A FAST ADAPTATION TECHNIQUE FOR BUILDING DIALECTAL MALAY SPEECH SYNTHESIS ACOUSTIC MODEL

2015 ◽  
Vol 77 (19) ◽  
Author(s):  
Yen-Min Jasmina Khaw ◽  
Tien-Ping Tan

This paper presents a fast adaptation technique to build a hidden Markov model (HMM) based dialectal speech synthesis acoustic model. Standard Malay is used as a source language whereas Kelantanese Malay is chosen to be target language in this study. Kelantan dialect is a Malay dialect from the northeast of Peninsular Malaysia. One of the most important steps and time consuming in building a HMM acoustic model is the alignment of speech sound. A good alignment will produce a clear and natural synthesize speech. The importance of this study is to propose a quick approach for aligning and building a good dialectal speech synthesis acoustic model by using a different source acoustic model. There are two proposed adaptation approaches in this study to synthesize dialectal Malay sentences using different amount of target speech and a source acoustic model to build the target acoustic model of speech synthesis system. From the results, we found out that the dialectal speech synthesis system built with adaptation approaches are much better in term of speech quality compared to the one without applying adaptation approach.

Author(s):  
Andika Wijaya ◽  
Gloria Christine Setiyowati

Song lyric translation is important because in these recent decades people can access songs worldwide. The aim of this research is to gain an understanding of the difference between singable translations made by an Indonesian translator and a foreign translator by investigating what translation procedures and methods occur in two translated songs from Indonesian to English using qualitative descriptive method. The result of this research indicates that the singable translation made by a foreign translator is more identical to the source language (SL) compared to the one made by an Indonesian translator. However, despite the differences, the two translated songs share something in common, for instance the singability and the length of lyrics. Taking the findings into consideration, it could be said that the foreign translator is more faithful to the source text (ST), while the Indonesian translator emphasizes the target language (TL) more.


2019 ◽  
Vol 34 (4) ◽  
pp. 349-363 ◽  
Author(s):  
Thinh Van Nguyen ◽  
Bao Quoc Nguyen ◽  
Kinh Huy Phan ◽  
Hai Van Do

In this paper, we present our first Vietnamese speech synthesis system based on deep neural networks. To improve the training data collected from the Internet, a cleaning method is proposed. The experimental results indicate that by using deeper architectures we can achieve better performance for the TTS than using shallow architectures such as hidden Markov model. We also present the effect of using different amounts of data to train the TTS systems. In the VLSP TTS challenge 2018, our proposed DNN-based speech synthesis system won the first place in all three subjects including naturalness, intelligibility, and MOS.


Babel ◽  
2006 ◽  
Vol 52 (2) ◽  
pp. 124-132
Author(s):  
Laurence Wong

Abstract This paper discusses the relationship between syntax and translatability, particularly in respect of literary texts. By translatability is meant the degree of ease with which one language lends itself to translation into another language. Through practice in the translation between Chinese and some of the major European languages, such as English, French, Italian, German, Spanish, Latin, and Greek, as well as between the European languages themselves, it can be found that translating between the European languages is much easier than translating between Chinese and any one of the European languages. Of all the factors that determine whether a language translates more readily or less readily into another language, syntactic differences constitute one of the most decisive. This is because the translator is, during the translation process, constantly dealing with syntax in two directions: the syntax of the source language on the one hand and the syntax of the target language on the other. As a result, problems arising from the syntactic differences between the two languages are bound to figure more prominently than those arising from the differences between individual lexical items and phrases or between cultures. In this paper, syntax will be studied and analysed with reference to Chinese, English, French, German, Italian, Spanish, Latin, and Greek texts. Finally, it will be shown that, mainly because of syntactic differences, there is a higher degree of translatability between any two of the above European languages (which are members of the Indo-European family) than between Chinese (which is a member of the Sino-Tibetan family) and any one of these European languages, and that the syntax of any one of these European languages can cope comfortably with Chinese syntax, but not the other way round. Résumé Cet article traite de la relation entre la syntaxe et la traduisibilité, en particulier, en ce qui concerne les textes littéraires. On entend par traduisibilité le degré de facilité avec laquelle une langue se prête à la traduction dans une autre. Par la pratique de la traduction entre le chinois et quelques-unes des principales langues européennes, comme l’anglais, le français, l’italien, l’allemand, l’espagnol, le latin et le grec, ainsi qu’entre les langues européennes mêmes, on s’aperçoit qu’il est beaucoup plus facile de traduire entre les langues européennes qu’entre le chinois et n’importe quelle langue européenne. Parmi tous les facteurs qui déterminent si une langue se traduit plus ou moins aisément dans une autre, les différences syntactiques comptent parmi les plus décisifs. Ceci est dû au fait que le traducteur, pendant le processus de traduction, est constamment confronté à une syntaxe dans deux directions : la syntaxe de la langue source, d’une part, et la syntaxe de la langue cible, d’autre part. En conséquence, les problèmes dus à des différences syntactiques entre les deux langues doivent nécessairement apparaître de manière plus évidente que ceux provenant de différences entre les syntagmes et éléments lexicaux individuels ou entre les cultures. Dans cet article, la syntaxe sera étudiée et analysée en référence à des textes en chinois, anglais, français, allemand, italien, espagnol, latin et grec. Enfin, il montrera qu’en raison des différences syntactiques surtout, la traduisibilité est plus grande entre deux langues européennes précitées quelles qu’elles soient (qui appartiennent à la famille indo-européenne) qu’entre le chinois (qui appartient à la famille sino-tibétaine) et une quelconque de ces langues européennes. Il montrera que la syntaxe de toute langue européenne peut sans difficulté venir à bout de n’importe quelle syntaxe chinoise, mais que l’inverse n’est pas vrai.


2009 ◽  
Vol 39 (3) ◽  
Author(s):  
Silvia Hansen-Schirra ◽  
Sandra Hansen ◽  
Sascha Wolfer ◽  
Lars Konieczny

This article examines the contrasts and commonalities between languages for specific purposes (LSP) and their popularizations on the one hand and the frequency patterns of LSP register features in English and German on the other. For this purpose corpora of expert-expert and expert-lay communication are annotated for part-of-speech and phrase structure information. On this basis, the frequencies of pre- and post-modifications in complex noun phrases are statistically investigated and compared for English and German. Moreover, using parallel and comparable corpora it is tested whether English-German translations obey the register norms of the target language or whether the LSP frequency patterns of the source language Ñshine throughì. The results provide an empirical insight into language contact phenomena involving specialized communication.


2017 ◽  
Vol 62 (2) ◽  
Author(s):  
Mohammad Hossein Haddadi

AbstractThere are many strategies and methods for translating texts of various genres. There are also strategies that are particularly useful for the translation of scientific Texts. When it comes to the translation of literary texts however there are plenty of thoughts, strategies, and methods that are proposed and which provoke a huge disagreement in this area. There is, particularly, a disagreement with regard to the translation of lyric poetry. Many critics and translation theorists agree that in the translation of lyric poetry (and generally in translation of literature texts) one should transfer, on the one hand, the style and the creative power of the author of the source language to the target language and, on the other hand, should meet the expectations of readers of the target language. They differ however over which strategy or method should be used in order to achieve this goal. When considering the translation of literature texts, and particularly the translation of lyric poetry, the spirit of the author and all existent relationships between form and contents should be transferred in the translation, since they have a close relationship with the spirit of the source language and thought of the author. The possibility of achieving this is however somewhat dubious because of the basic structural differences between the source language and the target language; and also because of those differences that may be between the content of the two mentioned languages. The following paper is based on the explanation of Rueckert’s idea of “world literature”, which is related to his thought about the transfer of the spirit of literature of the source language and the transfer of both form and contents. The paper seeks to answers this question: regarding the differences in the relationships of form and contents between the source language, Persian, and the German language, how can the spirit and content of Hafiz' lyrics be best transferred to the target language. It examines the structural and content differences between the Persian and German Languages and it illustrates the strategies and methods Rueckert uses to transfer the spirit and content of the poems of Hafiz to the German language.


Author(s):  
Ms Pratheeksha ◽  
Pratheeksha Rai ◽  
Ms Vijetha

The system used in Language to Language Translation is the phrases spoken in one language are immediately spoken in other language by the device. Language to Language Translation is a three steps software process which includes Automatic Speech Recognition, Machine Translation and Voice Synthesis. Language to Language system includes the major speech translation projects using different approaches for Speech Recognition, Translation and Text to Speech synthesis highlighting the major pros and cons for the approach being used. Language translation is a process that takes the conversational phrase in one language as an input and translated speech phrases in another language as the output. The three components of language-to-language translation are connected in a sequential order. Automatic Speech Recognition (ASR) is responsible for converting the spoken phrases of source language to the text in the same language followed by machine translation which translates the source language to next target language text and finally the speech synthesizer is responsible for text to speech conversion of target language.


2014 ◽  
Vol 6 (3) ◽  
pp. 381-389
Author(s):  
Doina Butiurca

AbstractOur research, Transparency and translatability of the terminological metaphor in the domain of the internet, is a contrastive analysis in the topic of the metaphor, especially. The relationship between the common and the special lexicon in the domain of the Internet in the English language as source language, the relationship between the common denominator between the source language and the semantic basis, of equivalence in the target language represent the aims of the research. The languages in which the analysis is carried out are different from the genealogical and typological point of view (the English language on the one hand, the Romance language and Hungarian on the other). The perspective is a descriptive-semasiological one, and the methods applied - the paradigmatic and syntagmatic analysis, the contrastive analysis - are adapted to this perspective. The transparency in the meaning, the degree of translatability, the motivated character of the terminological metaphor, the role of linguistics / of semantics in the terminology of the Internet are only some of the conclusions of the research.


Author(s):  
Feng Chen ◽  
Jian Yang ◽  
Lixuan Zhao

English as a second language is widely used in countries such as Malaysia and Indonesia, and it is common for English words to appear in Malay and Indonesian sentences. Malay and Indonesian have high homology and relatively few electronic language resources. We combine the corpus datasets of these two similar languages to design and implement a HMM–DNN-based cross-lingual speech synthesis system for Malay (including English words) and Indonesian (including English words). The methods used include: sharing synthesis units between Malay, Indonesian, and English, designing unified context attributes and question set in the process of acoustic model training, speaker-adaptive training with speech corpus of these three languages, and synthesizing speech using speaker-dependent Malay and Indonesian acoustic models. Experimental results show that the speech synthesis quality of the system is better than the traditional Hidden Markov model-based cross-lingual speech synthesis system.


Sign in / Sign up

Export Citation Format

Share Document