• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.28B no.10
• /
• pp.755-760
• /
• 1991

In this paper, we propose a new formant locus overlapping method which can effectively enhance a naturalness of synthetic speech produced by ddemisyllable based Korean text-to-speech system. At first, Korean demisyllables are divided into several number of segments which have linear formant transition characteristics. Then, database, which is composed of start point and length of each formant segments, is provided. When we synthesize speech with these demisyllable database, we concatenate each formant locus by using a proposed overlapping method which can closely simulate haman articulation mechanism. We have implemented a Korean text-to-speech system by using this method and proved that the formant loci of synthetic speech are similar to those of the natural speech. Finally, we could illustrate that the resulting spectrograms of proposed method are more similar to natural speech than those of conventional method.

• Phonetics and Speech Sciences
• /
• v.10 no.3
• /
• pp.31-39
• /
• 2018

This paper aims to present a taxonomy of non-standard words (NSW) for developing a French text normalization system and to propose a method for evaluating this system based on a corpus. The proposed taxonomy of French NSWs consists of 13 categories, including 2 types of letter-based categories and 9 types of number-based categories. In order to evaluate the text normalization system, a representative test set including NSWs from various text domains, such as news, literature, non-fiction, social-networking services (SNSs), and transcriptions, is constructed, and an evaluation equation is proposed reflecting the distribution of the NSW categories of the target domain to which the system is applied. The error rate of the test set is 1.64%, while the error rate of the whole corpus is 2.08%, reflecting the NSW distribution in the corpus. The results show that the literature and SNS domains are assessed as having higher error rates compared to the test set.

• Phonetics and Speech Sciences
• /
• v.5 no.4
• /
• pp.217-225
• /
• 2013

This paper describes the design and implementation of the kWebAnywhere system based on WebAnywhere, which assists people with severely diminished eye sight and the blind people to access Internet information through Web interfaces. The WebAnywhere is a server-based web reader which reads aloud the web contents using TTS(text-to-speech) technology on the Internet without installing any software on the client's system. The system can be used in general web browsers using a built-in audio function, for blind users who are unable to afford to use a screen reader and for web developers to design web accessibility. However, the WebAnywhere is limited to supporting only a single language and cannot be applied to Korean web contents directly. Thus, in this paper, we modified the WebAnywhere to serve multiple language contents written in both English and Korean texts. The modified WebAnywhere system is called kWebAnywhere to differentiate it with the original system. The kWebAnywhere system is modified to support the Korean TTS system, VoiceText$^{TM}$, and to include user interface to control the parameters of the TTS system. Because the VoiceText$^{TM}$ system does not support the Festival API used in the WebAnywhere, we developed the Festival Wrapper to transform the VoiceText$^{TM}$'s private APIs to the Festival APIs in order to communicate with the WebAnywhere engine. We expect that the developed system can help people with severely diminished eye sight and the blind people to access the internet contents easily.

• MALSORI
• /
• no.53
• /
• pp.61-74
• /
• 2005

In the present study, the intelligibility of the synthesized speech sounds was evaluated by using the psycholinguistic and fMRI techniques. In order to see the difference in recognizing words between the natural and synthesized speech sounds, word regularity and word frequency were varied. The results of Experiment1 and Experiment2 showed that the intelligibility difference of the synthesized speech comes from word regularity. In the case of the synthesized speech, the regular words were recognized slower than the irregular words, and there was smaller activation of the auditory areas in brain for the regular words than for the irregular words.

• Proceedings of the IEEK Conference
• /
• 2003.07c
• /
• pp.2649-2652
• /
• 2003

In this Paper, an Audio-Visual Teaching aid (AVTA) for use in a classroom and with Internet is presented. A system, which was designed and tested, consists of a wireless Microphone system, Text to Speech conversion Software, Noise filtering circuit and a Computer. An IBM compatible PC with sound card and Network Interface card and a Web browser and a voice and text messenger service were used to provide slightly delayed text and also voice over the internet for remote teaming, while providing scrolling text from a real time lecture in a classroom. The motivation for design of this system, was to aid Korean students who may have difficulty in listening comprehension while have, fairly good reading ability of text. This application of this system is twofold. On one hand it will help the students in a class to view and listen to a lecture, and on the other hand, it will serve as a vehicle for remote access (audio and text) for a classroom lecture. The project provides a simple and low cost solution to remote learning and also allows a student to have access to classroom in emergency situations when the student, can not attend a class. In addition, such system allows the student in capturing a teacher's lecture in audio and text form, without the need to be present in class or having to take many notes. This system will therefore help students in many ways.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1994.06a
• /
• pp.1082-1087
• /
• 1994

A set of text database is indispensable to the probabilistic models for speech recognition, linguistic model, and machine translation. We introduce an environment to canstruct text databases : an automatic tagging system and a set of tools for lexical knowledge acquisition, which provides the facilities of automatic part of speech recognition and guessing.

• International Journal of Computer Science & Network Security
• /
• v.24 no.2
• /
• pp.25-30
• /
• 2024

Various new technologies and aiding instruments are always being introduced for the betterment of the challenged. This project focuses on aiding the mute in expressing their views and ideas in a much efficient and effective manner thereby creating their own place in this world. The proposed system focuses on using various gestures traced into texts which could in turn be transformed into speech. The gesture identification and mapping is performed by the Kinect device, which is found to cost effective and reliable. A suitable text to speech convertor is used to translate the texts generated from Kinect into a speech. The proposed system though cannot be applied to man-to-man conversation owing to the hardware complexities, but could find itself very much of use under addressing environments such as auditoriums, classrooms, etc

• Phonetics and Speech Sciences
• /
• v.12 no.2
• /
• pp.29-37
• /
• 2020

Recently, the neural network-based deep learning algorithm has dramatically improved performance compared to the classical Gaussian mixture model based hidden Markov model (GMM-HMM) automatic speech recognition (ASR) system. In addition, researches on end-to-end (E2E) speech recognition systems integrating language modeling and decoding processes have been actively conducted to better utilize the advantages of deep learning techniques. In general, E2E ASR systems consist of multiple layers of encoder-decoder structure with attention. Therefore, E2E ASR systems require data with a large amount of speech-text paired data in order to achieve good performance. Obtaining speech-text paired data requires a lot of human labor and time, and is a high barrier to building E2E ASR system. Therefore, there are previous studies that improve the performance of E2E ASR system using relatively small amount of speech-text paired data, but most studies have been conducted by using only speech-only data or text-only data. In this study, we proposed a semi-supervised training method that enables E2E ASR system to perform well in corpus in different domains by using both speech or text only data. The proposed method works effectively by adapting to different domains, showing good performance in the target domain and not degrading much in the source domain.

• Speech Sciences
• /
• v.2
• /
• pp.9-23
• /
• 1997

Through iterative analysis and synthesis experiments on Korean monosyllables, the Korean text-to-speech system was implemented using the phoneme-based formant synthesis method. Since the formants of initial and final consonants in this system showed many variations depending on the medial vowels, the database for each phoneme was made up of formants depending on the medial vowels as well as duration information of transition region. These techniques were needed to improve the intelligibility of synthetic speech. This paper investigates also methods of concatenating the synthesis units to improve the quality of synthetic speech.

• The Journal of the Convergence on Culture Technology
• /
• v.7 no.1
• /
• pp.640-645
• /
• 2021

The deep learning based end-to-end TTS system consists of Text2Mel module that generates spectrogram from text, and vocoder module that synthesizes speech signals from spectrogram. Recently, by applying deep learning technology to the TTS system the intelligibility and naturalness of the synthesized speech is as improved as human vocalization. However, it has the disadvantage that the inference speed for synthesizing speech is very slow compared to the conventional method. The inference speed can be improved by applying the non-autoregressive method which can generate speech samples in parallel independent of previously generated samples. In this paper, we introduce FastSpeech, FastSpeech 2, and FastPitch as Text2Mel technology, and Parallel WaveGAN, Multi-band MelGAN, and WaveGlow as vocoder technology applying non-autoregressive method. And we implement them to verify whether it can be processed in real time. Experimental results show that by the obtained RTF all the presented methods are sufficiently capable of real-time processing. And it can be seen that the size of the learned model is about tens to hundreds of megabytes except WaveGlow, and it can be applied to the embedded environment where the memory is limited.