Text to Speech

This article presents a case study on the successful adaptation of an English language exam paper to visually diverse learners in the classroom. The study delves into the challenges confronted by a partially blind student with substantial sight loss and no prior Braille training when attempting traditional written English language exams. The identified obstacles encompass difficulties in navigating continuous text with text-to-speech (TTS) software, the reliance on images, and the impact of time constraints. The research underscores the effectiveness of adaptations made to the exam format to address these challenges and enhance accessibility and inclusivity. The findings reveal favourable outcomes for the student’s performance and well-being during the adapted exam. The study stresses the significance of longitudinal studies to monitor the progress and language proficiency development of visually diverse learners over time. Such investigations would enable comparisons with non-disabled peers, shedding light on the impact of exam accommodations on learning outcomes.

In the project HERON, we developed a framework for articulatory speech synthesis for European Portuguese. The system combines several modules, developed or adapted in the project. The Linguistic Processing model uses new syllabification and grapheme-to-phone modules developed in the project. The construction of the gestural scores is performed using an adapted version of TADA (TAsk Dynamic Application), made available for research by Haskins Laboratories. Other important result from HERON was the (first) comprehensive ...

Text-to-speech synthesis systems are concerned with the artificial generation of a natural and intelligible human voice from given text transcriptions. Despite the potential applications of the text-to-speech systems, it is a language-dependent discipline and most of the attempts are concerned with resourceful languages specifically the English language. Afaan Oromo is one of the under-resourced languages that have a shortage of language resources for developing a text-to-speech system. In this study, a speech dataset containing 8076 text and audio pairs was collected and prepared from legitimate sources to develop a text-to-speech synthesizer for Afaan Oromo. Apart from standard words and names, the proposed model incorporates nonstandard words including numbers, abbreviations, currency, and acronyms. The deep neural network is chosen for this work because it can map complex linguistic features into acoustic feature parameters. Several experiments are conducted to determine the best performing model. The attention error test and mean opinion score test are used for objective and subjective evaluations respectively used to assess the performance of the models. According to objective evaluation result, tacotron 2 made only 2 attention errors while deep voice 3 made 16 out of 148 words in the evaluation sentence list. In addition, tacotron 2 achieved the mos result of 4.32 and 4.21 out of five and deep voice 3 achieved 3.28 and 3.02 in terms of intelligibility and naturalness respectively. Consequently the tacotron 2 model provided an encouraging result, which makes the model appropriate for variety of applications such as recommendation systems, telephone inquiry services, and smart educations.

This work presents the frontema as a unit created to facilitate the study of phenomena regarding co-articulatory effects in word boundaries. Frontema is the orthographic representation of written language as well as of the acoustic unit corresponding to word boundaries that conventionally form a written text. The concept of frontema, as well as its materialization, simplify the process of discovery

Urdu language processing applications encounter non-Urdu text specifically English text frequently. The accuracy of these systems eg machine translation, text-to-speech etc. is highly undermined as they are unable to handle English text. One possibility could be addition ...

The paper present the issues encountered in processing spontaneous Czech speech in the MALACH project. Specific problems connected with a frequent occurrence of colloquial words in spontaneous Czech are analyzed; a partial solution is proposed and experimentally evaluated.

The paper describes the process of creation of domain-specific speech corpora containing air traffic control (ATC) communication prompts. Since the ATC domain is highly specific both from the acoustic point-of-view (significant level of noise in the signal, non-native English accents of the speakers, non-standard pronunciation of some frequent words) and the lexical and syntactic perspective (prescribed structure of utterances, rather limited vocabulary), it is useful to collect and annotate data from this specific domain. Actually, the ultimate goal of the research effort of our team was to develop a voice dialogue system simulating the responses of the pilot that could be used for training aspiring air traffic controllers. In order to do so, we needed – among other modules – a domain-specific automatic speech recognition (ASR) and text-to-speech synthesis (TTS) engines. This paper concentrates on the details of the ASR and TTS corpora creation process but also overviews their usag...

Millions of individuals around the globe have permanent visual impairment, underscoring the importance of facilitating their
understanding of people and the identification of essential daily-use products. To address this need, we propose the system to recognize
such items within their daily routines. Numerous initiatives are underway in this field to aid the visually impaired without end to end
deployment. The objective is to identify objects and translate them into auditory cues to inform individuals with visual impairment about
these items with the system comprises a camera, a speaker, and an image processing system. The primary focus of this study is the
amalgamation of real-time object detection and recognition using advanced deep learning techniques. The aim is to detect and label the
position and names of multiple objects captured by the camera through an object detection algorithm.

The integration of artificial intelligence (AI)
in architecture is transforming the design process, making
it faster, more efficient, and more sustainable. AI serves as
a starting point for conversation, allowing architects to
interact with data scientists and engineers. The study
recommends merging AI into architecture education to
improve designers' comprehension of design-related
concepts. The future role of architects in AI design and its
application raises questions about how AI design will
reflect their creativity and architectural touch. AI is
revolutionizing the design process by integrating Building
Information Modelling (BIM) techniques and enabling
real-time analysis and optimization. It also revolutionizes
architectural design education by generating initial project
forms and improving interaction with architects. AI
applications can revolutionize the way students learn and
create, allowing them to explore innovative designs they
may not have considered otherwise. As AI applications
continue to advance, they have the potential to evolve and
provide even more sophisticated design solutions.
AI-powered design tools have the potential to
revolutionize the way architects approach and create
architectural projects. The study explores the potential of
artificial intelligence in architectural design. It highlights
the integration of AI through generative design
algorithms, virtual reality tools, and machine learning
algorithms. These tools help architects explore design
options, optimize projects for energy efficiency and
structural integrity, and analyze large amounts of data for
informed decisions. AI-powered design tools can
understand aesthetics, architecture schools, and project
requirements, enhancing the initial design stage.
Integrating Building Information Modeling (BIM)
techniques can enable real-time analysis and optimization
of designs, saving time and resources. AI applications can
also revolutionize architectural design education by
generating initial project forms and allowing students to
develop innovative designs.

This thesis describes research on building a text-to-speech (TTS) framework that can accommodate the lack of linguistic information of under-resource languages by using existing resources from another language. It describes the adaptation process required when such limited resource is used. The main natural languages involved in this research are Malay and Iban language. The thesis includes a study on grapheme to phoneme mapping and the substitution of phonemes. A set of substitution matrices is presented which show the phoneme confusion in term of perception among respondents. The experiments conducted study the intelligibility as well as perception based on context of utterances. The study on the phonetic prosody is then presented and compared to the Klatt duration model. This is to find the similarities of cross language duration model if one exists. Then a comparative study of Iban native speaker with an Iban polyglot TTS using Malay resources is presented. This is to confirm th...

Abstract— A Malay speech synthesizer system will be discussed. This paper will cover the available Malay speech synthesis system, the underlying structure of our system, brief description of crucial modules, general evaluation of the system, the proposed enhancement and future work of ...

The paper presents the development of a phonetically balanced read speech corpus of code-mixed Hindi-English. Phonetic balance in the corpus has been created by selecting sentences that contained triphones lower in frequency than a predefined threshold. The assumption with a compulsory inclusion of such rare units was that the high frequency triphones will inevitably be included. Using this metric, the Pearson’s correlation coefficient of the phonetically balanced corpus with a large code-mixed reference corpus was recorded to be 0.996. The data for corpus creation has been extracted from selected sections of Hindi newspapers.These sections contain frequent English insertions in a matrix of Hindi sentence. Statistics on the phone and triphone distribution have been presented, to graphically display the phonetic likeness between the reference corpus and the corpus sampled through our method.

This patent is based on a novel model of human categorial perception of harmonic speech sounds, alternative to the currently dominant formant theory and to the NLP procedures based on the concepts of amplitude envelope and Mel-frequency scale. The field of the present invention relates to recognition or synthesis of human-uttered speech. In particular, computer-implemented methods are disclosed for recognizing or synthesizing human- uttered harmonic sounds.

In this paper we compare two different methods for phonetically labeling a speech database. The first approach is based on the alignment of the speech signal on a high quality synthetic speech pattern, and the second one uses a hybrid HMM/ANN system. Both systems have been evaluated on French read utterances from a speaker never seen in the training stage of the HMM/ANN system and manually segmented. This study outlines the advantages and drawbacks of both methods. The high quality speech synthetic system has the great advantage that no training stage is needed, while the classical HMM/ANN system easily allows multiple phonetic transcriptions. We deduce a method for the automatic constitution of phonetically labeled speech databases based on using the synthetic speech segmentation tool to bootstrap the training process of our hybrid HMM/ANN system. The importance of such segmentation tools will be a key point for the development of improved speech synthesis and recognition systems.

A comprehension derived from a sentence meaning which has syntactico-semantical relations between the word meanings is to be noted as the verbal cognition. In such verbal cognition among the many meaningful units, one is to be considered as principle qualificand or mukhya viśeṣya, whereas the other meaning elements are linked to it directly or indirectly. According to the logicians, the meaning of the word in a nominative case is the principle qualificand. But the Sanskrit grammatical tradition disagrees with the theory of logicians. The grammarians behold the idea that, an action denoted by a verbal root is the principle qualificand or mukhya viśeṣya in a sentence meaning or vākyārtha. Depending upon the shreds of evidence from mahābhāṣya provides explanations for these peculiar sentences; paśya mṛgo dhāvati pacati bhavati etc. it seems that the theory of the grammarians is preferable.  Following the theory of the grammarians, and using some technical terms like prakāratā, viśeṣyatā etc., after analysing the relation at the sentence level and word level as well, this paper tries to represent a special structure for the writing of a sentence. This representation indicates the principle qualificand in sentence level and discourse level, as well.

Introduction. This study aims to contribute to a better understanding of the learning that reading by listening requires from readers with print disabilities. Method. Sixteen semi-structured interviews were conducted with university students with blindness or vision impairment in Australia. Analysis. A theoretical reading of the interview material was conducted, with a basis in sociocultural learning theory. Results. The responsibility for learning to use audio-based reading tools was often left to the participants themselves. The process of appropriating audio-based reading tools included two important aspects: digital literacy and learning to manage the sequentiality of audio text. This process was related to trajectories of participation in academic communities of practice, English-speaking communities of practice, and blindness and visual impairments communities of practice. It also included learning to participate in practices where vision and reading by seeing is the norm. Conclusion. Three conclusions are drawn: 1) reading by listening is not a passive and effortless activity, it requires specific expertise, 2) socioeconomic and sociopolitical circumstances influence how reading by listening is practiced and the learning that this form of reading requires, and 3) reading by listening requires learning how to navigate and manage the effects of institutions and systems that are disabling.

This template gives you guidelines to prepare papers for SETIT 2005 Conference. Use this document as a template if you are using Microsoft Word 6.0 or later. Otherwise, use this document as an instruction set. Please note that use of SETIT 2005 templates is meant to assist authors in correctly formatting manuscripts for final submission. This template must be used

Follow this and additional works at: https://escholarship.umassmed.edu/chr_symposium Part of the Behavioral Medicine Commons, Civic and Community Engagement Commons, Communication Sciences and Disorders Commons, Community-Based Research Commons, Community Health and Preventive Medicine Commons, Health Communication Commons, Health Information Technology Commons, Health Services Administration Commons, Medical Education Commons, Online and Distance Education Commons, and the Translational Medical Research Commons

HLSyn is a quasi-articulatory synthesizer (independent of any particular language) with a small set of parameters, which model the complex acoustic consequences of vocal tract. The primary motivation of using HLSyn is to combine the ‘simplicity of control’ of articulatory synthesizers with ‘accuracy and efficiency’ of formant synthesizers. The aim of this paper is to analyze the quality of nonnasal and non-aspirated Urdu Stops, synthesized using HLSyn.

A multilayer perceptron (MLP) similar to that used in the NETtalk system is used to form a mapping between sequences of allophones and corresponding frames of LPC synthesizer control parameters. Three parameter sets equivalent to the LPC coefficients, line spectral pair (LSP), PARCOR and log area ratio, are evaluated. In addition to training a standard MLP, networks which have been decomposed according to phonetic class and by allophone, are trained. Decomposition is found to reduce training time and produce ...

We have added a dynamic bio-mechanical mapping layer that contains a model of the human vocal tract with tongue muscle activations as input and tract geometry as output to a real time gesture controlled voice synthesizer system used for musical performance and speech research. Using this mapping layer, we conducted user studies comparing controlling the model muscle activations using a 2D set of force sensors with a position controlled kinematic input space that maps directly to the sound. Preliminary user evaluation suggests that it was more difficult to using force input but the resultant output sound was more intelligible and natural compared to the kinematic controller. This result shows that force input is a potentially feasible for browsing through a vowel space for an articulatory voice synthesis system, although further evaluation is required.

According to the World Health Organization, Mute and deaf community are among the 5% of our world population that is approximately 430 million people. This number can increase to up to 2.5 billion till 2050 i.e. one among 10 people would suffer from some kind of hearing disorder. Often these differently abled people are neglected by society. These people face various problems like lack of basic education, platforms to express themselves to people who don’t understand sign language or able to earn jobs to fulfil their basic needs, etc. According WHO unaddressed issues faced by mute and deaf people were that, there are almost 32 million kids who suffer from hearing loss and these are from currently developing countries where there is limited budget to provide education, or any other facilities that could help these people to compete with world. Thus, in adults the unemployment rate is quite high. Further leading to social isolation, loneliness, depression. Thus, this survey majorly fo...