AI-based applications that bring benefits to librarian to enhance the virtual instructional services (AI). This study explores the existing papers to reveal the potential use of AI for research consultation, designing the instructional services and conducting evaluation of the program. Findings There are some AI-based applications that are available for free that will help instructional librarian jobs. Librarians use the AI to increase effectiveness of the services. The AI-based applications that can be used to support instructional services on research inquiries include virtual assistance, knowledge mapping and note making, and to support designing virtual instruction, librarians can use design apps, image generators, voice generator, grammar checker and paraphrasing. Originality/value There are many studies on AI at the library; however, it’s still rare a paper studied AI-based application that potentially will bring benefit for virtual instructional services. This paper will give overview of AI application that will help instructional librarian on transactions with users and help librarians to create innovative instructional media.

Kim Namjoon Speech at the UN Assembly: A Discourse Analysis

This research aims at revealing values, belief and assumptions are found in Kim Namjoon’s speech text on campaigning “Love Yourself” and to know types of figurative language used in the speech. This research used descriptive qualitative method. The instrument used in this research was the speech text transcript and speech video that retrieved from YouTube. Data collection techniques used in this research was library research techniques, thus seeing and recording as the main approach to collect the data. The qualitative data analysis was employed after following stages; analyzing data reduction, data display and verification.

A large-scale comparison of two voice synthesis techniques on intelligibility, naturalness, preferences, and attitudes toward voices banked by individuals with amyotrophic lateral sclerosis

Amyotrophic lateral sclerosis (ALS) commonly results in the inability to produce natural speech, making speech-generating devices (SGDs) important. Historically, synthetic voices generated by SGDs were neither unique, nor age- or dialect-appropriate, which depersonalized SGD use. Voices generated by SGDs can now be customized via voice banking and should ideally sound uniquely like the individual’s natural speech, be intelligible, and elicit positive reactions from communication partners. This large-scale 2 x 2 mixed between- and within-participants design examined perceptions of 831 adult listeners regarding custom synthetic voices created for two individuals diagnosed with ALS via two synthesis systems in common clinical use (waveform concatenation and statistical parametric synthesis). The study explored relationships among synthesis system, dysarthria severity, synthetic speech intelligibility, naturalness, and preferences, and also provided a preliminary examination of attitudes regarding the custom synthetic voices. Synthetic voices generated via statistical parametric synthesis trained on deep neural networks were more intelligible, natural, and preferred than voices produced via waveform concatenation, and were associated with more positive attitudes. The custom synthetic voice created from moderately dysarthric speech was more intelligible than the voice created from mildly dysarthric speech. Clinical implications and factors that may have contributed to the relative intelligibilities are discussed.

Optimizing medical personnel speech recognition models using speech synthesis and reinforcement learning

Text-to-Speech synthesis (TTS) can be used to generate training data for building Automatic Speech Recognition models (ASR). Access to medical speech data is because it is sensitive data that is difficult to obtain for privacy reasons. Speech can be synthesized by mimicking different accents, dialects, and speaking styles in a medical language. Reinforcement Learning (RL), in the context of ASR, can be used to optimize a model. A model can be trained to minimize errors in speech-to-text transcription, especially for technical medical terminology. In this case, the “reward” to the RL model can be negatively proportional to the number of transcription errors. The paper presents a method and experimental study from which it is concluded that the combination of TTS and RL can enable the creation of a speech recognition model suited to the specific needs of medical personnel, helping to expand the training data and optimize the model to minimize transcription errors. The learning process used reward functions based on Mean Opinion Score (MOS), a subjective metric for assessing speech quality, and Word Error Rate (WER), which evaluates the quality of speech-to-text transcription. [The Polish National Center for Research and Development (NCBR) supported the project: “ADMEDVOICE- Adaptive intelligent speech processing system of medical personnel with the structuring of test results and support of therapeutic process,” no. INFOSTRATEG4/0003/2022.]

Effects of Artificial Synthetic Speech Control of SNR and Speech Rate on the Intelligibility of Train Station Announcements

AbstractAn experimental study on the effect of the speech characteristics of the signal-to-noise ratio (SNR) and speech rate on the intelligibility of announcements at railway stations was conducted using an artificial synthetic voice. Synthesized speech has recently been used in noisy environments both indoors and outdoors, but unlike its use in quiet environments, when the environment is noisy, the intelligibility of announcements may be reduced. For railway station announcements, while natural spoken voices are currently used for multilingual announcements and disaster response broadcasts, deep neural network synthesized voices, which use deep learning, have also been adopted. However, the effect of the acoustic characteristics such as the SNR and speech rate on the intelligibility of reproduced announcements in noisy public spaces such as railway stations has not yet been clarified from a practical viewpoint. In this paper, in order to determine the appropriate SNR and speech rate for synthetic voice announcements in railway stations, auditory impressions of announcements with varying SNR and speech rate were evaluated by participants using a five-point scale. Based on the evaluations, the appropriate conditions for the broadcast of synthetic voice announcements at the ticket gate and on the platform of a station are discussed.

Deixis Study in Speech Text of President Joko Widodo on Independence Day

The deixis in the speech text of President Joko Widodo on Independence Day, August 17, 2022, is the subject of this article. This study's objectives are to identify deixis types, identify the dominant type in the speech text of President Joko Widodo, and explain deixis types and their causes. The study aims to identify the prevalent deixis type, understand the procedure, and determine why President Joko Widodo uses deixis in his speeches. This study employs the qualitative content analysis research design. The speech text of President Joko Widodo on Independence Day, August 17, 2022, serves as the data's primary source. The documentary technique is used for data collection. Documentary sheets are the tool used for data collection, and content analysis is the method used for data analysis. President Joko Widodo's speech text analysis on Independence Day reveals five different types of deixis. According to the analysis, social deixis is the most frequently used. The purpose of social deixis in a speech for Independence Day by President Joko Widodo demonstrates that there are distinctions in the roles of recipients of speech referrals and revealed reveals language politeness tactics, particularly those related to cultural factors, and depicts how a person's social status, such as age, occupation, education background, and social position, impacts a country's leader to deliver his Speech during the celebration of Independence Day.

Keterampilan Khithabah untuk Meningkatkan Maharah Kalam di Pondok Pesantren Darul Amin Bustanul Arifin

<p><em>Islamic Boarding School of Darul Amin Bustanul Arifin is one of the Islamic boarding schools located in the Mandailing Natal regency. Based on preliminary research in the field by interviewing the students, the Arabic speaking skill of the students at the Islamic Boarding School of Darul Amin Bustanul Arifin was considered as low. In consequence, the team of community service consisting of Arabic language and literature lecturers was motivated to organize a relevant program on Arabic language enhancement through training in speech or khithabah. This training was hoped to give an improvement on students’ Arabic speaking skills. In addition, this program was expected to trigger the students to be self-confident to deliver a speech in front of the audience, particularly when expressing ideas, sharing new knowledge, and preaching religious messages to society.<strong> </strong>This program applied two methods namely the speech method delivered by the team of community service consisting of Arabic language and literature lecturers and the communicative method through question-answer sessions related to the program’s materials which were strengthened by displaying the speech text and allowing students to practice the speech by turns.<strong> </strong>The program of community service conducted by the team of Arabic Language and literature lecturers at the Islamic boarding school of Darul Amin Bustanul Arifin run very well. The result of the program is expected to give benefit students’ Arabic speaking skills pertinently improving their self-confidence for speech delivery in front of the audience.<strong></strong></em></p><strong><em>Keywords: </em></strong><em>training of speech, Speaking Skills, Islamic boarding school, Darul Amin Bustanul Arifin.</em>

Deep Learning based Multilingual Speech Synthesis using Multi Feature Fusion Methods

The poor intelligibility and out-of-the-ordinary nature of the traditional concatenation speech synthesis technologies are two major problems. CNN's context deep learning approaches aren't robust enough for sensitive speech synthesis. Our suggested approach may satisfy such needs and modify the complexities of voice synthesis. The suggested model's minimal aperiodic distortion makes it an excellent candidate for a communication recognition model. Our suggested method is as close to human speech as possible, despite the fact that speech synthesis has a number of audible flaws. Additionally, there is excellent hard work to be done in incorporating sentiment analysis into text categorization using natural language processing. The intensity of feeling varies greatly from nation to country. To improve their voice synthesis outputs, models need to include more and more concealed layers & nodes into the updated mixture density network. For our suggested algorithm to perform at its best, we need a more robust network foundation and optimization methods. We hope that after reading this article and trying out the example data provided, both experienced researchers and those just starting out would have a better grasp of the steps involved in creating a deep learning approach. Overcoming fitting issues with less data in training, the model is making progress. More space is needed to hold the input parameters in the DL-based method.

SENTENCE PATTERN USED IN CALUM SCOTT’S SONG LYRICH

Sentence patterns and sentence types are an inseparable unit where each type of sentence contains several sentence patterns. In general, some researchers analyzed sentence patterns in a short story, speech text, news text and others. But in this study will analyze sentence patterns and sentence types in two song lyrics by Calum Scott. Usually in a song lyrics consists of phrases, clauses, and sentences, but the study will analyze how many sentences are there. This study aims to find sentence patterns and sentence types in the two song lyrics. This study uses a qualitative method. The results of the study show that there are six sentence patterns and six types of sentences found in the two songs. No Matter What's first song with details as follows; pattern I (S+VT+DO), pattern II (S+int.V), pattern, pattern IV (S+intr.V+SP), pattern V (S+com.V+DO+OP), and pattern IV (S+pre.V+PC). The second song is Dancing on My Own with the following details; pattern I (S+VT+DO), pattern II (S+int.V), pattern III (S+dist.V+DO+IO), pattern IV (S+intr.V+SP), pattern V (S+com.V+DO+OP), and pattern VI (S+pre.V+PC). While the types of sentences from the two songs, No Matter What and Dancing on My Own there are four types. They are simple sentence, compound sentence, complex sentence, and compound complex sentence.

Does the Dependence of Brand Voice on AI Restrict Freedom of Expression in Social Media?

Abstract With the advent of social media and a unique brand voice, brand communications frequently employ artificial intelligence (AI) to facilitate timely interactions across the many channels of communication, and often integrate communications. This research seeks to investigate the influence of an AI-regulated brand voice on the freedom of expression (FOE) when communicating on social media. A qualitative research design is implemented based on content analysis. The research gap was identified by reviewing the role of brand voice using AI, the application of AI to monitor freedom of expression, and the study of integrating FOE in brand voice generation using AI. Various case studies were used to conclude that AI-regulated brand voice has a substantial impact on the freedom of expression on social media, which is frequently governed by privacy rules. This study establishes that a brand’s communication in its social media portfolio is governed by national and international laws that regulate freedom of expression and thus have legal implications. This study suggests that AI must be fed information to prevent content such as hate speech, bullying, and so on, and produce a brand voice with checkpoints for human validation and verification to maintain general societal decorum.

Text to Speech Conversion based on Emotion using Recurrent Neural Network

Abstract: Emotion based text to speech conversion system will be proved as an improved version of traditional text to speech system. Emotions help us in recognizing the message conveyed by the conveyor in more effective way. ETTS (Emotion based text to speech conversion system) will stand out from TTS (Text to speech) as it will have variation of voice according to the emotions detected in the text. ETTS will detect four basic and most used emotions of humans. These emotions are ‘happy’, ‘sad’, ‘anger’, ‘neutral’. As ETTS will use RNN (Recurrent Neural Network) foridentifying emotions, emotions will be identified with high accuracy. GRU (Gated Recurrent Unit) model of recurrent neural network helps in maintaining the high accuracy of prediction. Emotion based text to speech system will be beneficial for educational purpose like listening stories from storytelling applications for young budding children. Emotion based text to speech is going to be serviceable for visually impaired individuals. It can also be used in robots forbetter communication.

THE USE OF ANTCONC TO INVESTIGATE THE PATTERN OF CORPUS REPRESENTATION IN JOKO WIDODO’S PRESIDENTIAL SPEECH

This research investigates the patterns of representation in Joko Widodo’s corpus of presidential speech. The research method used is a mixed method that combines quantitative and qualitative analysis. The data of the research are transcriptions of President Joko Widodo’s speech text in 2015 and in 2018. In order to be processed with a corpus analysis tool, the data is stored in a text file or .txt format with UTF-8 encoding. The instrument which is used in this study is AntConc. It also brings up words that often appear and keywords in the corpus and analyzes text patterns with certain letter arrangements. Regarding patterns in the text, several appear from words contained in the corpus. Among them are the existence of words that have prefixes, suffixes, prefixes and suffixes, and repetition. Further, from the frequency and keyword analysis, it can be seen that Joko Widodo is focusing his attention on achieving the goals for the Indonesian nation and state, both in the 2015 and 2018 speech texts. It can be concluded, therefore, that Widodo is consistent in his speech patterns. On the other hand, this may be considered a new breakthrough because the speech from 2018 uses the same speech patterns as the speech from three years earlier.

A Deep Learning Approach for ASL Recognition and Text-to-Speech Synthesis using CNN

Abstract: Sign language is a visual language that is used by the deaf and hard-of-hearing community to communicate. However, sign language is not universally understood by non-signers, which can create communication barriers for the deaf and hard-ofhearing individuals. In this paper, we present a novel application for American Sign Language (ASL) to text to speech conversion using deep learning techniques. Our app aims to bridge the communication gap between hearing-impaired individuals who use ASL as their primary mode of communication and individuals who do not understand ASL. The app comprises three main components: a hand gesture recognition module that detects and recognizes ASL signs, a text-tospeech synthesis module that converts recognized ASL signs to text, and a speech synthesis module that converts the text to speech

Время начала озвончения взрывных смычных согласных в устной речи монолингвов в России

The aim of the study is to identify the features of unvoiced and voiced plosive occlusives realised in the speech of monolinguals in Russia when reading a phonetically representative text aloud. It should be noted that the study is novel in that it is the first in Russian linguistics to use the parameter of voice onset time (VOT) to determine the voicing/devoicing of plosive occlusives. The authors consider VOT to be the most reliable acoustic parameter. As a result of the study, it was found that in the speech of Russian monolinguals, VOT was most often implemented incorrectly (positive VOT instead of negative), which is especially exemplified by the material of words starting with the sound [d] and one word in the speech of mostly adults, which is associated with speech fluency. The presence of pauses helped children to pronounce unvoiced and voiced consonants more correctly. Unvoiced consonants were pronounced instead of voiced ones in the speech of adult monolinguals. It should be said that other trends were noted in the speech of bilinguals. Basically, the range of voicing did not correspond to the reference values in both languages (Russian and German). Bilingual adults and children had their own system of parameters. As for unvoiced consonants, their values almost completely corresponded to the reference values of the Russian language, except for the consonant /k/.

Reflections on Building a Multi-Country AAC Implementation Guide.

Augmentative and Alternative Communication (AAC) implementation at any level is a multifaceted process that requires selection of relevant and appropriate systems to suit individual users who may have complex communication needs and other co-occurring difficulties. Careful and systematic action may be required to develop skills and abilities in the use of chosen technologies with suitable ongoing support within a wide range of settings. The wider milieu in which services are provided must also be considered in order to provide a firm foundation for capacity building alongside considerations for multilingual and multicultural factors. UNICEF with the Global Symbols team supported by local professionals working with AAC users, their families and carers set out to collaboratively provide an implementation guide based on their experiences in several Eastern European countries. The aim of the guide was to illustrate work already being undertaken in the area and to ensure the sharing of knowledge and resources where gaps were discovered. The result became a series of linked webpages in an online framework that covered practical aspects for the development of policies and procedures to support early intervention for those with severe speech, language and communication needs across countries of differing cultures and languages. The actual AAC implementation required ingenuity on all sides with translations for pictographic symbol and software adaptations with Cyrillic and Latin alphabets, new synthetic voices alongside deployment and capacity development. Considerable local support was forthcoming and captured with interviews by those working with AAC users as technology was introduced and outcomes measured. As the guide was completed several videos were shared publicly by carers with examples of AAC and assistive technology use. Policies and procedures were also shared in the form of tables, charts, symbol sets, communication boards and software that illustrated not only the occurrence of knowledge transfer and the use of open licenses, but also differences in strategies and the way they were adapted to suit the range of settings in the various countries.

Nadiem Makarim’s Speech in the Freedom of Learning Series: Critical Discourse Analysis of Social Situations and Ideology

This research aims to (1) Explain the transitivity system in Nadiem Makarim speech, (2) Explains the ideology reflected in the speech text. The main character's speech at this time is frequent used as topic study in study Language. A number of study cross disciplinary appear along development role And definition The proposed transitivity system Halliday is an analytical tool for analyzing linguistic data that has been used by Fairclough in his theoretical framework of critical discourse analysis. Based on the purpose of describing the transitivity system and Nadiem Makarim's speech ideology, this research can be categorized as descriptive research. Research result shows the six processes used by Nadiem Makarim in representing his experience, viz material processes, mental processes, relational processes, verbal processes, behavioral processes (behavioral), existential processes. The process that is often used by Nadiema Makarim is the material process followed by the relational process and mental processes. In expressing his ideas, Nadiem Makarim uses a lot of material processes as well show in principle that more Good Act than only just speak. Nadeim Makarim delivered his speech on National Teacher's Day in 2019. And then continued with Nadiem Makarim's speech at the 2020 HGN in the midst of the pandemic.